UvA-DARE (Digital Academic Repository) Surgery for colorectal … · patients with CRC. In Chapter...

UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Surgery for colorectal cancer: improving staging by the sentinel lymph node procedure

van der Zaag, E.S.

Link to publication

Citation for published version (APA):van der Zaag, E. S. (2011). Surgery for colorectal cancer: improving staging by the sentinel lymph nodeprocedure.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 29 Jun 2020

https://dare.uva.nl/personal/pure/en/publications/surgery-for-colorectal-cancer-improving-staging-by-the-sentinel-lymph-node-procedure(ab5f2866-4e71-4f72-9187-59050a061841).html

SURGERY FORCOLORECTAL CANCER

Improving staging by the sentinel lymph node procedure

E.S. van der Zaag

SURGERY FOR COLORECTAL CANCERImproving staging by the sentinel lymph node procedure

E.S. van der Zaag

The publication of this thesis was financially supported by ConvaTec, Welland Nederland BV, Johnson&Johnson, Covidien

Coverdesign/layout by In Zicht Grafisch Ontwerp, www.promotie-inzicht.nlPrinted by Ipskamp Drukkers B.V., www.ipskampdrukkers.nl

ISBN 978-90-9026386-1

© E.S. van der Zaag, 2011 No part of this thesis may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or otherwise without permission of the author.

SURGERY FOR COLORECTAL CANCERImproving staging by the sentinel lymph node procedure

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctoraan de Universiteit van Amsterdamop gezag van de Rector Magnificus

prof. dr. D.C. van den Boomten overstaan van een door het college voor promoties

ingestelde commissie,in het openbaar te verdedigen in de Aula der Universiteit

op vrijdag 4 november 2011, te 11:00 uur

door

Edwin Silvester van der Zaaggeboren te Amsterdam

Promotiecommissie

Promotor Prof. dr. W.A. Bemelman

Copromotoren Dr. C.J. Buskens Dr. W.H. Bouma

Overige leden Prof. dr. J.F. Lange Prof. dr. J.P. Medema Prof. dr. D.J. Richel Dr. P.J. Tanis Prof. dr. M.J. van de Vijver

Faculteit der Geneeskunde

Stay! speak, speak! I charge thee, speak!Horatio. Elsinore. (A platform before the castle)

Hamlet, Shakespeare 1603

Voor Sophie, Pelle, Annemein en Hester

Table of contents

General introduction and aim of this thesis ........................................................................................................................... 9

1 Systematic review of sentinel lymph node mapping in colorectal cancer. .....................................................19

E.S. van der Zaag, W.H. Bouma, P.J. Tanis, D.T. Ubbink, W.A. Bemelman, C.J. Buskens.

Submitted Annals of Surgical Oncology

2 Improving staging accuracy in colon and rectal cancer by sentinel lymph node mapping: ..............41 a comparative study.

E.S. van der Zaag, C.J. Buskens, N. Kooij, H. Akol, H.M. Peters, W.H. Bouma, W.A. Bemelman.

Eur J Surg Oncol. 2009 Oct;35(10):1065-70.

3 Diagnosing occult tumour cells and their predictive value in sentinel nodes of ........................................55

histologically negative colorectal carcinomas.

E.S. van der Zaag, N. Kooij, M.J. van de Vijver, W.A. Bemelman, H.M. Peters, C.J. Buskens.

Eur J Surg Oncol. 2010 Apr;36(4):350-7.

4 Categorization of occult tumour cells in lymph nodes in patients with colorectal cancer ...................73

not reliable enough.

E.S. van der Zaag, L. Welling, M.J. van de Vijver, W.A. Bemelman, H.M. Peters, C.J. Buskens.

Translated from Ned Tijdschr Geneeskd 2011; 155(7):293-8.

5 Decreased incidence of tumour cells after laparoscopic resection of colorectal cancer. .......................85

E.S. van der Zaag, H.M. Peters, C.J. Buskens, M.S. Vlug, W.A. Bemelman.

Surg Endoscopy 2011 Jun 24.[Epub]

6 Implications of sentinel lymph node mapping on nodal staging and prognosis in .................................97

colorectal cancer.

E.S. van der Zaag, W.H. Bouma, H.M. Peters, W.A. Bemelman, C.J. Buskens.

Accepted for publication in Colorectal Disease

7 The effect of colonoscopic tattooing on lymph node retrieval and sentinel node mapping. .........115 S.A.L. Bartels, E.S. van der Zaag, H.M. Peters, E. Dekker, C.J. Buskens, W.A. Bemelman.

Submitted Gastrointestinal Endoscopy

Summary, conclusions and future perspectives ............................................................................................................ 131

Samenvatting ....................................................................................................................................................................................... 137

Dankwoord ............................................................................................................................................................................................ 147

Curriculum vitae ................................................................................................................................................................................. 151

General introduction and outline of this thesis

11

general introduction and aim of this thesis

Colorectal cancer

Epidemiology In the Netherlands colorectal cancer (CRC) is the second most common type of cancer with more than 12.000 new patients in 2009. In 2008 approximately 4800 patients deceased from the effects of colorectal cancer, making it the second leading cause of death from cancer in the Netherlands after lung cancer. The incidence of colorectal cancer increases with one percent each year.1-3 In 2010 the Health Council of the Netherlands advised to start a screening program for colorectal cancer. If this program will start a further increase in incidence can be expected. Therefore colorectal cancer will remain a major health problem and all efforts must be made to optimize treatment strategy and increase survival of these patients.

Surgical treatment The initial treatment for patients with nonmetastatic CRC is surgical resection. In 2010 more than 8100 resection for primary CRC were performed in the Netherlands.4 The resection can either be performed via an open or laparoscopic approach. Over the years the number of laparoscopic resections increase: in 2010, 42% of the elective resections were laparoscopically performed.4 During surgery a complete resection of the primary tumor with en-bloc resection of the regional lymph node bearing the mesentery should be performed. The role of regional lymphadenectomy in CRC is well established: it involves local-regional control, cancer staging, adjuvant treatment planning, and it affects overall survival. Nodal involvement is the most important prognostic parameter, and is the pillar in consensus-driven treatment decision- making for adjuvant chemotherapy.5 The American Joint Committee on Cancer (AJCC) recommends at least 12 harvested lymph nodes per resection specimen for accurate nodal staging, but in daily clinical practice the nodal yield varies with over 50% of resection specimens containing fewer than 12 lymph nodes.6 In the Netherlands 22 % of the resection specimens contained less than ten lymph nodes.4 This relates to a clinically significant understaging in CRC.

Complete resection of early stage CRC without involvement of lymph node metastases should cure the patient. However, up to 30% of the patients with early node-negative CRC will have recurrent or metastatic disease following potentially curative resection.7,8 Apart from incomplete surgical resection, inadequate staging of CRC may be caused by insufficient pathologic regional nodal retrieval, sampling error or overlooked small volume nodal metastasis by conventional methods. Serial sectioning and additional immunohistochemical analyses or reverse transcriptase-polymerase chain reaction (RT-PCR) could diagnose lymphatic spread more accurately.9 Ideally, all regional lymph nodes should be examined with these

12 13

general introduction and aim of this thesis general introduction and aim of this thesis

staging without clearly defined prognostic impact and therapeutic implications so far. Numerous, generally small and single-institution, studies assessed the feasibility of SN with varying conclusions. The SN procedure for CRC has not been standardised, and the methods, materials, and patient selection vary by institution and surgeon.13-36

Outline of this thesis

In this thesis the role of the sentinel lymph node procedure in CRC surgery is described. The aim of our research is to assess the accuracy of SN-mapping in staging patients with CRC. In Chapter 1 we performed a systematic review to determine the diagnostic accuracy of this procedure from published data and to identify factors that contribute to the conflicting reports. In Chapter 2 we made a prospective comparison between colon and rectal cancer to analyse whether differences in anatomy and pre-operative treatment of rectal cancer affect the predictive value of the SN procedure. In addition, the incidence of micrometastatic disease in histologically negative lymph nodes was assessed, and the ability to refine staging by additional immunohisto chemical analysis compared to conventional histopathological examination. However, in the literature additional analysis is limited to the SN only. In Chapter 3 we analysed the real diagnostic accuracy of the SN procedure in upstaging patients. Therefore the incidence of OTC in SNs was compared to the presence of these cells in all histologically negative lymph nodes of these patients and related to clinical pathological characteristics. Also, three different antibodies used for the immuno histochemical analysis were compared to determine the sensitivity and specificity of these antibodies. The detected OTC are analysed and classified by the pathologist. This classification can be performed using the International Union Against Cancer (UICC) TNM-classification with both quantitative and qualitative criteria. According to this definition OTC are isolated tumour cells without clinical consequence if they are smaller than 0.2mm, or show no sign of activity (no proliferation, no desmoplastic stroma-reaction) and are localised in the lymphatic sinus. On the other hand if isolated tumour cells are detected at the parenchyma of the lymph node, they are considered micrometastases (pN1mi+). However, in the 6thedition of the American Joint Committee on Cancer (AJCC) TNM-classification only quantitative criteria are used. Therefore, in Chapter 4, we examined the interobserver variability among pathologists in classifying OTC in lymph nodes in CRC. Isolated tumour cells smaller than 0.2 mm are supposed to be of no clinical relevance. These cells have a limited life-span and levels of circulating tumour cells can be demonstrated in peripheral and portal blood during intra-operative manipulation of colorectal tumours. During a laparoscopic resection the tumour is not manipulated

techniques for occult tumour cells (OTC), but this would be too expensive and time-consuming and therefore not feasible in everyday practice.

Sentinel node procedure The sentinel-lymph-node concept (SN) could offer a solution.10 This concept is based on the stepwise lymphatic spread. The sentinel node procedure identifies lymph nodes most likely to harbour metastases. During the sentinel node procedure patent blue is injected around the tumour for lymph mapping. In this way lymphatic flow to the first blue lymph nodes can be visualised. These lymph nodes are identified as sentinel nodes. (Figure 1) This facilitates the pathologist to study only the few SNs removed in greater detail for tumour burden, compared to the conventional Haematoxyline and Eosin (H&E) staining of all lymph nodes which is currently done in routine daily practice. The SN procedure could refine staging, thereby identifying a patient group with tumour cells in the resected lymph nodes which is overlooked by conventional analysis. Potentially these patients have a higher risk for tumour recurrence or metastatic disease and therefore might benefit from adjuvant chemotherapy. In breast cancer and melanoma, the SN procedure is an important step in treatment strategy.11,12 If detailed analysis of the sentinel lymph node reveals no metastasis the lymphadenectomy will not be performed. In contrast to breast cancer and melanoma, regional lymphadenectomy is an integral part of the surgical procedure in CRC. Therefore, SN biopsy in CRC is only an adjunct potentially improving

Figure 1 Blue sentinel lymph node

14 15


References

1. Siesling S, van der Aa MA, Coebergh JW, et al. Time-space trends in cancer incidence in the Netherlands in

1989–2003.Int J Cancer 2008;122:2106–14.

2. Comprehensive Cancer Centres. http://www.cancerregistry.nl [Accessed 30.3.2011].

3. Elferink MA, Pukkala E, Klaase JM, Siesling S. Spatial variation in stage distribution in colorectal cancer in the

Netherlands. Eur J Cancer. 2011 Jul 29.

4. DSCA. Annual Report 2010 Outcome of care registration; transparency, quality and outcomes of care.

http://www.dsca.nl

5. Cohen Am, Kelsen D, Saltz L, et al. Adjuvant therapy for colorectal cancer. Curr Probl Cancer. 1997; 34:601-76.

6. Mitchell PJ, Ravi S, Grifftiths B, Reid F, Speake D, Midgley C, Mapstone N. Multicentre review of lymph node

harvest in colorectal cancer: are we understaging colorectal cancer patients? Int J Colorectal Dis. 2009

Aug; 24:915-21.

7. International multicentre pooled analysis of B2 colon cancer trials (IMPACT B2). J Clin Oncol 1999; 17:1356-63.

8. Figueredo A, Coombes ME, Mukherjee S. Adjuvant therapy for completely resected stage II colon cancer.

Cochrane Database Syst Rev 2008; 3:CD005390.

9. Koyanagi K, Bilchik AJ, Saha S, et al. Prognostic relevance of occult nodal micrometastases and circulating

tumor cells in colorectal cancer in a prospective multicenter trial. Clin Cancer Res 2008; 14:7391–96.

10. Cabanas RM. An approach for the treatment of penile carcinoma. Cancer 1977; 39:456-66.

11. Morton DL, Cochran AJ, Thompson JF, et al. Sentinel node biopsy for early-stage melanoma: accuracy and

morbidity in MSLT-I, an international multicenter trial. Ann Surg 2005; 242:311-3.

12. Veronesi U, Paganelli G, Viale G, et al. A randomized comparison of sentinel-node biopsy with routine

axillary dissection in breast cancer. N Engl J Med 2003; 349:546-53.

13. Matter M, Winckler M, Aellen S, Bouzourene H. Detection of metastatic disease with sentinel lymph node

dissection in colorectal carcinoma patients. Eur J Surg Oncol 2007; 33:1183-90.

14. Kelder W, Braat AE, Karrenbeld A, Grond JA, De Vries JE, Oosterhuis JW, Baas PC, Plukker JT. The sentinel

node procedure in colon carcinoma: a multi-centre study in The Netherlands. Int J Colorectal Dis 2007;

22:1509-14.

15. Bembenek AE, Rosenberg R, Wagler E, Gretschel S, Sendler A, Siewert JR, Nährig J, et al. Sentinel lymph

node biopsy in colon cancer: a prospective multicenter trial. Ann Surg 2007; 245:858-63.

16. Stojadinovic A, Nissan A, Protic M, et al. Prospective randomized study comparing sentinel lymph node

evaluation with standard pathologic evaluation for the staging of colon carcinoma: results from the United

States Military Cancer Institute Clinical Trials Group Study G1-01. Ann Surg 2007; 245:864-6.

17. Van Schaik PM, Van der Linden JC, Ernst MF, Gelderman WA, Bosscha K. Ex vivo sentinel lymph node

“mapping” in colorectal cancer. Eur J Surg Oncol 2007; 33:1177-82.

18. Bianchi PP, Ceriani C, Rottoli M, Torzilli G, Roncalli M, Spinelli A, Montorsi M. Laparoscopic lymphatic

mapping and sentinel lymph node detection in colon cancer: technical aspects and preliminary results.

Surg Endosc 2007; 21:1567-71.

19. Thomas KA, Lechner J, Shen P, Waters GS, Geisinger KR, Levine EA. Use of sentinel node mapping for cancer

of the colon: ‘to map or not to map’. Am Surg 2006; 72:606-11.

20. Smith J, Hwang H, Wiseman KW, Filipenko D, Phang PT. Ex vivo sentinel lymph node mapping in colon

cancer: improving the accuracy of pathologic staging? Am J Surg 2006; 191:665-8.

21. Saha S, Seghal R, Patel M, et al. A multicenter trial of sentinel lymph node mapping in colorectal cancer:

prognostic implications for nodal staging and recurrence. Am J Surg 2006;191:305-10.

22. Codignola C, Zorzi F, Zaniboni A, Mutti S, Rizzi A, Padolecchia E, Morandi GB. Is there any role for sentinel

node mapping in colorectal cancer staging? Personal experience and review of the literature. Jpn J Clin

Oncol 2005;35:645-50.

23. Braat AE, Oosterhuis JW, Moll FC, de Vries JE, Wiggers T. Sentinel node detection after preoperative

short-course radiotherapy in rectal carcinoma is not reliable. Br J Surg 2005;92:1533-8.

using the no-touch isolation technique. It can be hypothesised that during a laparoscopic resection less OTC are distributed into the lymphatic flow passing the lymphatic sinus. In Chapter 5 we assessed the effect of the surgical approach (i.e. open lateral to medial versus laparoscopic no touch medial to lateral approach) on the levels of OTC in sentinel lymph nodes of patients with early stage I and II CRC. Although upstaging of colorectal patient with SN procedure has been described, the clinical impact of these findings remains controversial because survival analyses of these patients are lacking. In Chapter 6 we describe our prospective analysis of the effect of SN mapping, in patients with CRC, on nodal staging and its prognostic impact in terms of disease recurrence and survival. During the SN procedure patent blue is injected around the tumour for lymph mapping. In pre-operative work up for CRC colonoscopic tattooing with Indian ink is performed to mark the tumour site before surgery. Colonoscopic tattooing is necessary to localise the tumour during surgery, particularly in laparoscopic resection where palpation of the tumour is not possible. Finally, in Chapter 7 we compared colonoscopic tattooing to the SN procedure and analysed if it can contribute to staging accuracy by analysing whether it leads to a higher lymph node yield per specimen, and to determine its diagnostic accuracy.

16 17


24. Cox ED, Kellicut D, Adair C, Marley K, Otchy DP, Peoples GE. Sentinel lymph node evaluation is technically

feasible and may improve staging in colorectal cancer. Curr Surg 2002;59:301-6.

25. Smith FM, Coffey JC, Khasri NM, et.al. Sentinel nodes are identifiable in formalin-fixed specimens after

surgeon-performed ex vivo sentinel lymph node mapping in colorectal cancer. Ann Surg Oncol

2005;12:504-9.

26. Bell SW, Mourra N, Flejou JF, Parc R, Tiret E. Ex vivo sentinel lymph node mapping in colorectal cancer. Dis

Colon Rectum 2005;48:74-9.

27. Paramo JC, Summerall J, Poppiti R, Mesko TW. Validation of sentinel node mapping in patients with colon

cancer. Ann Surg Oncol 2002;9:529-31.

28. Fitzgerald TL, Khalifa MA, Al Zahrani M, Law CH, Smith AJ. Ex vivo sentinel lymph node biopsy in colorectal

cancer: a feasibility study. J Surg Oncol 2002; 80:27-32.

29. Bendavid Y, Latulippe JF, Younan RJ, et al. Phase I study on sentinel lymph node mapping in colon cancer:

a preliminary report. J Surg Oncol 2002; 7981-4.

30. Joosten JJ, Strobbe LJ, Wauters CA, Pruszczynski M, Wobbes T, Ruers TJ. Intraoperative lymphatic mapping

and the sentinel node concept in colorectal carcinoma. Br J Surg 1999;86:482-6.

31. Bilchik AJ, Hoon DS, Saha S, et al. Prognostic impact of micrometastases in colon cancer: interim results of

a prospective multicenter trial. Ann Surg 2007; 246:568-75.

32. Bertagnolli M, Miedema B, Redston M, et al. Sentinel node staging of respectable colon cancer: results of a

multicenter study. Ann Surg 2004;240:624-8.

33. Faerden AE, Sjo OH, Andersen SN, et.al. Sentinel node mapping does not improve staging of lymph node

metastasis in colonic cancer. Dis Colon Rectum 2008; 51:891-6.

34. Yagci G, Unlu A, Kurt B, et al. Detection of micrometastases and skip metastases with ex vivo sentinel node

mapping in carcinoma of the colon and rectum. Int J Colorectal Dis 2007; 22:167-73.

35. Bembenek A, Schneider U, Gretschel S, Fischer J, Schlag PM. Detection of lymph node micrometastases

and isolated tumor cells I sentinel and nonsentinel lymph nodes of colon cancer patients. World J Surg

2005;29:1172-5.

36. Turner RR, Nora DT, Trocha SD, Bilchik AJ. Frequency and nature of cytokeratin-positive cells in sentinel and

nonsentinel lymph nodes. Arch Pathol Lab Med 2003;127:673-9.

1Systematic review of sentinel lymph node mapping in colorectal cancer

We determined the accuracy of this procedure from published data and identified

factors that contribute to the conflicting reports.

21

1systematic review of sentinel lymph node mapping in colorectal cancer

Introduction

Although improvements in screening and treatment have contributed to reduced disease-specific incidence and mortality, colorectal cancer (CRC) remains the third leading cause of cancer-related deaths.1 The primary treatment for nonmetastatic CRC is surgical resection of the primary tumour with en-bloc resection of the regional node bearing mesentery. The role for regional lymphadenectomy in CRC is well established: local-regional control, cancer staging, adjuvant treatment planning, and overall survival. Nodal involvement is the most important prognostic parameter, and is the pillar in consensus-driven treatment decision-making for adjuvant chemo-therapy.2 The American Joint Committee on Cancer (AJCC) recommends at least 12 harvested lymph nodes per resection specimen for accurate nodal staging, but in daily clinical practice the nodal yield varies with over 50% of resection specimens containing fewer than 12 lymph nodes.3 This relates to a clinically significant understaging in CRC. Up to 30% of early, presumably node-negative, patients will develop recurrences or distant metastases following potentially curative resection.4 Apart from incomplete surgical resection, inadequate staging of CRC may result from insufficient pathologic regional nodal retrieval, sampling error or overlooked small volume nodal disease. Serial sectioning and additional immunohistochemistry or reverse transcriptase(RT)-polymerase chain reaction (PCR) could diagnose lymphatic spread more accurately.5 Ideally, all regional lymph nodes should be examined with these techniques, but this would be too expensive and time consuming and therefore not feasible in everyday practice. The sentinel lymph node concept (SN) could offer a solution.6 This procedure allows the pathologist to study the few SNs removed in greater detail for tumour burden compared to the conventional H&E staining currently used in routine daily practice. Therefore, SN procedure could refine staging, possibly identifying a patient group benefitting from adjuvant chemotherapy. In contrast to breast cancer and melanoma treatment SN procedure is not routinely used for surgery in colorectal cancert7, 8. Regional lymphadenectomy is an integral part of the resection in CRC. Therefore, SN biopsy is an adjunct potentially improving staging without defined prognostic impact at this point in time. Numerous, generally small and single- institution studies assessed the feasibility of SN with varying conclusions. The SN procedure for CRC has not been standardized, and the methods, materials, and patient selection vary by institution and surgeon. In this report, we present a systematic review of all published studies of SN in CRC to analyze the diagnostic accuracy of this procedure. Our study provides a thorough assessment of the test performance characteristics of the SN procedure reported in the literature and explores the reasons for the observed heterogeneity in study results.

22 23

1chapter 1 systematic review of sentinel lymph node mapping in colorectal cancer

SN accuracy parameters were recalculated from the quantitative data presented in the original manuscript without taking immunohistochemical or PCR results into account. Standard definitions were used to facilitate comparison across studies. Outcome parameters may, therefore, differ from the original manuscript. The following definitions were used to describe the performance rates of SN biopsy:

Detection rate refers to the number of times a SN was actually identifiable = (number of successful attempts to retrieve a SN / number of attempts to retrieve a SN) x 100%. Accuracy rate refers to the ability of the SN to reflect the overall status of the lymphatic basin = (number of correct predictions of the nodal status by SN biopsy / number of patients with successful SN biopsy) x 100%. Sensitivity refers to the number of times the sentinel reflects the fact that nodal disease is present = (number of patients with tumour involved SNs / number of patients with any lymph node containing tumour) x 100% The false negative rate reflects the proportion of patients in whom no cancer was identified in the SN but who had nodal deposits found in their non-SNs compared to the total number of patients with nodal metastases = (number of false negative patients / number of true positive cases + number of false negative cases) x 100%. Upstaging rate refers to the number of cases in which additional serial sectioning and immunohistochemistry or PCR reveals tumour deposits in lymph nodes that would have been classified as pN0 with conventional staging techniques = (number of patients with micrometastases or isolated tumour cells / number of patients classified as pN0 with routine histopathological examination) x100%.

Assessment of methodological quality Each of the studies identified was assessed for validity criteria laid down by QUADRAS (an evidence base tool for the assessment of the quality of diagnostic studies).10 The criteria for validity were adjusted for this review and the following items were scored.1; prospective study, 2; consecutive patients, 3; specifications of inclusion and exclusion criteria 4; SN criteria and detection procedures (i.e. whether the procedure was described in sufficient detail to permit replication), 5; valid reference test (histology), 6; at least 20 procedures every year in each participating centre, 7; outcome parameters reduced to stage of disease or location of disease, 8; use of additional immunohistochemistry or PCR-techniques with classification of upstaging.

Statistical analysis A random effects model with an exact likelihood approach was used to calculate pooled SN accuracy parameters and 95% confidence intervals (CIs). The variation in sensitivity in the individual studies was displayed graphically as a forest plot using review manager version 5.0. Bivariate correlations between continuous measures

Material and methods

Literature search A systematic review of all published literature was undertaken independently by two investigators (ESZ and CJB) to identify reports regarding lymphatic mapping in CRC patients. The PubMed and Embase databases and the Cochrane Library were search until July 2011. The following expanded Medical Subject Headings terms were used: ‘sentinel node’, ‘lymphatic mapping’, ‘colon cancer’, ‘rectal cancer’, ‘colorectal cancer/tumo(u)r’. References from included studies, review articles and editorials were cross-checked for additional relevant publications. Data from meeting abstract were not studied as these were judged unlikely to present sufficient detail for data extraction required by our study protocol.

Inclusion and exclusion criteria for identified literature Only English language publications analysing lymphatic mapping in human patients with CRC were included. Duplicate articles based on the same group of patients were excluded. For follow-up studies that included a subset of previously reported patients, only the most recent article was included. Studies describing the use of indocyanine green were judged to be experimental and were therefore excluded. Finally, if quantitative results were not presented or SN performance parameters could not be extracted from the presented data, studies were also excluded.

Data extraction All data extraction was performed by two authors (ESZ and CJB) with cross- checking to ensure validation. When there was a discrepancy between the results, the original article was reanalysed. The fields for data capture were pre-specified before analysis and included extensive information on publication details, patient demographics, methodology, and SN efficacy by binary classification (i.e. detection, accuracy, sensitivity and false negative rates as well as negative predictive values). The quantitative results were used to build 2x2 contingency tables comprising true positive, true negative, and false negative. The term ‘false positive’ (and hence the calculation of specificity) is not appropriate because the presence of metastases in the SN confers node positivity. The term upstaging is used to describe the immu-nohistochemical or PCR findings in SNs in the absence of tumour-positive non-SNs. The American Joint Committee on Cancer (AJCC) definition for occult tumour cells was used, classifying micrometatases (lesions between 0.2mm and 2.0mm) as true upstaging to pN1 whereas patients with isolated tumour cells (ITC)(tumour cell deposits smaller than 0.2mm) are still considered pN0 (pN0itc+).9

24 25


SN identification rate In 3643 specimens (92.4%) one or more SN could be identified. The proportion of successful lymphatic mapping across studies ranged from 58% to 100%, with the majority of studies (41/57) demonstrating an identification rate over 90%. The detection rate in studies analysing colon carcinomas was significantly higher than studies including only rectal cancers (93.1% versus 83.1% resp., p=0.03). Identification rate was significantly higher in studies including more than 100 patients (mean 94.6%) compared to smaller studies (89.5%, p=0.02). In the 32 studies employing the in vivo method, the SN was significantly less often successfully identified than in the 18 studies with ex vivo procedures (89.2% and 93.7% resp., p=0.04). In vivo identification did not improve in the 11 studies were colloid was added to blue dye. Eighteen in vivo studies commented on aberrant lymphatic drainage with a mean rate of 3.9%.13, 24, 25, 27,

28, 30, 31, 34-36, 42, 44, 47, 49, 61-63, 67 No SNs were found outside the planned resection area in 11 studies. When analysing other parameters, no predictive factors for identification rate could be identified (i.e. multicentre studies, tracer used).

were based on the Pearson coefficient. To correlate sensitivity results to various study parameters, a linear regression model was used. For sensitivity analysis of individual patient data, a logistic-regression analysis was applied. When appropriate, cut-off points for continuous variables were selected a priori based on clinically relevant criteria or reporting convention. All data were processed with SPSS version 16.0, and a p-value of 0.05 or less was used as the level of statistical significance.

Results

Included studies The literature search yielded 98 publications on SN biopsy in humans with CRC between January 1999 (the year of the earliest series) and July 2011. Of these, two studies did not assess the SN procedure, two articles were not in English, 34 articles were either duplicate studies or had more recent data updating the principal studies, and from three articles the quantitative data could not be retrieved. The remainder of this analysis is based on the 57 included studies that were available for data extraction (Figure 1).11-68

Study characteristics Selected study characteristics from the 57 articles are presented in Table 1. In total, 3934 patients were enrolled across all studies with 3944 SN procedures performed. There were ten multicenter studies.23, 27-29, 31, 39, 41, 49, 54, 56 Most studies analysed a limited number of patients, with only ten studies including more than 100 patients. There were 21 studies that solely included patients with colon cancer and three studies that examined rectal cancer only. However, the majority of studies (33) included both patient with colon and rectal cancer. In 32 studies the in vivo technique was used, in 18 studies the ex vivo technique and in seven studies both methods were used to identify SNs. The two methods employed to identify SNs were blue dyes or radiolabeled tracers. Three blue dyes were used: patent blue dye V (25 studies), isosulfan blue (lymphazurin) 1% (15 studies) and 1% methylene blue (four studies). Two studies used 99mTc sulfur colloid for SN identification and 11 studies used a combination of blue dye and radioactive colloid.The mean number of harvested lymph nodes reported was 16.7, and ranged from 7.5 to 30.0 across studies. The overall average number of SNs identified was 2.8 (range between studies from 1.0 to 7.1). All studies had a prospective design and used a valid reference test (histology) with clear SN criteria and detection protocols. However, only 20 of the 57 eligible studies met at least three out of the five other validity criteria (consecutive patients, inclusion and exclusion criteria clearly described, >20 SN procedures per year, separate presentation of outcome parameters for colon versus rectum or early versus advanced carcinomas, use of additional upstaging techniques with classification of upstaging).11, 18-21, 24, 25, 28, 29, 32, 39, 41, 43-45, 56, 57, 59, 64, 67

Figure 1 Flow chart showing the selection and exclusion of publications.

Publications retrieved for detailedassessment

(n=98)

Publications analysed for SNperformance parameters

(n=60)

Not in English (n=2)

Duplicate/updates available (n=34)

Publications analysed(n=57)

No quantitative data available (n=3)

Medline and Embase searchSN studies colorectal cancer

(n=453)

Reports excluded based on title or abstract

1. review articles (n=94)2. irrelevant or non-comparative (n=251)3. indocyanine green tracer (n=10)

SN procedure not used (n=2)

26 27


Tabl

e 1

R

esul

ts o

f lym

pha

tic m

app

ing

in p

atie

nts

with

col

orec

tal c

ance

r

Stud

yN

o pa

tient

sCo

lon

Rect

umN

o ln

Trac

erM

etho

dN

o SN

True

pos

True

neg

Fals

e ne

gA

naly

sis

SNU

psta

ging

Vilc

ea11

4322

21N

RM

ethy

lene

blu

eIn

viv

o/ E

x vi

vo1,

08

195

HE,

IHC

of S

N

9.4%

Cera

nic12

45N

RN

R22

,9M

ethy

lene

blu

eEx

viv

o1,

714

237

HE,

IHC

of S

N

22,0

Rett

er13

3131

021

,5Pa

tent

blu

eIn

viv

o1,

34

168

HE,

IHC

of S

N

20,8

%

Fina

n1458

058

12,1

Isos

ulfa

n bl

ueEx

viv

o2,

215

277

HE,

IHC

of S

N

0

Som

mar

iva15

6954

1512

,4Pa

tent

blu

eEx

viv

o5,

012

469

HE,

IHC

of S

N

12,0

Dra

gan16

6060

0N

RIs

osul

fan

blue

Ex v

ivo

4,1

3226

0H

E, IH

C o

f SN

26

,9

Ivan

ov17

103

4855

NR

Pate

nt b

lue

In v

ivo

NR

4852

3H

E, IH

C o

f SN

20

,0

Nor

dgär

d1813

113

10

13,8

Pate

nt b

lue

Ex v

ivo

4,0

2983

13H

E, R

T-PC

R of

SN

21

,4

Van

der Z

aag19

132

100

3215

,4Pa

tent

blu

eEx

viv

o2,

033

1173

HE,

IHC

of S

N

28,8

Park

2069

4524

18,5

Met

hyle

ne b

ueIn

viv

o/ E

x vi

vo2,

526

276

HE,

IHC

of S

N

18,5

Cha

n2131

1912

7,5

Met

hyle

ne b

lue

In v

ivo

1,3

1111

3H

E, IH

C o

f SN

0

Qua

dros

2252

2230

19,0

Pate

nt b

lue/

99

mTc

In v

ivo

3,5

1516

8H

E, IH

C o

f SN

37

,5

Faer

den23

199

200

013

,0Pa

tent

blu

eIn

viv

o4,

032

125

28H

E, IH

C o

f al

l ln

29,8

Lim

2412

012

00

20,0

Isos

ulfa

n bl

ue/

99m

Tc

In v

ivo

4,0

2971

20H

E, IH

C o

f SN

11,3

Sand

rucc

i2535

305

8,7

Pate

nt b

lue/

99

mTc

In v

ivo

2,2

2111

3H

E, IH

C o

f SN

36

,4

Köks

al26

1913

610

,0Is

osul

fan

blue

In v

ivo/

Ex

vivo

1,8

213

3H

E, IH

C o

f SN

18

,7

Keld

er27

6969

011

,0Pa

tent

blu

eIn

viv

o2,

315

493

HE,

IHC

of S

N

18,4

Bem

bene

k2831

531

50

20,0

Pate

nt b

lue

In v

ivo

2,0

7415

638

HE,

IHC

of S

N

21,3

Stoj

adin

ovic

2984

840

18,2

Isos

ulfa

n bl

ueEx

viv

o2,

718

568

HE,

IHC

of S

N26

,8

Mat

ter30

5236

1230

,0Pa

tent

blu

eIn

viv

o2,

78

3010

HE,

IHC

of

all l

n19

,4

Tiffe

t3164

4915

18,0

Pate

nt b

lue/

99

mTc

In v

ivo

2,8

1235

12H

E, IH

C o

f SN

5,

7

Van

scha

ik32

4427

1710

,5Pa

tent

blu

eEx

viv

o5,

026

160

HE,

IHC

of S

N

30,3

Mur

awa33

2713

14N

RPa

tent

blu

eIn

viv

oN

R2

221

HE,

IHC

of S

N

9,1

Libe

rale

3411

871

4720

,0Pa

tent

blu

eIn

viv

o/ E

x vi

vo2,

022

7614

HE,

IHC

of S

N

9,5

Cova

relli

3520

200

17,1

Pate

nt b

lue/

99

mTc

In v

ivo

1,3

612

1H

E, IH

C o

f SN

7,

7

Bian

chi36

2222

023

,3Pa

tent

blu

eIn

viv

o2,

35

161

HE,

IHC

of S

N

12,5

Yagc

i3747

2027

18,6

Pate

nt b

lue

Ex v

ivo

5,9

1627

4H

E, IH

C o

f SN

14

,8

Thom

as38

6963

615

,8Is

osul

fan

blue

In v

ivo

2,1

1238

14H

E, IH

C o

f SN

5,3

Terw

issc

ha39

5356

09,

0Pa

tent

blu

e/

99m

TcIn

viv

o2,

212

352

HE,

IHC

of

all l

n14

,8

Smith

4017

170

16,0

Isos

ulfa

n bl

ueEx

viv

o1,

88

08

HE,

IHC

of S

N

10

Saha

4150

040

892

15,0

Isos

ulfa

n bl

ueIn

viv

o2,

218

621

121

HE,

IHC

of S

N

26,1

Tuec

h4230

340

20,0

Pate

nt b

lue

In v

ivo/

Ex

vivo

1,8

1020

2H

E, IH

C o

f SN

12

,0

Khaf

agy43

530

53N

RPa

tent

blu

eIn

viv

oN

R31

88

HE,

IHC

of

all l

n46

,0

28 29


Tabl

e 1

C

ontin

ued

Codi

gnol

a4456

524

21,0

Pate

nt b

lue

In v

ivo

2,0

3713

6H

E, IH

C o

f SN

37

,5

Braa

t4591

5734

7,7

Pate

nt b

lue

Ex v

ivo

1,8

2351

8H

E, IH

C o

f SN

10

,5

Smith

4640

298

16,9

Isos

ulfa

n bl

ueEx

viv

o4,

410

236

HE,

IHC

of S

N24

,0

Dah

l4730

300

17,4

Pate

nt b

lue/

99

mTc

In v

ivo

2,2

1018

2H

E on

ly

Bell48

5846

1229

,9Pa

tent

blu

eEx

viv

o2,

99

3314

HE,

IHC

of S

N

and

equa

l no

of o

ther

ln

6,1

Bert

agno

lli49

7272

017

,3Is

osul

fan

blue

In v

ivo

2,1

1042

14H

E, s

eria

l se

ctio

ns o

f al

l ln

Dem

irbas

5041

2516

8,6

Pate

nt b

lue

Ex v

ivo

3,0

1817

2H

E, IH

C o

f SN

11

,8

Won

g5112

411

012

30,0

Isos

ulfa

n bl

ueEx

viv

o3,

827

6924

HE,

IHC

of

all l

n27

,3

Patt

en52

5757

014

,4Is

osul

fan

blue

/ 99

mTc

In v

ivo

3,5

1431

11H

E, IH

C o

f SN

14

,3

Bem

bene

k5348

048

18,0

99m

TcEx

viv

o3,

07

309

HE,

IHC

of S

N

0

Bert

oglio

5426

206

12,4

Pate

nt b

lue

In v

ivo

2,9

715

2H

E, s

eria

l se

ctio

n of

SN

on

ly

Rose

ano55

2314

914

,6Pa

tent

blu

eIn

viv

o/ E

x vi

vo2,

52

192

HE,

IHC

of S

N0

Vieh

l5631

310

21,6

Isos

ulfa

n bl

ueIn

viv

o2,

56

156

HE,

IHC

of S

N

NR

Bilc

hik57

120

102

1814

,0Is

osul

fan

blue

/ 99

mTc

In v

ivo

1,8

3773

5H

E, IH

C o

f SN

29

,5

Brod

eric

k-Vi

lla58

5046

5N

RIs

osul

fan

blue

In v

ivo/

Ex

vivo

1,5

1027

10H

E, IH

C o

f SN

NR

Kita

gaw

a5956

1937

23,9

99m

TcIn

viv

o3,

518

294

HE

only

Nas

tro60

88

0N

RPa

tent

blu

e/

99m

TcIn

viv

oN

R4

20

HE,

IHC

of S

N

50%

Para

mo61

5555

012

,1Is

osul

fan

blue

In v

ivo

1,9

1430

1H

E, IH

C o

f SN

20,0

%

Cox62

1717

017

,5Is

osul

fan

blue

In v

ivo/

Ex

vivo

5,5

314

0H

E, IH

C o

f SN

24

,0

Gan

dy63

198

1116

,0Pa

tent

blu

eEx

viv

o7,

15

61

HE

only

Esse

r6431

265

15,0

Isos

ulfa

n bl

ueIn

viv

o1,

72

151

HE

only

Mer

rie65

2526

017

,5Pa

tent

blu

e/

99m

TcIn

viv

o3,

04

163

HE,

RT-

PCR

of

all l

n25

,0

Cse

rni66

25N

RN

R15

,5Pa

tent

blu

eIn

viv

o4,

08

115

HE

only

Joos

ten67

5044

614

,0Pa

tent

blu

eIn

viv

o3,

08

1512

HE,

IHC

of S

N

13,3

SN =

sen

tin

el ly

mp

h n

od

e, H

E =

hae

mat

oxyl

in a

nd

eosi

n st

aini

ng

, IH

C =

imm

unoh

isto

chem

istr

y, R

T-PC

R =

real

-tim

e p

oly

mer

ase

chai

n re

acti

on, N

R =

not

rep

orte

d

30 31


Test performance measures The pooled sensitivity of the recalculated outcome parameters was 69.6% (64.7-74.6), with an accompanying false negative rate of 30.4 (25.5-35.3). The overall pooled accuracy of the SN procedure was 88.2% (86.0-90.3). The sensitivity ranged from 33.3% to 100% across studies (Figure 2). There was a strong correlation between the sensitivity and the number of identified SNs (Pearson correlation 0.37, p=0.007, Figure 3), with a good predictive accuracy for lymph node involvement when four or more SNs were identified (mean sensitivity <4 SNs = 66.3% versus ≥4 SNs = 85.2%, p=0.003). There was no relation between the sensitivity and the method of SN detection (ex vivo versus in vivo) or tracer used (blue versus radioactive colloid). Apart from the 24 studies only analysing colon or rectal malignancies, another 11 studies presented the separate sensitivity results for the two types of cancer.11, 15, 9-12, 32,

37, 41, 45, 64 Combing the outcome parameters a significantly higher sensitivity of SN procedures in colon cancer was found (77,6% versus 65.7% for rectal carcinomas, p=0.04). In 6 studies, sensitivity results could be calculated separately for early (T1/ T2) and advanced (T3/ T4) carcinomas.15, 21, 31, 41, 53, 59 Statistically, a significantly higher sensitivity was seen in the early group compared to advanced carcinomas (93.4% versus 58.8%, p=0.01), but the number of patients with positive lymph nodes in early carcinomas in these studies was small. Unfortunately, the reported parameters associated with heterogeneity in the sensitivity results could not be analysed in a prediction model since different studies commented on different variables. Overall the sensitivity of 20 high quality studies was higher than the other studies, although not statistically significant (57.6% versus 66.4% resp., p=0.07). (Table 2)

Upstaging In 46 studies, immunohistochemistry staining was performed on specimens histologically classified as pN0.11-17, 19-46, 48, 50-53, 55, 57, 60-62, 67 Most studies only analyzed the SN when the haematoxylin and eosin staining results were negative. In two studies RT-PCR on the SN was used after a negative haematoxylin and eosin staining result.18,

65 Sectioning and staining of the lymph nodes was not uniformly undertaken. Serial sectioning was used in most of the studies with intervals ranging from 20–500 μm with a large variety of monoclonal and polyclonal antibodies used (e.g. cytokeratin markers, cell surface glycoproteins, tumour specific proteins). Mean upstaging was 18.9% (0-50%). However, only ten studies classified these finding into micrometastases or isolated tumour cells according to the AJCC.13, 19, 22, 27, 28,

31, 34, 36, 38, 59 In these studies the true upstaging rate was significantly lower with 7.7% (0-15.4%). If the immunohistochemical findings of the SN were included in the accuracy parameters, an increased mean sensitivity of 80.2% was found (20 studies15, 18, 19,

11-15, 30, 33, 35, 39, 43-46, 48, 49, 52, 57, 60).

Figure 2 Weighted sensitivity for the 57 sentinel lymph node studies included in the meta-analysis.

TP= true positive, FP= false positive FN= false negative TN= true negative

32 33


Discussion

Our meta-analysis of 57 studies analyzing the SN procedure for CRC shows an overall acceptable identification rate (92%). The higher detection rate in studies including over 100 patients indicates the existing of a learning curve. Furthermore, technical issues may influence successful SN detection given the superiority of the ex vivo technique over the in vivo method. Usually, the in vivo technique is propagated since this procedure has the advantage of identifying aberrant lymphatic drainage with the possibility to adjust the planned resection. However, in the 18 studies analyzing aberrant drainage, a SN outside the planned resection margins was only found in 4% of the patients. With a minority of these nodes being tumour-positive, it will have limited impact on staging. The addition of radiocolloid to blue dye did not improve the in vivo results. Also considering the complexity of the procedure, these observations make ex vivo mapping the method of choice.

Table 2 Sensitivity results of sentinel lymph node mapping in patients with colorectal cancer.

Sensitivity (%)

95% CI Sensitivity (%)

95% CI p-value

Overall results

HE analysis(57 studies, n=3934)

69.6 64.7 - 74.6

HE + IHC analysis(20 studies, n=1477)

80.2 74.3 – 86.1

Subgroup analysisEx vivo 18 studies, n=1169

72.9 63.1 – 80.9 In vivo32 studies, n=2415

68.3 61.4 – 75.3 0.5

Blue dye44 studies, n=3246

69.4 63.7 – 75.2 Colloid13 studies, n=688

70.4 59.5 – 81.2 0.9

Number of SN ≥ 4(10 studies, n=434)

85.2 73.5 – 96.9 Number of SN < 4(45 studies, n=3320)

66.3 61.2 – 71.4 0.003

Colon cancer(31 studies, n=2224)

77.6 71.3 – 83.8 Rectal cancer(14 studies, n=468)

65.7 54.7 – 76.7 0.04

Early carcinoma6 studies, n=156

93.1 73.7 - 100 Advanced carcinoma(6 studies, n=612

58.8 37.1 – 80.5 0.01

High validity study(20 studies, n=2009)

75.6 68.5 – 82.6 Low validity study(37 studies, n=1925)

66.4 59.9 – 73.0 0.07

Figure 3 Scatter plot illustrating the correlation between the sensitivity of the sentinel lymph node (SN) procedure across studies and the number of identified SNs (Pearson correlation 0.37, p=0.004). A fitted linear-regression equation is shown with 95% confidence intervals.

34 35


patients. Another report demonstrated that pN0 patients with or without isolated tumour cells in lymph nodes show similar survival rates, whereas patients with micro-metastases had lower survival rates.79 As long as the prognostic significance is not sorted, the AJCC recommends additional treatment only in patients with micrometas-tases and the 19% upstaging found in this study should be regarded as overestimation. Apart from improving staging by additional staining, SN mapping has also been described to increase the yield of harvested lymph nodes with corresponding upstaging. The number of lymph nodes analysed has been recognized as a prognostic factor for a long time.80 It has been demonstrated that SN mapping results in an increased proportion of N1 patients with a corresponding better prognosis of the pN0 patient group81, 82 which would be an additional reason to recommend SN mapping in patients with early staged colorectal cancer. It should be borne in mind that, in contrast to breast cancer and melanomas, this procedure is not used for therapeutic purposes but mainly to refine staging. The SN procedure is quite safe (especially ex vivo) and the procedure is not a difficult technique, a learning curve of 20–30 cases is described.18 A major drawback of our study is the tremendous diversity across reports in patient selection, technical details of the SN procedures, and pathological analysis. The results should therefore be interpreted with caution. As with any meta-analysis, the possibility of publication bias should be taken into account.

Conclusion

In conclusion, our meta-analysis demonstrates an overall disappointing sensitivity of SN mapping in colorectal patients. However, in early staged colon cancer the SN procedure has acceptable accuracy rates and refines staging. We recommend that SN mapping should always be considered in addition to conventional resection in colon cancer.

A low pooled sensitivity of 69.6% for predicting lymph node metastases was found with an accompanying false negative rate of 30.4%. Therefore pathological examination of only the SN cannot replace routine examination of the complete mesentery. However, subgroups could be identified with higher sensitivity rates. Sensitivity improved when the number of identified SNs was higher. The latter has been previously established as independent predictive factors of false negative mapping in a prediction model created with Bayesian Network Analysis.69 Another factor predictive of sensitivity in our study was depth of invasion. Early (T1 and T2) carcinomas had higher accuracy parameters when compared to advanced carcinomas. Our results are in line with a review of two prospectively maintained databases describing a sensitivity of 89% in for T1/ T2 carcinomas. A recent review with comparable overall sensitivity rates did not find a relation with T-stage, but the authors stratified for individual T-stages in colon and rectal cancer separately which might yield different results.70,71 The lower sensitivity in advanced cancers is probably due to obstruction of afferent lymph vessels or nodes by tumour growth, changing the lymphatic drainage. Since the aim of SN mapping is to refine staging, high accuracy is less important in advanced stage CRC that already meet criteria for adjuvant chemotherapy. Because of progress in diagnostic technology and screening programmes, diagnosis of CRC will occur at earlier stages.72, 73 It is especially important in these early cancers to identify the small subgroup of high risk patients who may benefit from adjuvant systemic treatment. The additional value of SN mapping is most debatable in rectal cancer. Both identification rate and sensitivity were lower in comparison to colon cancer. Neo-adjuvant treatment in rectal cancer may change tumour status of the SN and hampers lymph node retrieval in general. Furthermore, the clinical consequences of node-positivity in rectal cancer, especially after neo-adjuvant chemo radiotherapy, are less clear than in colon cancer with respect to adjuvant systemic treatment. Previously, we demonstrated that occult tumour cells are predominantly found in the SN.74 The mean upstaging rate of 19% found in this meta-analysis with most studies only analyzing the SN is therefore probably an accurate estimate of the percentage of patients with occult tumour cells. However, only ten studies classified these cells in isolated tumour cells or micrometastases (12,3% pN0itc+ versus 7,7% pN1mi+) as recommended by the AJCC staging manual, which makes the upstaging results difficult to interpret.9 The prognostic value of isolated tumour cells is still unclear. A reduced survival is described associated with the presence of occult tumour cells in a consecutive series of patients with stage II colon cancer, suggesting also clinical significance of isolated tumour cells.75 This would be in line with the prognostic role of isolated tumour cells established in breast, colon, and prostate cancer patients.76-78 However, it is conceivable that isolated tumour cells in patients with distant metastasis have different prognostic value when compared to the pN0

36 37


24. Lim SJ, Feig BW, Wang H, et al. Sentinel lymph node evaluation does not improve staging accuracy in colon

cancer. Ann Surg Oncol 2008; 15:46-51.

25. Sandrucci S, Mussa B, Goss M, et al. Lymphoscintigraphic localization of sentinel node in early colorectal

cancer: results of a monocentric study. J Surg Oncol 2007; 96:464-9.

26. Köksal H, Bostanci H, Mentes BB. Importance of sentinel lymph nodes in colorectal cancer: a pilot study.

Adv Ther 2007; 24:583-8.

27. Kelder W, Braat AE, Karrenbeld A, et al. The sentinel node procedure in colon carcinoma: a multi-centre

study in The Netherlands. Int J Colorectal Dis 2007; 22:1509-14.

28. Bembenek AE, Rosenberg R, Wagler E, et al. Sentinel lymph node biopsy in colon cancer: a prospective

multicenter trial. Ann Surg 2007; 245:858-63.






31. Tiffet O, Kaczmarek D, Chambonniere ML, et al. Combining radioisotopic and blue-dye technique does not

improve the false-negative rate in sentinel lymph node mapping for colorectal cancer. Dis Colon Rectum

2007; 50:962-70.

32. Van Schaik PM, Van der Linden JC, Ernst MF, Gelderman WA, Bosscha K. Ex vivo sentinel lymph node

“mapping” in colorectal cancer. Eur J Surg Oncol 2007; 33 :1177-82.

33. Murawa D, Filas V, Breborowicz J, Spychala A, Dworzecka K, Murawa P. Evaluation of the sentinel node

biopsy in colorectal carcinoma including the results of immunohistochemical examinations. Acta Chir Belg

2007; 107:45-8.

34. Liberale G, Lasser P, Sabourin JC, et al. Sentinel lymph nodes of colorectal carcinoma : reappraisal of 123

cases. Gastroenterol Clin Biol 2007; 31:281-5.

35. Covarelli P, Cristofani R, Boselli C, et al. Preliminary study on radioguided sentinel node identification in

colon cancer. Am Surg 2007; 73:222-6.


mapping and sentinel lymph node detection in colon cancer : technical aspects and preliminary results.

Surg Endosc 2007; 21:1567-71.

37. Yagci G, Unlu A, Kurt B, et al. Detection of micrometastases and skip metastases with ex vivo sentinel node

mapping in carcinoma of the colon and rectum. Int J Colorectal Dis 2007; 22:167-73.

38. Thomas KA, Lechner J, Shen P, Waters GS, Geisinger KR, Levine EA. Use of sentinel node mapping for cancer

of the colon: ‘to map or not to map’. Am Surg 2006; 72:606-11.

39. Terwisscha Van Scheltinga SE, Den Boer FC, Pijpers R, et al. Sentinel node staging in colon carcinoma: value

of sentinel lymph node biopsy with radiocolloid and blue staining. Scand J Gastroenterol Suppl 2006;

243:153-7.


cancer: improving the accuracy of pathologic staging? Am J Surg 2006; 191:665-8.


prognostic implications for nodal staging and recurrence. Am J Surg 2006; 191:305-10.

42. Tuech JJ, Pessaux P, Di Fiore F, Nitu V, Lefebure B, Colson A, Michot F. Sentinel node mapping in colon

carcinoma: in-vivo versus ex-vivo approach. Eur J Surg Oncol 2006; 32:158-61.

43. Khafagy W, El-Dawoody A, El-Ghawalby N, El-Shobaky M, Atwan N. Ultrastaging of rectal cancer based on

identification of micrometastases in sentinel lymph node. Coloproctology 2005; 27:153–60.



Oncol 2005; 35:645-50.

References

1. Cancer Facts and Figures 2007. Atlanta, GA: American Cancer Society; 2007.

2. Cohen Am, Kelsen D, Saltz L, et al. Adjuvant therapy for colorectal cancer. Curr Probl Cancer. 1997; 34:601-76.

3. Mitchell PJ, Ravi S, Grifftiths B, Reid F, Speake D, Midgley C, Mapstone N. Multicentre review of lymph node

harvest in colorectal cancer: are we understaging colorectal cancer patients? Int J Colorectal Dis. 2009 Aug;

24:915-21.

4. Figueredo A, Coombes ME, Mukherjee S. Adjuvant therapy for completely resected stage II colon cancer.

Cochrane Database Syst Rev 2008; 3:CD005390.

5. Koyanagi K, Bilchik AJ, Saha S, et al. Prognostic relevance of occult nodal micrometastases and circulating

tumor cells in colorectal cancer in a prospective multicenter trial. Clin Cancer Res 2008; 14:7391–96.

6. Cabanas RM. An approach for the treatment of penile carcinoma. Cancer 1977; 39:456-66.


morbidity in MSLT-I, an international multicenter trial. Ann Surg 2005; 242:311-3.


axillary dissection in breast cancer. N Engl J Med 2003; 349:546-53.

9. Singletary SE, Greene FL, Sobin LH. Classification of isolated tumor cells: Clarification of the 6th edition of the

American Joint Committee on Cancer Staging manual. Cancer 2003; 98:2740-1.

10. Whiting P, Rutjes AW, Reitsma JB, et.al. The development of QUADAS: a tool for the quality assessment of

studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003; 3:25.

11. Vîlcea ID, Vasile I, Mirea CS, et al. Sentinel lymph node study in colorectal cancer using serial sectioning and

Hematoxylin-Eosin staining: importance and limitations. Rom J Morphol Embryol. 2011; 52:379-83.

12. Ceranic MS, Kecmanovic DM, Pavlov MJ, Nale DP, Micev MT, Kovacevic PA, Stamenkovic AB. Validation and

feasibility of ex vivo sentinel lymph node “mapping” by methylene blue in colorectal cancer. Hepatogastro-

enterology 2010; 57:1113-8.

13. Retter SM, Herrmann G, Schiedeck TH. Clinical Value of Sentinel Node Mapping in Carcinoma of the Colon.

Colorectal Dis. 2010; 13:855-9.

14. Finan KR, Lewis JS Jr, Winslow E, Mutch MG, Birnbaum EH, Fleshman JW. Ex vivo sentinel lymph node

mapping in patients undergoing proctectomy for rectal cancer. Dis Colon Rectum. 2010; 53:243-50.

15. Sommariva A, Donisi PM, Gnocato B, Vianello R, Stracca Pansa V, Zaninotto G. Factors affecting false-negative

rates on ex vivo sentinel lymph node mapping in colorectal cancer. Eur J Surg Oncol 2010; 36:130-4.

16. Dragan R, Nebojsa M, Dejan S, et al. Clinical application of sentinel lymph node biopsy for staging,

treatment and prognosis of colon and gastric cancer. Hepatogastroenterology 2009; 56:1606-11.

17. Ivanov K, Kolev N, Ignatov V, Madjov R. Intraoperative sentinel lymph node mapping in patients with

colorectal cancer. Hepatogastroenterology 2009; 56:99-105.

18. Nordgard O, Oltedal S, Korner H, Aasprong OG, Tjensvoll K, Gilje B, Heikkila R. Quantitative RT-PCR detection

of tumor cells in sentinel lymph nodes isolated from colon cancer patients with an ex vivo approach. Ann

Surg 2009; 249:602-7.

19. Van der Zaag ES, Buskens CJ, Kooij N, Akol H, Peters HM, Bouma WH, Bemelman WA. Improving staging

accuracy in colon and rectal cancer by sentinel lymph node mapping: a comparative study. Eur J Surg

Oncol 2009; 35:1065-70.

20. Park JS, Chang IT, Park SJ, et al. Comparison of ex vivo and in vivo injection of blue dye in sentinel lymph

node mapping for colorectal cancer. World J Surg 2009; 33:539-46.

21. Chan SH, Ng C, Looi LM. Intraoperative methylene blue sentinel lymph node mapping in colorectal cancer.

ANZ J Surg 2008; 78:775-9.

22. Quadros CA, Lopes A, Araujo I, Fregnani JH, Fahel F. Upstaging benefits and accuracy of sentinel lymph

node mapping in colorectal adenocarcinoma nodal staging. J Surg Oncol 2008; 98:324-30.

23. Faerden AE, Sjo OH, Andersen SN, et al. Sentinel node mapping does not improve staging of lymph node

metastasis in colonic cancer. Dis Colon Rectum 2008; 51:891-6.

38 39


70. Van der Pas MH, Meijer S, Hoekstra OS, et al. Sentinel-lymph-node procedure in colon and rectal cancer: a

systematic review and meta-analysis. Lancet Oncol 2011; 12:540-50.

71. Cahill RA, Bembenek A, Sirop S, et al. Sentinel node biopsy for the individualization of surgical strategy for

cure of early-stage colon cancer. Ann Surg Oncol 2009; 16:2170-80.

72. De Wijkerslooth TR, De Haan MC, Stoop EM, , et al. Study protocol: population screening for colorectal

cancer by colonoscopy or CT colonography: a randomized controlled trial. BMC Gastroenterol 2010;

19;10-47.

73. Atkin WS, Edwards R, Kralj-Hans I, et al. UK Flexible Sigmoidoscopy Trial Investigators. Once-only flexible

sigmoidoscopy screening in prevention of colorectal cancer: a multicentre randomised controlled trial.

Lancet 2010; 375:1624-33.

74. Van der Zaag ES, Kooij N, Van de Vijver MJ, Bemelman WA, Peters HM, Buskens CJ. Diagnosing occult tumour

cells and their predictive value in sentinel nodes of histologically negative patients with colorectal cancer.

Eur J Surg Oncol 2010; 36:350-7.

75. Bukholm IR, Bondi J, Wiik P, Nesland JM, Andersen SN, Bakka A, Bukholm G. Presence of isolated tumour cells

in mesenteric lymph nodes predicts poor prognosis in patients with stage II colon cancer. Eur J Surg Oncol

2003;29:862–6.

76. Cristofanili M, Budd GT, Ellis MJ, et al. Circulating tumor cells, disease progression, and survival in metastatic

breast cancer. N Engl J Med 2004; 351:781-91.

77. Cohen SJ, Punt CJ, Iannotti N, et al. Prognostic significance of circulating tumor cells in patients with

metastatic colorectal cancer. Ann Oncol 2009; 20:1223-9

78. Moreno JG, Miller MC, Gross S, Allard WJ, Gomella LG, Terstappen LW. Circulating tumor cells predict survival

in patients with metastatic prostate cancer. Urology 2005; 65:713-8.

79. Messerini L, Cianchi F, Cortesini C, Comin CE. Incidence and prognostic significance of occult tumor cells in

lymph nodes from patients with stage IIA colorectal carcinoma. Hum Pathol 2006; 37:1259-67.

80. Hashiguchi Y, Hase K, Ueno H, Mochizuki H, Kajiwara Y, Ichikura T, Yamamoto J. Prognostic significance of

the number of lymph nodes examined in colon cancer surgery. Ann Surg 2010; 251:872-81.



States Military Cancer Institute Clinical Trials Group Study GI-01. Ann Surg 2007; 245:846-57.

82. Van der Zaag ES, Bouma WH, Peters HM, Bemelman WA, Buskens CJ. Implications of sentinel lymph node

mapping for nodal staging and prognosis in colorectal cancer. Accepted for publication in Colorectal Disease


short-course radiotherapy in rectal carcinoma is not reliable. Br J Surg 2005; 92:1533-8.

46. Smith FM, Coffey JC, Khasri NM, et al. Sentinel nodes are identifiable in formalin-fixed specimens after sur-

geon-performed ex vivo sentinel lymph node mapping in colorectal cancer. Ann Surg Oncol 2005;

12:504-9.

47. Dahl K, Westlin J, Kraaz W, Winqvist O, Bergkvist L, Thörn M. Identification of sentinel nodes in patients with

colon cancer. Eur J Surg Oncol 2005; 31:381-5.


Colon Rectum 2005; 48:74-9.

49. Bertagnolli M, Miedema B, Redston M, et al. Sentinel node staging of resectable colon cancer: results of a

multicenter study. Ann Surg 2004; 624:628-30.

50. Demirbas S, Ince M, Baloglu H, Celenk T. Should sentinel lymph node mapping be performed for colorectal

cancer? Turk J Gastroenterol 2004; 15:39-44.

51. Wong JH, Johnson DS, Namiki T, Tauchi-Nishi P. Validation of ex vivo lymphatic mapping in hematoxylin-

eosin node negative carcinoma of the colon and rectum. Ann Surg Oncol 2004; 11:772-7.

52. Patten LC, Berger DH, Rodriquez-Bigas M, et al. A prospective evaluation of radiocolloid and immunohisto-

chemical staining in colon carcinoma lymphatic mapping. Cancer 2004; 15:2104-9.

53. Bembenek A, Rau B, Moesta T, et al. Sentinel lymph node biopsy in rectal cancer: not yet ready for routine

clinical use. Surgery 2004; 135:498-505.

54. Bertoglio S, Sandrucci S, Percivale P, et al. Prognostic value of sentinel lymph node biopsy in the pathologic

staging of colorectal cancer patients. J Surg Oncol 2004; 85:166-170.

55. Roseano M, Scaramucci M, Ciutto T, et al. Sentinel lymph node mapping in the management of colorectal

cancer: preliminary report. Tumori 2003; 89:412-6.

56. Viehl CT, Hamel CT, Marti WR, et al. Identification of sentinel lymph nodes in colon cancer depends on the

amount of dye injected relative to tumor size. World J Surg 2003; 27:1285-90.

57. Bilchik AJ, Nora DT, Sobin LH, Turner RR, Trocha S, Krasne D, Morton DL. Effect of lymphatic mapping on the

new tumor-node-metastasis classification for colorectal cancer. J Clin Oncol 2003; 21:668-72.

58. Broderick-Villa G, Ko A, O’Connell TX, Guenther JM, Danial T, DiFronzo LA. Does tumor burden limit the

accuracy of lymphatic mapping and sentinel node biopsy in colorectal cancer? Cancer J 2002; 8:445-50.

59. Kitagawa Y, Watanabe M, Hasegawa H, et al. Sentinel node mapping for colorectal cancer with radioactive

tracer. Dis Colon Rectum 2002; 45:1476-80.

60. Nastro P, Sodo M, Dodaro CA, Gargiulo S, Acampa W, Bracale U, Renda A. Intraoperative radiochromoguided

mapping of sentinel lymph node in colon cancer. Tumori 2002; 88:352-3.




feasible and may improve staging in colorectal cancer. Curr Surg 2002; 59:301-6.

63. Gandy CP, Biddlestone LR, Roe AM, O’Leary DP. Intra-operative injection of Patent Blue V dye to facilitate

nodal staging in colorectal cancer. Colorectal Dis 2002; 4:447-9.

64. Esser S, Reilly WT, Riley LB, Eyvazzadeh C, Arcona S. The role of sentinel lymph node mapping in staging of

colon and rectal cancer. Dis Colon Rectum 2001; 44:850-4.

65. Merrie AE, van Rij AM, Phillips LV, Rossaak JI, Yun K, McCall JL. Diagnostic use of the sentinel node in colon

cancer. Dis Colon Rectum 2001; 44:410-7.

66. Cserni G, Vajda K, Tarjan M, Bori R, Svebis M, Baltas B. Nodal staging of colorectal carcinomas from

quantitative and qualitative aspects. Can lymphatic mapping help staging? Pathol Oncol Res 1999; 5:291-6.

67. Joosten JJ, Strobbe LJ, Wauters CA, Pruszczynski M, Wobbes T, Ruers TJ.

68. Intraoperative lymphatic mapping and the sentinel node concept in colorectal carcinoma. Br J Surg 1999;

86:482-6.

69. Nissan A, Protic M, Bilchik A, Eberhardt J, Peoples GE, Stojadinovic A. Predictive model of outcome of

targeted nodal assessment in colorectal cancer. Ann Surg 2010; 251:265-74.

2Improving staging accuracy in colon and rectal cancer by sentinel lymph node mapping: a comparative study

We analysed whether differences in anatomy and pre-operative treatment of rectal

cancer affect the predictive value of the sentinel lymph node procedure in comparison

to patients with colon cancer.

43

2

improving staging accuracy in colon and rectal cancer by sentinel lymph node mapping

Introduction

lymph node status is the most significant predictor for recurrence and survival in colorectal cancer patients. However, 10-25% of patients with stage I or II (absence of lymph node metastases; according to the American Joint Committee on Cancer (AJCC)) will have local recurrences or distant metastases within five years.1 Metastatic nodes can be missed if an insufficient number of lymph nodes has been examined, or if occult metastases have not been identified by standard histopathological examination. Current data show that adequate staging of colon cancer requires examination of at least 12 lymph nodes per patients.2 However, to achieve an 85% probability of true node negativity even 40 lymph nodes should be examined.3 This has prompted the use of sentinel lymph node (SN) mapping for colorectal cancer in an effort to select the nodes most likely to harbour metastases. This technique accurately predicted nodal status in more than 90%, and resulted in upstaging 10 to 30% of patients with stage I or II disease to stage III disease4, 5, thereby making these patients candidates for adjuvant chemotherapy. The rectum is anatomically quite different from the colon, with a more bulky mesentery and an infraperitoneal location that makes the in vivo visualisation of the SN procedure difficult. Therefore, ex vivo SN mapping is generally performed in these carcinomas. In the Netherlands, preoperative short-course radiotherapy is widely used in rectal cancer because it lowers the local recurrence rate. 6 Although radiotherapy does not alter tumour and node classification, it can induce fibrosis and cause lymphoid tissue to disappear.7, 8 In contrast to breast cancer, where subsequent lymph nodes will only be harvested when the SN shows metastases, the main purpose of SN mapping in patients with colorectal cancer is to use this technique for a more concentrated pathologic assessment of lymph nodes that are most likely to harbour metastases. Occult tumour cells have been detected in up to 40% of histologically negative nodes in patients with colorectal cancers. The AJCC nowadays recommends discrimination of occult tumour cells in micrometastases and isolated tumour cells, since the latter do not seem to have prognostic significance.9,10

Detection of occult tumour cells can be done using immunohistochemistry or polymerase chain reaction (PCR) techniques. However, since PCR techniques do not allow morphological determination of identified staining and will detect isolated tumour cells, micrometastases, and benign marker-positive mesothelial cells with equal sensitivity, immunohistochemistry is the preferred technique and most frequently used.11-13

The aim of this study was to investigate the accuracy of SN-mapping in patients with colorectal cancer. Specifically we want to analyse whether differences in anatomy and pre-operative treatment of rectal cancer affect the predictive value of the SN

44 45

2

chapter 2 improving staging accuracy in colon and rectal cancer by sentinel lymph node mapping

Pathological examination Immediately after resection, the fresh specimen was processed using a standardised protocol. Tumours were staged according to the AJCC TNM classification 2002.9 All lymph nodes (SNs and non-SNs) identified by the pathologist were collected in separate boxes and marked according to location, then cut in two with both sides stained with haematoxylin and eosin and evaluated for tumour involvement. If the SN was negative for metastases, serial sectioning was performed at three levels of 500mm intervals. Three sections of each level were stained with three monoclonal antibodies. The anti-epithelial cell antibody Ber-EP4 (DAKO, The Netherlands), was combined with two anti-cytokeratin antibodies: the anti-CK20 antibody, with its expression limited to gastrointestinal epithelial cells (Euro Diagnostica, Arnhem, The Netherlands), and the anticytokeratin marker Cam5.2 directed against cytokeratin 7 and 8 expressed in all epithelial cells (Becton and Dickinson, Alphen aan den Rijn, The Netherlands).14 Appropriate positive and negative controls were added on each automated run, to confirm the sensitivity and specificity of the antibodies (sections of colonic carcinoma tissue served as positive controls; negative controls were obtained by omitting the primary antibody).

Occult tumour cells Tumour deposits within lymph nodes were classified and staged according to the revised guidelines set by the AJCC.10 Tumour cell deposits between 0.2mm and 2.0mm were referred to as micrometastases and those smaller than 0.2mm as isolated tumour cells. False-positive non-neoplastic haematopoeitic cells (e.g. reticular cells and plasma cells which can also show staining for cytokeratins), were discriminated from isolated tumour cells by microscopic morphological and nuclear differences. The immunohistochemically stained slides were evaluated blind and independently by two experienced pathologists, who were unaware of the clinical data.

Statistical analsysisThe following definitions were used for calculation:- Identification rate (%): number of patients with successfully retrieved SN X 100 /

number of patients enrolled.- Accuracy (%): number of patients with correct prediction of nodal status X 100 /

number of patients with successfully retrieved SN.- Sensitivity (%): number of patients with tumour involved SN X 100 / number of

patients with successfully retrieved SN and macrometastases in any lymph node.- Negative predictive value (%): number of nodal negative patients with successfully

retrieved SN X 100 / (number of nodal negative patients + number of false-negative patients).

procedure in rectal cancer in comparison to patients with colon cancer. In addition, the incidence of micrometastatic disease in histologically negative nodes was assessed, and the ability to refine staging by immunohistochemistry compared to conventional histopathological examination.

Material and methods

Patients Patients with colorectal cancer who were operated on with curative intent between November 2006 and May 2008 were considered for inclusion in the study. Exclusion criteria were invasion of other organs (T4 carcinomas), or two adjacent colorectal carcinomas. Five patients with locally advanced rectal cancer undergoing neoadjuvant chemoradiotherapy were excluded since this treatment downstages tumour and lymph node status, thereby introducing possible false negative findings. Six patients who underwent a colon resection for benign disease (e.g. Crohn’s disease or diverticulosis), were used as a negative control group. Resection was either performed via an open or laparoscopic approach. Patients with rectal cancer underwent total mesorectal excision (TME) after preoperative radiotherapy comprising five fractions of five Gray. Patients were generally operated on within five days, but no later than 10 days after the last fraction of radiation. A small number of patients with rectal cancer did not receive radiotherapy due to imminent occlusion or the pre-operative diagnosis of villous adenoma with high grade dysplasia. The study was done in accordance with the guidelines of the local ethics committee.

Sentinel lymph node mapping An ex vivo SN technique was used in all patients. After standard resection, 0.5-2 ml (depending on the volume of the tumour) patent blue V (Guerbet, Gorinchem, the Netherlands) was injected around the tumour with the colonic specimen left intact, after which the injection site was gently massaged. For colon carcinomas, the mesocolon was inspected and the first one to four blue lymph nodes were identified as SNs and either dissected or marked with a suture. For rectal carcinomas, the identification of blue nodes was done at the department of pathology to keep the circumferential resection margin intact. In order not to have the results affected by this discrepancy in SN identification, the specimen was sent to the pathologist in fresh state immediately after resection.

46 47

2


Predictive value of SN mapping A SN was identified in 117 of the 132 patients. Identification of at least one SN was possible in 92 of 100 patients with colon cancer and in 25 of the 32 patients with rectal cancer (p=0.03). Overall, the SN accurately predicted nodal status in 106 patients with 11 false negative SN procedures (accuracy 91%). There was a significant discrepancy in staging accuracy for colon and rectal carcinomas (95% versus 76%, respectively; p=0.005). The sensitivity of the SN procedure was also higher in colon carcinomas (83%) than in rectal carcinomas (57%), although this was not statistically significant (p=0.06). The SN was negative in 67 patients with colon cancer and 17 patients with rectal cancer. In the colon carcinoma group five patients had a false negative procedure, whereas six patients in the rectal carcinoma group had a false negative procedure, resulting in negative predictive values of 93% and 65% respectively (p=0.002) (Table 2). In both groups more unidentified or false negative SNs were seen with an increasing number of positive lymph nodes (p=0.01) and with depth of tumour infiltration (T-stage), although the latter was not significant. Eleven of the 13 unsuccessful or false negative SN procedures occurred in patients with rectal cancer who had had pre-operative short-course radiotherapy (p=0.05) (Table 2a and 2b).

Identification of micrometastatic disease in pN0 carcinomas A total of 62 colon cancer patients and 11 rectal cancer patients classified as nodal negative by routine H&E staining with successfully retrieved SN underwent serial step sectioning and focussed immunohistochemical analysis of the SN. In eight patients a micrometastasis was detected in the sinus of the SN, resulting in an upstaging rate of 11%. Another 13 patients revealed isolated tumour cells, resulting in an overall upstaging rate of 29%. There was no significant difference in overall upstaging between patients with colon cancer (27%) or rectal cancer (20%). The pathologists agreed in 67 of the 73 patients (92%) All micrometastases were identified by both pathologists independently. In six patients there was a disagreement about the occurrence of isolated tumour cells. The presence of micrometastases was significantly related to lymphangio invasion (p<0.0001), whereas this correlation could not be found for the occurrence of isolated tumour cells and lymphangio invasion (p=0.4). Although six out of the eight patients with micrometastases had a tumour invading the adventitia (T3), no correlation with depth of tumour infiltration and the presence of micrometastases could be demonstrated (p=0.7) (Table 3). No occult tumour cells were found in the SN of the negative control group.

Results are expressed as percentages or mean ± SD. All statistical analyses were performed using SPSS version 14.0 (SPSS INC., Chicago, IL, USA). The association between clinicopathological features and detection rate, negative predictive value, sensitivity, and the presence of micrometastases was analysed using Student's t-test (continuous variables) or Chi-squared test (categorical variables). p-values of 0.05 or less were considered statistically significant.

Results

Patient and tumour characteristics A total of 100 patients with colon cancer and 32 patients with rectal cancer were included. Of the 32 patients with a rectal cancer, 23 received pre-operative short-course radiotherapy. Baseline clinicopathological characteristics were comparable for both groups, except for the mean number of lymph nodes and SNs analysed per patient which was higher for patients with colon cancer (Table 1).

Table 1 Baseline characteristics of the 132 included patients with colorectal cancer.

Characteristics Colonn=100

Rectumn=32

p-value

Age (yrs) 70.4±15 66.4±12 0.2

Gender Male 42 17 0.3

Female 58 15

Type of resection Open 64 17 0.3

Laparoscopic 36 15

AJCC Classification Dukes 1 22 7 0.08

Dukes 2 45 8

Dukes 3 33 17

Total lymph nodes identified 16.3±8 12.4±7 0.01

Total SN identified 2.3±3 1.1±1 0.04

48 49

2


Discussion

Predicitive value of SN mapping This study demonstrates that SN mapping can reliably predict nodal status in patients with colon carcinomas. A significantly lower identification rate, accuracy, and negative predictive value of SN procedures was found for patients with rectal cancer particularly after radiotherapy. A recent review describes varying accuracy rates of SN lymph node mapping with accuracy rates between 78 and 100% and sensitivity rates varying between 25 and 100%.6 Interestingly, most studies with disappointing results included both patients with colon and rectal cancer.15-18 Our findings of the inferior results of the SN procedure in patients with rectal cancer can be an explanation for the conflicting results in previous studies. There are several possible reasons why the SN technique is less reliable in patients with rectal cancer. In contrast to the mesorectum, the mesocolon can easily be inspected from two sides after resection which facilitates identification of the SN. In addition, the mesorectum has a perirectal fascia and this circumferential resection margin should be kept intact during the resection for adequate pathological examination and to reduce the risk of local recurrence, making SN identification more difficult.6,19 In our study, SN identification of rectal carcinomas was done at the department of pathology immediately after resection. Not directly identifying SNs after patent blue injection harbours the risk of increasing the number of sentinel nodes by including second echelon lymph nodes.20 Analysis of a large number of SNs by serial step immunohistochemistry is impractical. However, our study demonstrates that even after postponed SN identification in rectal carcinoma, the number of SNs is still significantly lower than for colon carcinomas.

Table 2a The predictive value of SN mapping in percentages.

Overalln=132

Colon cancern=100

Rectal cancern=32

p-value

Detection rate 89 92 78 0.03

Accuracy 91 95 76 0.005

Sensitivity 75 83 57 0.06

Negative predictive value 87 93 65 0.002

Table 3 Correlation of SN micrometastases and clinicopathological findings in 73 conventional N0 patients with colorectal cancer.

No micrometastasesn=65

Micrometastases n=8

p-value

Depth of tumour invasion T1 6 0 0.7

T2 17 2

T3 42 6

Differentiation grade Well 7 0 0.6

Moderate 48 6

Poor 10 2

Lymphangio-invasion No 63 3 0.0001

Yes 2 5Table 2b Factors influencing SN accuracy in 100 patients with colon cancer

and 32 patients with rectal cancer.

SN correct

SN not identified/ incorrect

p-value

Colon cancer n=100 n=87 n=13Depth of tumour invasiona T1 21 1 0.06

T2 41 4

T3 25 8

Number of positive nodesb N0 62 5 0.01

N1 16 3

N2 9 5

Tumour size mean cm 5.1 4.8 0.7

Rectal cancer n=32 n=19 n=13

Depth of tumour invasion T1 3 4 0.03

T2 8 0

T3 8 9

Number of positive nodes N0 11 4 0.01

N1 6 4

N2 2 5

Tumour size mean cm 3.3 3.7 0.4

Pre-operative radiotherapy No 7 2 0.05

Yes 12 11

50 51

2


have been published with conflicting results. Six of the 12 reported studies found that the presence of nodal micrometastases correlates with significantly worse survival or recurrent disease, whereas the other six show no relation with clinical outcome parameters.5 Comparison of results between studies is hampered by the substantial variation in detection method, staining protocols and antibodies. RT-PCR is more sensitive in detecting marker-positive cells and less subject to sampling error and inter-observer variation when compared to immunohistochemistry.11,12 However, this technique does not allow morphological determination, and will detect isolated tumour cells, hyperplastic epithelial cells and benign marker-positive mesothelial cells with equal sensitivity. In addition, polyclonal antibodies are frequently used, which have been demonstrated to be associated with more false-positive staining.13 To prevent false positive staining, immunohistochemistry with three different monoclonal antibodies was used to detect micrometastatic disease in this study. Another explanation for the conflicting data is the variation in terminology and definitions of micrometastatic disease. In some studies all nodes with positive immuno histochemical findings are included, while more recent studies consider isolated tumour cells different from micrometastases, as suggested by the revised 6th edition of the TNM system by the AJCC, since clusters of cells smaller than 0.2 mm seem to have a different level of clinical significance.30 Although our study was not designed to comment on the prognostic significance of micrometastatic disease, the relation between micrometastases and lymphangio invasion does suggest a clinical relevance, especially since a correlation between lymphangio invasion and isolated tumour cells could not be established.

Conclusion

We show that SN mapping can accurately predict nodal status in patients with colon cancer. Immunohistochemical analysis of the SN detects micometastases in 11% of N0 patients, possibly identifying a patient group that may benefit from adjuvant therapy. The results suggest that SN mapping with focussed histopatho-logical examination should be considered as a mandatory step towards optimal staging in patients with colon cancer. In contrast, SN mapping is less reliable in patients with rectal cancer who underwent TME after pre-operative short course radiotherapy. To improve staging in rectal cancer, other studies should be performed investigating other techniques or tracers. Further studies must focus on univocal definitions of occult tumour cells and survival analysis of patients with accurately staged stage I and II colorectal cancer.

Another possible reason why the SN identification is unsuccessful in patients with rectal cancer is pre-operative radiotherapy. Radiotherapy has been demonstrated to induce fibrosis, causing the already small rectal lymph nodes to disappear.7,8,21 This might alter lymphatic flow and lead to false-positive staining of non-SNs. In addition, it can also explain the significant lower number of SNs found in rectal carcinomas leading to decreased accuracy. So far, only a few other studies with a small number of patients have concentrated on SN mapping in rectal carcinomas after neo-adjuvant therapy with similar disappointing results.22,23 These results suggest that a different approach should be sought for SN mapping for patients with rectal cancer after pre-operative radiotherapy.

Factors influencing false negative results Our study demonstrates better results of SN mapping in patients with colon cancer, with a sensitivity of 83% and a negative predictive value of 93%. In literature however, there is a persistent controversy over the value of SN lymph node mapping in these patients. False negative findings are attributable to a number of factors including extent of the disease, mapping technique, timing and method of pathological processing, number of SN lymph nodes evaluated, and method of ultrastaging. Especially local tumour invasion of adjacent lymph nodes has been indicated as a predictor of false negative SN mapping24,25, which is confirmed by our data in which a strong correlation between false negative results and number of involved lymph nodes was found. In contrast to our study, other studies also suggest a negative association with tumour volume and number of SNs identified.25,26 However, these false negative procedures can be regarded as technical failures. Viehl et al. identified dye volume as a significant predictor of successful SN identification and demonstrated that when the volume is adjusted to tumour size (0.5 ml per 1.0 cm tumour) technical failures can be prevented.24 Therefore, standardisation of working definitions, mapping technique and pathological processing are critical to the success of SN lymph node mapping for patients with colon cancer. We used the ex vivo SN procedure since only this technique can be performed easily in rectal carcinomas. The ex vivo technique has been demonstrated to be as accurate as the in vivo technique.27,28 In addition, it facilitates standardised, uniform specimen processing and assessment, recognising that the principal limitation of this technique is its inability to detect the few (2-10%) cases with aberrant lymphatic drainage.28,29

Detection of occult tumour cells Although almost all studies analysing SN mapping for colorectal cancer demonstrate the detection of micrometastases, the clinical significance of this micro-metastatic disease is still unknown. A number of studies with respect to this subject

52 53

2


23. Braat AE, Oosterhuis JW, De Vries JE, Tollenaar RA. Lymphatic staging in colorectal cancer: pathologic,

molecular, and sentinel node techniques. Dis Colon Rectum 2005; 48:371-83.

24. Viehl CT, Hamel CT, Marti WR, et al. Identification of sentinel lymph nodes in colon cancer depends on the

amount of dye injected relative to tumor size. World J Surg 2003;27:1285-90.


evaluation with standard pathlogic evaluation for the staging of colon carcinoma. Ann Surg 2007;

245:846-857.

26. Stojadinovics A, Allen PJ, Protic M, et al. Colon sentinel lymph node mapping: practical surgical applications.

J Am Coll Surg 2005; 201:297-313.

27. Teuch JJ, Pessaux P, Fiore FD, et al. Sentinel node mapping in coloncarcinoma : in-vivo versus ex-vivo

approach. Eur J Surg Oncol 2006 ; 32 :158-161

28. Wood TF, Saha S, Morton DL, et al. Validation of lymphatic mapping in colorectal cancer: in vivo, ex vivo, and

laparoscopic techniques. Ann Surg Oncol 2001; 8:150-157.

29. Patten LC, Berger DH, Rodriquez-Bigas M, et al. A prospective evaluation of radiocolloid and immunohisto-

chemical staining in colon carcinoma lymphatic mapping. Cancer 2004; 100:2104-2109.



References


2. Goldstein NS, Sanford W, Copffey M, et al. Lymph node recovery from colorectal resection specimens

removed for adenocarcinoma. Am J Clin Pathol 1996; 106:209-16.

3. Joseph NE, Sigurdson ER, Hanlon AL, et al. Accuracy of determining nodal negativity in colorectal cancer on

the basis of the number of nodes retrieved on resection. Ann Surg Oncol 2003; 10:213-18.

4. Saha S, Wiese D, Badin J, et al. Technical details of sentinel lymph node mapping in colorectal cancer and

its impact on staging. Ann Surg Oncol 2000; 7:120-4.

5. De Haas RJ, Wicherts DA, Hobbelink MG, et al. Sentinel lymph node mapping in colon cancer: current

status. Ann Surg Oncol 2007; 14:1070-80.

6. Kapiteijn E, Marijnen CA, Nagtegaal ID, et al. Preoperative radiotherapy combined with total mesorectal

excision for resectable rectal cancer. N Engl J Med 2001; 345:638-46.

7. Marijnen CA, Nagtegaal ID, Klein KE, et al. No downstaging after short-term preoperative radiotherapy in

rectal cancer patients. J Clin Oncol 2001; 19:1976-84.

8. Nagtegaal ID, Marijnen CA, Kranenbarg EK, et al. Short-term preoperative radiotherapy interferes with the

determination of pathological parameters in rectal cancer. J Pathol 2002; 197:20-7.

9. Greene FL, Page DL, Fleming ID. AJCC Cancer staging manual (ed 6). New York, NY, Springer, 2002.



11. Kelder W, Braat AE, Van den berg A, et.al. Value of RT-PCR analysis of sentinel nodes in determining the

pathological nodal status in colon cancer. Anticancer Res 2007; 27: 2855-9.

12. Doekhie FS, Kuppen PJ, Peeters KC, et al. Prognostic relevance of occult tumour cells in lymph nodes in

colorectal cancer. Eur J Surg Oncol. 2006;32:253-8.

13. Borgen E, Beiske K, Trachsel S, et al. Immunocytochemical detection of isolated epithelial cells in bone

marrow: non-specific staining and contribution by plasma cells directly reactive to alkaline phosphatase. J

Pathol 1998;185:427-434.

14. Buskens CJ, Ten Kate FJ, Obertop H, Izbicki JR, van Lanschot JJ. Analysis of micrometastatic disease in

histologically negative lymph nodes of patients with adenocarcinoma of the distal esophagus or gastric

cardia. Dis. Esophagus 2008; 21:488-95

15. Joosten JJ, Strobbe LJ, Wauters CA, et al. Intraoperative lymphatic mapping and the sentinel node concept

in colorectal carcinomas. Br J Surg 1999; 86:482-486.

16. Saha S, Bilchik A, Wiese D, et al. Ultrastaging of colorectal cancer by sentinel lymph node mapping

technique – a multicenter trial. Ann Surg Oncol 2001; 8:94S-98S.

17. Wong JH, Steineman S, Calderia C, Bowles J, Namiki T. Ex vivo sentinel node mapping in carcinoma of the

colon and rectum. Ann Surg 2001; 233:515-521.

18. Tiffet O, Kaczmarek D, Chambonniere ML, et al. Combining radioisotopic and blue-dye technique does not

improve the false-negative rate in sentinel lymph node mapping for colorectal cancer. Dis Colon Rectum

2007; 50 : 962-70.

19. Quirke P, Durdey P, Dixon MF, Williams NS. Local recurrence of rectal adenocarcinoma due to inadequate

surgical resection. Histopathological study of lateral tumour spread and surgical excision. Lancet 1986;

ii:996-999.

20. Gandy CP, Biddlestone LR, Roe AM, O’Leary DP. Intra-operative injection of patent blue V to facilitate nodal

staging in colorectal cancer. Colorectal Dis 2002; 4:447-449.

21. Topor B, Acland R, Kolodko V, et al. Mesorectal lymph nodes: their location and distribution within the

mesorectum. Dis Colon Rectum 2003; 46:779-785.

22. Baton O, Lasser P, Sabourin JC, et al. Ex vivo lymph node study for rectal adenocarcinoma : preliminary

study. World J Surg 2005; 29:1166-70.

3Diagnosing occult tumour cells and their predictive value in sentinel nodes of histologically negative patients with colorectal cancer

The incidence of occult tumour cells in sentinel lymph nodes was compared to the

presence of these cells in all histologically negative lymph nodes.

57

3

diagnosing occult tumour cells in lymph nodes

Introduction

Sentinel lymph node (SN) mapping is an increasingly popular technique to improve staging accuracy for patients with colorectal cancer by identifying nodes with the highest likelihood to harbour metastatic disease. This technique has shown promising results for colon carcinomas with SN mapping accurately predicting nodal status in 80-100%.1-24 An advantage of SN mapping is the potential to use this technique for a more accurate and cost-effective pathologic assessment of lymph node status. The additional stepwise sectioning combined with immunohistochemical analysis of the SN has been demonstrated to detect occult tumour cells (OTC) in up to 40% of histologically negative patients with colorectal cancer1-24, possibly identifying a patient group that may benefit from adjuvant systemic therapy. However, previously reported results are difficult to interpret because most studies publishing SN results include immunohistochemical analysis of the SN only. This might result in flattering negative predicting values and sensitivity rates. Moreover, most studies published so far have not subdivided the finding of OTC into micrometastases and isolated tumour cells (ITC) on the basis of their dimensions as recommended by the American Joint Committee on Cancer (AJCC) staging manual.25 This possibly results in an overestimation of the number of upstaged patients and makes the clinical significance of the detected OTC difficult to evaluate. The widespread use of SN immunohistochemical analysis in patients with breast cancer has demonstrated that stage migration due to SN mapping harbours the danger of potential misleading effects and artefacts in prognostication.26 This highlights the need to differentiate correctly between clinically relevant (micro)metastases (0.2-2 mm)(pN1mi+) and immunohistochemically detected isolated tumour cells without prognostic significance (<0.2mm)(pN0itc+). This is particularly so since the main purpose of SN mapping in patients with colorectal cancer is to establish a more accurate staging of patients at risk of recurrence due to lymphogenic spread and who could benefit from adjuvant chemotherapy.

For clinical application with therapeutical consequences, an immunohistochemical marker for OTC has to be both highly sensitive (i.e. detect the majority of tumour cells), and specific (i.e. absence of false positive staining of cells). Since most studies use different antibodies without a gold standard and uniform definition of OTC, the antibody of first choice for the detection of clinically relevant micrometastases in patients with colorectal cancer is still difficult to determine on the basis of published data. We performed this study to evaluate the real diagnostic accuracy of upstaging N0 patients with colorectal cancer by the SN procedure with immunohistochemical analysis. Therefore the incidence of OTC in SNs was compared to the presence of

58 59

3

chapter 3 diagnosing occult tumour cells in lymph nodes

of blue nodes was performed at the department of pathology immediately, to keep the circumferential resection margin intact.

Immunohistochemical staining The surgical resection specimens were analysed at the pathology department using a standardised protocol. All lymph nodes (SNs and non-SNs) were collected in separate boxes and marked according to location, then cut in two with both sides stained with hematoxylin and eosin and evaluated for tumour involvement. Of the 58 histologically N0 patients, serial sectioning was performed at 500μm intervals of all lymph nodes from formalin-fixed and paraffin-embedded archival tissue blocks. Three serial sections at three separate levels were immunohistochemically stained with three different monoclonal antibodies: the anti-epithelial cell antibody Ber-EP4 (DAKO, The Netherlands), was combined with two anti-cytokeratin antibodies: the anti-CK20 antibody, with its expression limited to gastrointestinal epithelial cells (Euro Diagnostica, Arnhem, The Netherlands), and the anticytokeratin marker Cam5.2 directed against cytokeratin 7 and 8 expressed in all epithelial cells (Becton and Dickinson, Alphen aan den Rijn, The Netherlands). Appropriate positive and negative controls were added on each automated run, to confirm the sensitivity and specificity of the antibodies (sections of colonic carcinoma tissue served as positive controls; negative controls were obtained by omitting the primary antibody). The staining procedures have been described in detail previously.29

Microscopic evaluation and definitions All slides were reviewed by two independent experienced pathologists who were

unaware of the clinical data. The type of involvement (single tumour cells or clusters), topography (lymph-angio invasion and/or lymph sinuses and/or parenchyma) and extension of the lesion was assessed in each case (Figure 1). Occult tumour cells were classified as either an ITC (diameter < 0.2 mm) or a mi-crometastasis (diameter between 0.2 mm and 2 mm). Patients were then restaged according to the AJCC classification.25 OTC comprised of micrometastases (0.2-2 mm)(pN1mi+) and ITC (<0.2mm)(pN0itc+). In case of disagreement, the slides were re-evaluated in a consensus meeting. False-positive non-neoplastic haematopoeitic cells (e.g. reticular cells and plasma cells which can also show staining for cytokeratins), were discriminated from ITC on the basis of histopathologic features.

Statistical analyses The following definitions were used for calculations (Figure 2). Identification rate is the number of patients with one or more SN identified (b) divided by the total number of procedures (a) x 100%. Sensitivity is the number of patients with an immunohistochemically positive SN (c) divided by the total number of patients with

these cells in all histologically negative lymph nodes of these patients. Two experienced pathologists independently assessed the presence of lymph node metastases. The detection of OTC was analysed by three different antibodies to determine the sensitivity and specificity of these antibodies. In addition, the presence of OTC was correlated to clinicopathological parameters.

Patients and methods

Patients Between November 2006 and July 2007, an ex vivo SN procedure was performed in 100 patients with colorectal carcinoma who were operated on with curative intent. Exclusion criteria for SN mapping were invasion of other organs (T4 carcinomas), or two simultaneous colorectal carcinomas. Patients with locally advanced rectal cancer who underwent neoadjuvant chemoradiotherapy were excluded because of stage migration before surgical treatment. Of the 100 patients 42 patients were shown to have macro metastases in one or more lymph nodes. The remaining 58 patients were histologically staged as N0 (Dukes A or B), and they comprise the population of the current study. All patients with colon cancer underwent an oncological resection, including the mesocolon of the vascular trunk. Patients with rectal carcinoma underwent total mesorectal excision (TME) after preoperative short-course radiotherapy comprising five fractions of five Gray. In addition, the mesenteric lymph nodes of 6 patients were included who underwent colonic resection for benign disease (e.g. Crohn’s disease or diverticulosis) as a negative control group. The study was done in accordance with the guidelines of the local ethics committee.

Sentinel lymph node mapping In this study, the ex vivo lymph mapping was used since only this technique can also be performed easily in patients with rectal carcinomas. The ex vivo technique has been demonstrated to be as accurate as the in vivo technique in identifying SNs27, 28, and it facilitates standardised, uniform specimen processing and assessment, recognising that the principal limitation of this technique is its inability to detect a small percentage (1-10%) of cases with aberrant lymphatic drainage.2, 16, 19 After standard resection, 0.5-2 ml (depending on the volume of the tumour) patent blue V (Guerbet, Gorinchem, The Netherlands) was injected around the tumour with the colonic specimen left intact.29 For colonic carcinomas, the mesocolon was inspected and the first one to four blue lymph nodes were identified as SNs and either dissected or marked with a suture. For rectal carcinomas the specimen was sent to the department of pathology immediately after resection in fresh state. The identification

60 61

3


detected OTC (c+e) x 100%. Negative SNs were considered false negative if one of the other regional lymph nodes (non-SNs) demonstrated micrometastases or ITC (e). The accuracy of the immunohistochemical SN analysis suggests a conformity of the SN status and the regional nodal status, i.e. the total number of patients with an immuno-histochemically positive SN (c) plus the number of patients with a true-negative SN (f )/ the number of patients with an identified SN (b) x 100%.

Figure 1 Representative examples of immunohistochemically detected occult tumour cells. A: micrometastasis > 0.2mm. B: isolated tumour cell, cluster of cells > 0.2mm. C: isolated tumour cells, cluster of cells < 0.2 mm with characteristic nuclear features. D: false-positively staining cell.

Figu

re 2

F

low

char

t. Pr

edic

tive

valu

e of

sen

tinel

nod

e m

app

ing

for o

ccul

t tu

mou

r cel

ls in

his

tolo

gica

lly n

egat

ive

lym

ph

nod

es.

Patie

nts

with

no

SN id

entifi

ed

(n=

5)

Imm

uno

nega

tive

SN

(n=

38)

(d

)

Neg

ativ

e no

n-SN

(n

=36

) (f)

Excl

usio

n of

N1

patie

nts

Stud

y po

pula

tion

(n=

100)

His

tolo

gica

lly N

0 pa

tient

s

(n=

58)

(a

)

Patie

nts

with

SN

iden

tified

(n

=53

)

(b)

Occ

ult t

umou

r cel

ls in

SN

(n=

15)

(c

)

Mic

rom

etas

tase

s (n

=7)

IT

C

(n=

8)

Posi

tive

non

-SN

(n

=2)

(e

)

ITC

(n=

2)

A B

C D

62 63

3


All statistical analyses were performed using the Statistical Software Package version 14.0 (SPSS INC., Chicago, IL, USA). Associations between clinicopathological features and the presence of micrometastases were analysed using Student’s t-test (continuous variables) or Chi-squared test (categorical variables). p-values of 0.05 or less were considered statistically significant.

Results

Restaging patients by detection of occult tumour cells Three serial sections of all 908 lymph nodes from 58 patients with N0 carcinomas were examined with three different antibodies (mean 15.7 lymph nodes per patient, range 6-42). In 27 patients none of the lymph nodes had metastases after immuno-histochemical staining and in 8 patients one or more lymph nodes demonstrated false positive staining unanimously recognised by both pathologists. There was consensus on 6 patients with micrometastases and nine patients with ITC; overall agreement 86% (50 of 58 patients). For the remaining eight patients there was discrepancy in the diagnosis by the two pathologists. In these eight patients, one pathologist diagnosed one micrometastasis and 10 ITC in 11 lymph nodes, whereas the other pathologist classified two lesions as micrometastases, six lesions as ITC and three lesions as false positive staining. It was noted that the discrepancy between micrometastases and ITC was mainly due to differences in measuring multiple foci. According to the AJCC 6th edition of TNM classification of malignant tumours, clusters of tumour cells are considered as one focus if the nests are separated by a few cells, and when in doubt the lower category should be selected to prevent upstaging. In a consensus meeting the following was agreed upon: 52 lymph nodes with OTC were found in 19 of the 58 patients (33%); seven patients had micrometastases (12% pN1mi+) and 12 patients had ITC (21% pN0itc+)(Table 1).

The presence of OTC (micrometastases and ITC combined) was not related to any clinicopathological characteristic (Table 2). Although micrometastases and ITC were found more frequently in tumours infiltrating the serosa and in tumours larger than 5cm, this was not statistically significant. Since micrometastases are thought to represent a different level of clinical significance, a correlation between pN1(mi+) versus pN0 (itc+/-) and clinicopathological characteristics was also analysed. The presence of micrometastases was also not related to tumour infiltration depth, differentiation grade, and extent of tumour spread, but a significant relation with lymphangio-invasion could be demonstrated (p<0.001).

Table 1 Classification of immunohistochemical staining results by two independent pathologists of 58 histologically negative patients with colorectal cancer.

Pathologist 1

Micrometastasis Isolated tumour cells

False positive staining

Absent

Pathologist 2 Micrometastasis 6 1 0 0

Isolated tumour cells

1 9 4 0

False positive staining

0 5 8 3

Absent 0 0 2 27

Consensus 7 12 10 29

Table 2 The presence of occult tumour cells correlated to tumour characteristics of 58 patients.

Occult tumour cells

Tumour characteristics n=58Micrometastasis

n=7ITC

n=12Absent n=39

p-value

Depth of invasiona T1 (4) 0 0 4 ns

T2 (18) 2 3 13

T3 (36) 5 9 22

Differentiation grade well (4) 0 0 4 ns

moderate (45) 7 10 28

poor (9) 0 2 7

Extend of tumour spread < 5cm (34) 2 6 26 ns

> 5cm (24) 5 6 13

Lymphangio-invasion absent (51) 3 12 36 0.001

present (7) 4 0 3

a T1: tumour limited to the (sub)mucosa, T2: tumour infiltrates muscularis propria, but not adventitia,

T3: tumour infiltrates adventitia. ITC = Isolated tumour cells

64 65

3


Tabl

e 3

P

redi

ctiv

e va

lue

of b

lue

dye

SN

map

pin

g in

pat

ient

s w

ith c

olor

ecta

l can

cer i

nclu

ding

imm

unoh

isto

chem

ical

ana

lyse

s (c

omp

aris

on o

f exi

stin

g lit

erat

ure

until

May

200

8; e

xclu

ding

dup

licat

e se

ries

and

stud

ies

with

up

date

d p

ublic

atio

ns).

Stud

yN

No

LNs

(mea

n)N

0(%

)Id

entifi

catio

n (%

)A

ccur

acy

(%)

Fals

e ne

g (%

)Se

nsiti

vity

(%)

Ups

tagi

ng (%

)IT

Ca

Ant

ibod

y

Imm

unoh

isto

chem

ical

ana

lysi

s of

SN

onl

y

Mat

ter1

5230

c63

92-

25d

44d

23N

CK1

1/C

EA/

Ca19

-9

Keld

er2

6911

7397

9630

d89

d, f

18N

CK7

/8/1

8

Bem

bene

k331

520

6985

8646

545/

16g

YM

NF1

6

Stoj

adin

ovic

416

218

6598

90d

31d

69d,

f11

NA

E1/A

E3

Cam

5.2

Van

Scha

ik5

4410

.552

9410

00

100

30N

LU-5

Bian

chi6

2222

c73

100

9517

8313

NC

K?

Thom

as7

6916

6293

7854

46f

5/-e,

gY

CEA

/ C

K

Smith

817

1641

e94

100

010

020

dN

MN

F116

Saha

950

015

50d

9896

1090

26N

AE1

/AE3

Codi

gnol

a1056

2159

100

89e,

f11

d86

e, f

38d

NCa

m5.

2

Braa

t1191

7.7e

62e

90e

90e

26e

74e,

f12

eN

CK7

/8

Cox12

1718

8210

010

00

100f

29N

AE1

/AE3

Smith

1340

1756

9895

e12

88f

29d

NCa

m5.

2

Bell14

58b

3057

9784

e16

d64

e4

NC

K

Para

mo15

5512

67e

8298

3d93

e11

dN

Cam

5.2

Fitz

gera

ld16

2615

87e

8891

d9d

33e

10N

Cam

5.2

Bend

avid

1720

-55

90-

--

25N

AE1

/AE3

Joos

ten18

5014

c48

7066

e60

4013

NLU

-5/

CK1

9

Bilc

hik19

141

1773

9991

3961

e8/

22g

YC

K

Bert

agno

lli20

7217

6692

8054

46-

NA

E1/A

E3 C

EA

Imm

unoh

isto

chem

ical

ana

lysi

s of

all

lym

ph n

odes

Faer

den21

199

13c

6893

8547

5330

NCa

m5.

2

Yagc

i2247

1957

9891

e20

8015

NA

E1/A

E3 /

CEA

Bem

bene

k2355

26c

7085

-4

9639

NM

NF

116

Turn

er24

5114

c55

e10

088

e26

e74

e29

NA

E1/A

E3

a D

iscr

imin

atio

n b

etw

een

mic

rom

etas

tasi

s an

d is

ola

ted

tum

our c

ells

(N=

no,

Y=

yes)

b Sp

ecim

ens

c M

edia

nd

Dis

crep

ancy

bet

wee

n re

por

ted

valu

e an

d ca

lcul

ated

val

ue (r

epor

ted

valu

es a

re m

enti

oned

)e C

alcu

late

d va

lue

f Imm

unoh

isto

chem

istr

y re

sult

s of

SN

incl

uded

in s

ensi

tivi

tyg %

mic

rom

etas

tase

s (t

rue

upst

agin

g)/

% is

ola

ted

tum

our c

ells

66 67

3


Discussion

Value of SN procedure in refining staging of colorectal patients Our study emphasises the potential value of SN procedure for optimising staging of patients with colorectal cancer. The examination of 908 lymph nodes by stepwise sections and immunohistochemical analysis revealed OTC in 33% of histologically N0 patients (12% micrometastases and 21% ITC). OTC were predominantly found in SNs with an overall sensitivity of SN mapping for OTC of 88%. A large number of studies already addressed the value of SN mapping for colorectal carcinomas (Table 3).1-24 In reviewing the literature, most studies demonstrate identification rates and accuracy rates varying from 80 to 100%, and conclude that this technique can predict lymph node status with good accuracy rates. However, in these studies the percentage of histologically N0 patients is often more than 50% which makes the presented accuracy rates difficult to interpret.1-7, 9-17, 19-23 This is merely because accuracy rates up to 90% may be accompanied by sensitivity rates and false negative rates of 50%.3, 7, 16, 19-21

For clinical use the SN procedure should be both highly sensitive and specific. Evaluating upstaging percentages in studies lacking immunohistochemically analysis of non-SNs harbours the danger of an underestimation of sensitivity rates and the presentation of flattering negative predictive values. The primary role of SN procedure in patients with colorectal cancer is to increase the accuracy of staging N0 patients by identifying and analysing those nodes with the highest likelihood to harbour metastatic disease in a cost-effective manner. Therefore, it is of great importance that OTC are predominantly found in SNs. We demonstrate that the incidence of OTC is more than four times higher in SNs compared to non-SNs. Our results are in agreement with two previous studies that showed SNs to have the highest probability of harbouring OTC.23, 24

Classification and detection of occult tumour cells However, our study also emphasises that the classification of OTC remains difficult. Reliable and reproducible discrimination between micrometastases (pN1) and ITC (pN0) is an essential component of adjuvant treatment planning. In analysing discrepant diagnoses for clinically relevant micrometastases between the two pathologists, we observed that the most challenging issue was whether multiple tumour clusters should be measured as one. A useful guideline in the AJCC system is that when controversy exists, the lower category should be selected. This recommendation was used during the consensus meeting. Another problem was the distinction between ITC and false positively staining (non-tumour) cells. Although this has no impact on treatment recommendations, the inter-observer variation highlights the need for well-defined antibodies.31 Our data suggest that Cam5.2 would be the

Sensitivity and specificity of three different antibodies No epithelial cells with atypical nuclear features characteristic of carcinoma cells were identified in lymph nodes from six control patients operated upon for benign disease, although occasional staining for keratin was observed in stellate cells located in the interfollicular regions of the nodes consistent with staining of reticulum cells. However, this pattern was easily distinguishable from the staining of carcinoma cells. Each micrometastasis was detected with all three antibodies, as were 39 of 45 ITC-positive lymph nodes. All ITC were detected with BerEP4. Six OTC were not detected with CK20 and two were missed with Cam5.2. It was noted that the intensity of staining with Cam5.2 and CK20 was related to differentiation of the primary tumour, and the undetected ITC were found in lymph nodes of patients with a dediffer-entiated tumour. In 18% of the patients false-positive staining was detected with BerEP4 (e.g.

reticulocytes, fibroblasts and mast cells) and this antibody yielded the most discordant classifications of ITC and false-positive stained cells between the two pathologists. False-positive staining was less frequently encountered with Cam5.2 (although there was occasional staining of dendritic cells) and was hardly seen with CK20 immuno-histochemical staining. These results would make Cam5.2 the antibody of first choice since it demonstrated the most consistent staining with high sensitivity and acceptable specificity.

Diagnostic value of SN mapping The SN (median two SNs, range 1-4) was found in 53 of the 58 N0 patients (identification rate 91%). Of the five patients in which the SN could not be identified, two patients demonstrated OTC in lymph nodes; these cells were all ITC . The remaining 53 patients were evaluated for the calculation of the diagnostic value of the SN. Overall, OTC were found in 17 patients; of which 15 were found in the SN and two in a non SN. Resulting in a sensitivity of 88% (15/17) and a negative predictive value of 94% (34/36)(Figure 2). OTC were found in 18 of the 105 SNs (17%), compared to 34 of 803 non-SNs (4%); this difference was statistically significant (p<0.001). Because all seven micrometastases were found in the SNs exclusively, the accuracy of the SN procedure in upstaging patients is 100%.

68 69

3


References



2. Kelder W, Braat AE, Karrenbeld A, Grond JA, De Vries JE, Oosterhuis JW, Baas PC, Plukker JT. The sentinel node

procedure in colon carcinoma: a multi-centre study in The Netherlands. Int J Colorectal Dis 2007; 22:1509-14.

3. Bembenek AE, Rosenberg R, Wagler E, et al. Sentinel lymph node biopsy in colon cancer: a prospective

multicenter trial. Ann Surg 2007; 245:858-63.




5. Van Schaik PM, Van der Linden JC, Ernst MF, et al. Ex vivo sentinel lymph node “mapping” in colorectal

cancer. Eur J Surg Oncol 2007;33:1177-82.


mapping and sentinel lymph node detection in colon cancer: technical aspects and preliminary results.

Surg Endosc 2007; 21:1567-71.

7. Thomas KA, Lechner J, Shen P, et al. Use of sentinel node mapping for cancer of the colon: ‘to map or not to

map’. Am Surg 2006;72:606-11.


cancer: improving the accuracy of pathologic staging? Am J Surg 2006;191:665-8.


prognostic implications for nodal staging and recurrence. Am J Surg 2006; 191:305-10.



Oncol 2005; 35:645-50.


short-course radiotherapy in rectal carcinoma is not reliable. Br J Surg 2005; 92:1533-8.


feasible and may improve staging in colorectal cancer. Curr Surg 2002; 59:301-6.

13. Smith FM, Coffey JC, Khasri NM, Walsh MF, Parfrey N, Gaffeny E, Stephens R, Kennedy MJ, Kirwan W,

Redmond HP. Sentinel nodes are identifiable in formalin-fixed specimens after surgeon-performed ex vivo

sentinel lymph node mapping in colorectal cancer. Ann Surg Oncol 2005; 12:504-9.


Colon Rectum 2005 ; 48 :74-9.



16. Fitzgerald TL, Khalifa MA, Al Zahrani M, et al. Ex vivo sentinel lymph node biopsy in colorectal cancer: a

feasibility study. J Surg Oncol 2002; 80:27-32.

17. Bendavid Y, Latulippe JF, Younan RJ, Leclerc YE, Dube S, Heyen F, Morin M, Girard R, et al. Phase I study on

sentinel lymph node mapping in colon cancer: a preliminary report. J Surg Oncol 2002; 7981-4.

18. Joosten JJ, Strobbe LJ, Wauters CA, Pruszczynski M, Wobbes T, Ruers TJ. Intraoperative lymphatic mapping

and the sentinel node concept in colorectal carcinoma. Br J Surg 1999; 86:482-6.


a prospective multicenter trial. Ann Surg 2007; 246:568-75.

20. Bertagnolli M, Miedema B, Redston M, et al. Sentinel node staging of resectable colon cancer: results of a

multicenter study. Ann Surg 2004; 240:624-8.


metastasis in colonic cancer. Dis Colon Rectum 2008;51:891-6.

antibody of first choice for the detection of OTC. The anticytokeratin-marker BerEP-4 is more sensitive but has an almost 20% false positive rate with high inter-observer variation. The use of the highly specific anti-CK20 with its expression limited to gastrointestinal epithelial cells cannot be recommended due to its loss of expression in dedifferentiated tumours, resulting in low specificity.

True upstaging rate of SN mapping and predictive value Comparing the incidence of clinically relevant upstaging by SN mapping with immunohistochemical analysis to literature is difficult. Before the revised TNM classification was introduced in 2003 micrometastases were not distinguished from ITC.31 Only three studies discriminated micrometastatic disease from ITC3, 7, 19 with low incidence of micrometastatic disease varying from 5 to 10%. Other studies reporting higher upstaging rates all included ITC in their findings.5, 10, 12, 13, 23, 24 This variation in definition of micrometastases makes it difficult to reach well-founded conclusions on the role of OTC and could explain the controversial prognostic significance of OTC. So far, only two studies have analysed the prognostic significance of micrometastatic disease in SNs with a decreased disease-specific survival in the pN1(mi+) group.10,11 All other studies commenting on prognostic significance of OTC did not perform SN mapping.33 Most studies failed to demonstrate a relation between the presence of OTC and tumour recurrence or survival. Messerini et al. demonstrated that the inclusion of ITC should probably be regarded as a confounding factor in these studies.33 Although our study was not designed to comment on the prognostic significance of OTC, the relation between lymphangio-invasion of tumour cells and micrometastases is suggestive of clinical relevance. Especially since this relation could not be demonstrated between lymphangio-invasion and ITC.

Conclusion

Occult tumour cells are predominantly found in the SN with a true upstaging rate of 12% in patients with pN0 colorectal cancer. For immunohistochemical staining Cam5.2 seems the antibody of choice with an overall agreement of pathologist of 88% in staging patients. The SN procedure has the potential to refine the staging system for patients with colorectal cancer. Further survival analyses should be awaited to reveal the real prognostic relevance of these findings.

70 71

3


22. Yagci G, Unlu A, Kurt B, Can MF, Kaymakcioglu N, Cetiner S, Tufan T, Sen D. Detection of micrometastases

and skip metastases with ex vivo sentinel node mapping in carcinoma of the colon and rectum. Int J

Colorectal Dis 2007; 22:167-73.

23. Bembenek A, Schneider U, Gretschel S, Fischer J, Schlag PM. Detection of lymph node micrometastases and

isolated tumor cells I sentinel and nonsentinel lymph nodes of colon cancer patients. World J Surg 2005;

29:1172-5.

24. Turner RR, Nora DT, Trocha SD, Bilchik AJ. Frequency and nature of cytokeratin-positive cells in sentinel and

nonsentinel lymph nodes. Arch Pathol Lab Med 2003; 127:673-9.

25. Green FL, Page DL, Fleming ID, et al. AJCC cancer staging manual. 6th ed. New York, Springer 2002.

26. Singletary SE, Allred C, Ashley P, et al. Revision of the American Joint Committee on Cancer staging system

for breast cancer. J Clin Oncol 2002; 20: 3628-36.

27. Teuch JJ, Pessaux P, Fiore FD, et al. Sentinel node mapping in coloncarcinoma : in-vivo versus ex-vivo

approach. Eur J Surg Oncol 2006 ; 32 :158-161.

28. Wood TF, Saha S, Morton DL, et al. Validation of lymphatic mapping in colorectal cancer: in vivo, ex vivo, and

laparoscopic techniques. Ann Surg Oncol 2001; 8:150-157.

29. van der Zaag ES, Buskens CJ, Kooij N, Akol H, Peters HM, Bouma WH, Bemelman WA. Improving staging


Oncol 2009; Mar 3.



Pathol 1998;185:427-434.

31. Redston M, Compton CC, Miedema BW, et al. Analysis of micrometastatic disease in sentinel lymph nodes

from resectable colon cancer: results of cancer and leukaemia group B trial 80001. J Clin Oncol 2006;

24:878-883.



4Categorization of occult tumour cells in lymph nodes in patients with colon cancer not reliable enough

We examined the interobserver variability among Dutch pathologists in classifying

occult tumour cells in lymph nodes in colon cancer.

75

4

categorization of occult tumour cells

Introduction

Lymph node status is the most important predictor for recurrence and survival in patients with colorectal cancer, and thus an essential part of the TNM-classification.1 After resection of colon carcinoma hematoxylin and eosin (H&E) staining of lymph nodes is used to detect lymphogenic metastases. However, 15-25% of patients with favourable early-stage I and II colon cancer (according to the American Joint Committee on Cancer (AJCC) absence of lymph node metastases) will have local recurrences or distant metastases within five years.2 This has prompted the use of sentinel lymph node (SN) mapping for colorectal cancer in an experimental setting. This procedure identifies lymph nodes most likely to harbour metastases and additional immunohistochemical analyses can be performed to detect occult tumour cells (OTC).3-5 Recently, OTC in lymph nodes were associated with a reduced 5-year rate of disease-free survival in women with breast cancer.6 Also, the use of the SN procedure in patients with breast cancer showed that 20% of the histologically negative patients were restaged.7 In the most recent TNM classification for breast cancer there is a distinction between macro metastases (>2mm, pN1), micrometastases (0.2-2mm, pN1mi) and isolated tumour cells (ITC)(<0.2mm, pN0).8-11

The International Union Against Cancer (UICC) TNM-classification uses both quantitative and qualitative criteria.11 According to this definition OTC are ITC without clinical consequence if they are smaller than 0.2mm, or show no sign of activity (no proliferation, no desmoplastic stromal-reaction) and are localised in the lymphatic sinus. On the other hand if ITC are detected at the parenchyma of the lymph node, they are considered micrometastases (pN1mi+). However, in the 6thedition of the American Joint Committee on Cancer (AJCC) TNM-classification only quantitative criteria are used7: clusters of tumour cells between 0.2mm and 2.0mm are micro-metastases pN1(mi+) and all lesions smaller than 0.2mm are defined as ITC pN0(itc+). Furthermore, in immunohistochemical techniques sometimes a false positive staining of hematopoietic stem cells arise that has no meaning in staging.12 If the occurrence of OTC in the SN demonstrated by immunohistochemical staining will be used for the administration of adjuvant chemotherapy in patients with colorectal cancer, it is crucial that these findings are consistent and reproducible. Therefore, we have examined the interobserver variability among Dutch pathologists in classifying OTC in lymph nodes in colon cancer. Characteristics that are associated with inconsistent diagnoses were identified in order to describe recommendations which could clarify the guidelines of the TNM classification.

76 77

4

chapter 4 categorization of occult tumour cells

Patients & methods

Patients The study population consisted of a single centre consecutive series of patients operated on for colorectal cancer with curative intent between November 2006 and July 2007. After conventional H&E staining 82 patients were histologically staged as pN0. All lymph nodes of these patients were immunohistochemically analysed for the occurrence of OTC.

Immunohistochemical analysis Serial sectioning was performed at 500μm intervals of all lymph nodes from formalin-fixed and paraffin-embedded archival tissue blocks. To decrease falls positive staining three different monoclonal antibodies were used: The anti-epithelial cell antibody Ber-EP4 (DAKO, The Netherlands) was combined with two anti-cytokeratin antibodies: the anti-CK20 antibody, with its expression limited to gastrointestinal epithelial cells (Euro Diagnostica, Arnhem, The Netherlands), and the anti-cytokeratin marker Cam5.2 directed against cytokeratin 7 and 8 expressed in all epithelial cells (Becton and Dickinson, Alphen aan den Rijn, The Netherlands). Details of the immuno-histochemical analysis are described elsewhere.13

Study design Of 30 immunohistochemically detected lesions representable pictures were made by a digital camera (Olympus) fixed on a conventional light microscope. One to three images with different magnifications and a scale for each photo were placed on a secured website (http://gelre.wockey.nl). Together with these pictures seven pictures were added of small volume lesions detected by H&E staining. Forty randomly selected pathologists were asked to judge the pictures and to categorise the lesions into ‘micrometastases’, ‘isolated tumour cells’ or ‘different’ and whether they were certain of their diagnosis. At each photograph they had the opportunity to comment. A selection of pictures to be judged is shown on Figure 1.

Definitions According to the latest TNM classification OTC found in lymph nodes are divided into clinically relevant micrometastases (pN1mi+) and ITC without clinical consequences (pN0itc+) .8-11 The pathologists had the option to use the UICC TNM-classification based on both quantitative and qualitative criteria, or the AJCC TNM-classification which uses quantitative criteria only.7,11

Figure 1 Examples of pictures placed at the website. A: Picture 1 (classified as lymph-angio invasion, isolated tumour cell or micrometastasis), B: Picture 6 (unanimously classified as micrometastasis), C: Picture 11 (mostly classified as isolated tumour cell, sometimes as micrometastasis based on localization and stroma cell reaction), D: Picture 15 (classified as isolated tumour cell or false-positive staining), E: Picture 21 (unanimously classified as isolated tumour cell), F: Picture 25 (classified as isolated tumour cell or contamination based on localization), G: Picture 33 and H: Picture 34 (classified as micrometastasis or isolated tumour cell, depending on whether or not lesions can be added up).

A B

C D

E F

G H

78 79

4


about their judgment, the amount of pictures with a good to perfect agreement increased to 18 out of 37 pictures (49%)(Table 2).

Tumour cell characteristics correlated with discrepant diagnoses Characteristics of the judged lesions of 26 pictures with a Kendall-W coefficient < 0.6 were analysed. In 16 lesions it was a relevant difference (ITC pN0 or artefact versus micrometastasis pN1). Predominantly this was associated with the occurrence of multifocal lesions in eight pictures, in which the pathologists indicated that it appeared unclear if these multifocal lesions should be added up. In three lesions the poor agreement was caused by the fact that some pathologists classified lesions less than 0.2 mm consciously as micrometastasis on the basis of tumour cell characteristics or localisation in lymphatic parenchyma. Two lesions with clinically relevant difference in diagnosis were approximately 0.2 mm and therefore discrepancy was probably based on differences of measurement. One lesion was scored by some pathologists as a micrometastasis while others classified these cells as contamination based on their position at the border of the lymph node. Two pictures with a poor agreement (Kendall-W 0.06) showed a tumour embolism at a blood vessel. This lesion was judged in three different manners; as a micrometastasis based on its localisation, as an isolated tumour cell because of its size and as a vascular invasion (no metastatic disease). The other ten lesions with a poor agreement were predominately different judgments in isolated tumour cells, artefacts or contamination.

Statistical analysis The interobserver agreement between pathologists judging OTC was calculated by the Kendall-W coefficient. This coefficient calculates the reproducibility of responses and risk-corrected level of agreement in the case of multiple evaluators.14 The Kendall-W coefficient of agreement has a range of 0,0 to1,0. A coefficient greater than 0,8 stands for (almost) excellent agreement between multiple observers (Table 1).15 All statistical analyses were performed using SPSS version 15.0 (SPSS INC., Chicago, IL, USA).

Results

Thirty-seven out of 40 pathologists took part in the survey (response rate of 93%). Two responses were excluded; one case because the query was incomplete, and the second case because the pathologist did not classify any lesion, arguing that in colon cancer the distinction between micrometastasis and ITC cannot be made. In total 35 responses could be used for analysis. All participants were certified pathologists, three pathologists were specialised in gastro-intestinal oncology. At the time of participation eight pathologists worked at a university hospital or oncological centre and 27 pathologists at a general hospital. In five cases (14%) there was complete agreement in classification of the lesion. Three lesions were categorised unanimously as micrometastases and two small clusters of tumour cells localised in the lymphatic sinus as isolated tumour cells. In 17 out of the 37 pictures (46%) the degree of agreement was poor to moderate, nine pictures (24%) showed a moderate degree of agreement and in 11 pictures (30%) a good to perfect agreement was found (Kendall-W coefficient > 0.61). When the analysis was performed only on diagnoses in which the pathologists were confident

Table 1 Kendall-W coefficient; risk-corrected level of agreement in the case of multiple evaluators.

Coefficient Level of agreement

0.00 - 0.20 Poor

0.21 - 0.40 Fair

0.41 - 0.60 Moderate

0.61 - 0.80 Good

0.81 - 1.0 Excellent

Table 2 Kendall-W degree of agreement of pathologists judging small volume lesions.

Kendall-W All pictures N

Pictures marked as certain*N

0.00 - 0.20 8 4

0.21 - 0.40 9 5

0.41 - 0.60 9 10

0.61 - 0.80 6 5

0.81 - 1.0 5 13

*Pictures marked by the pathologists as convinced about their diagnosis.

80 81

4


or ITC (n=10). In the first group the moderate agreement was related to the uncertainty in the TNM classification concerning adding up multifocal lesions. More specific guidelines could solve this problem. Another problem classifying small volume lesion is the difference in AJCC and UICC criteria. In the 6th edition of the AJCC TNM- classification only quantative criteria are used; all lesions smaller than 0.2 mm are defined as ITC pN0(itc+).7 Besides this criterion the UICC also uses malignant activity and localisation in the lymph node to describe OTC, lesions smaller than 0.2 mm can also be classified as clinically relevant micrometastasis.9 Some pathologists addressed that they deliberately used the latter classification because some lesion smaller than 0.2 mm were yet judged as clinically relevant. To come to a uniform judgement it is desirable to classify small lesions with the use of the AJCC criteria and to judge all lesions smaller than 0.2 mm as ITC. This is also consistent with the TNM classification; in doubt a lower category should be chosen.7 In patients with breast cancer a previous research showed that an uniform division between ITC and micrometastases can be made by establishing clear guidelines and training of pathologists.18 This does not apply to distinguishing false-positive staining cells from ITC. Although this difference is of little relevance, the poor agreement in ten pictures indicates that selective monoclonal antibodies are mandatory to prevent false positive staining.22

Conclusion

We concluded that categorisation of micrometastases and ITC in lymph nodes of patients with colon cancer is not unambiguous. In case occurrence of OTC will be used for the decision to give systemic chemotherapy in patients with colon cancer, better definitions and a more accurate classification for OTC are needed.

Comments on the pictures Although the pathologists confirmed that the survey consisted of clear pictures and a representative series of lesions, it appeared to be more difficult to judge these lesions on pictures compared to the use of a light microscope in daily practice. They addressed that this was mainly caused by the inability to see depth and they could not compare the lesion with the primary tumour. Off course this will influence the poor agreement shown in ten lesions for which they doubted between isolated tumour cells, contamination or artefact. However this will only partially explain the poor to moderate agreement in 16 pictures diagnosing micrometastasis or ITC. This is because for this distinction the main criteria are the localisation and the size of the lesion and whether or not multiple tumour clusters should be measured as one. These aspects can be well assessed on a picture with a reference line. Besides, the pathologists also stated in their comments that it is unclear which criteria should be used in daily practice.

Discussion

We show that OTC are difficult to classify reliably in lymph nodes of patients with colon cancer. In 70% of the 37 pictures the judgement of 35 pathologists showed a poor to moderate risk-corrected level of agreement (Kendall-W < 0.6). Two recent studies describing the reproducibility of the classification of OTC also showed disappointing results.17, 18 The pathologists who participated indicated that it was difficult to give a reliable diagnosis because they could not see depth at the pictures and a comparison with the original tumour was not possible either. This is partly an explanation for the moderate reliability. However, in contrast to daily practice, in our study investigators did not have to search for the lesions themselves which should enhance the reliability of immunohistochemically stained lesions in lymph nodes. Previously performed studies showed that the detection of small volume lesions is difficult and depends on concentration and experience of the pathologist.19,20 Besides, the pathologists had the opportunity to indicate at each picture whether they were confident about the diagnosis. If the impaired judgment of lesions at the pictures was the cause of the poor agreement, one would expect results to improve if only those lesions had been analysed in which the pathologists had indicated that they were convinced about the diagnosis. Although the percentage of pictures with a good agreement (Kendall-W coefficient 0.6-1.0) increased from 30% to 49%, the agreement in half of the judged pictures remained poor to moderate. Two groups within the 26 pictures of moderate agreement were distinguished during analysis of the results: clinically relevant disagreement in diagnosis ITC (pN0itc) versus micrometastasis (pN1mi) (n=16) and the discrimination of false-positive staining

82 83

4


References

1. Wiggers T, Arends JW, Schutte B, Volovics L, Bosman FT. A multivariate analysis of clinical and pathologic

prognostic indicators in large bowel cancer. Cancer 1988; 61:386-95


3. De Haas RJ, Wicherts DA, Hobbelink MG, Borel Rinkes IH, Schipper ME, Van der Zee JA, Van Hillegersberg R.

Sentinel lymph node mapping in colon cancer: current status. Ann Surg Oncol 2007; 14:1070-80.

4. Pickren JW. Significance of occult metastases. A study of breast cancer. Cancer 1961; 14:1266-71.

5. Messerini L, Cianchi F, Cortesini C, Comin CE. Incidence and prognostic significance of occult tumour cells

in lymph nodes from patients with stage IIA colorectal carcinoma. Hum Pathol 2006; 37:1259-67.

6. de Boer M, van Deurzen CH, van Dijck JA, et.al. Micrometastases or isolated tumor cells and the outcome of

breast cancer. N Engl J Med. 2009;361(7):653-63.

7. Singletary SE, Allred C, Ashley P, et al. Revision of the American Joint Committee on Cancer staging system

for breast cancer. J Clin Oncol 2002; 20: 3628-36.

8. Green FL, Page DL, Fleming ID, et al. AJCC cancer staging manual. 6th ed. New York, Springer 2002.

9. Singletary SE, Greene FL, Sobin LH. Classification of isolated tumour cells: Clarification of the 6th edition of

the American Joint Committee on Cancer Staging manual. Cancer 2003; 98:2740-1.

10. Sobin LH, Wittekind C. International Union Against Cancer classification of malignant tumours (ed 6). New

York, Wiley-Liss, 2002.

11. Hermanek P, Hutter RV, Sobin LH, Wittekind C. International Union Against Cancer. Classification of isolated

tumor cells and micrometastasis. Cancer 1999; 86:2668-73.



Pathol 1998;185:427-434.

13. Van der Zaag ES, Buskens CJ, Kooij N, Akol H, Peters HM, Bouma WH, Bemelman WA. Improving staging


Oncol 2009;

14. Siegel S Castellan N. Non parametric statistics for the behavioral sciences. New York: McGraw-Hill education;

1988.

15. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-

174.

16. Tanis PJ, Steller EP. Sentinel node metastasis in the case of colorectal cancer. No role for sentinel node

biopsy at present. Ned Tijdschr Geneeskd 2004; 148:1805-8

17. Cserni G, Bianchi S, Boecker W, Decker T, Lacerda M, Rank F, Wells CA. Improving the reproducibility of

diagnosing micrometastases and isolated tumour cells. Cancer 2005; 103:358-67.

18. Turner RR, Weaver DL, Cserni G, et al. Nodal stage classification for breast carcinoma: improving interobserver

reproducibility through standardized histologic criteria and image based training. J Clin Oncol 2008;

26:258-63.

19. Weaver DL, Krag DN, Manna EA, Ashikaga T, Harlow SP, Bauer KD. Comparison of pathologist-detected and

automated computer-assisted image analysis detected sentinel lymph node micrometastases in breast

cancer. Mod Pathol 2003; 16:1159-63.

20. Mesker W, Torrenga H, Sloos W, et al. Supervised automated microscopy increases sensitivity and efficiency

of detection of sentinel node micrometastases in breast cancer patients. J Clin Pathol 2004; 57:960-64.

21. De Mascarel I, MacGrogan G, Debled M, Brouste V, Mauriac L. Distinction between isolated tumor cells and

micrometastases in breast cancer. Cancer 2008; 112:1672-8.

22. Bleiweiss IJ, Nagi CS, Jaffer S. Axillary sentinel lymph nodes can be falsely positive due to iatrogenic

displacement and transport of benign epithelial cells in patients with breast carcinoma. J Clin Oncol 2006;

24:2013-8.

5Decreased incidence of isolated tumour cells in lymph nodes after laparoscopic resection for colorectal cancer

We assessed the effect of the surgical approach (i.e. open lateral to medial versus

laparoscopic no touch medial to lateral approach) on the levels of occult tumour cells

in sentinel lymph nodes of patients with stage I and II colorectal cancer.

87

5

less isolated tumour cells after laparoscopic resection

Introduction

During intra-operative manipulation of colorectal carcinomas, levels of circulating tumour cells (CTC) can be demonstrated in peripheral and portal blood.1 Recently, it has been demonstrated that the cumulative percentage of CTC in peripheral and portal blood was significantly lower after laparoscopic resection in which no-touch technique is used.2 This term refers to the principal of early lymphovascular ligation before manipulation of the tumour. In 1966, Turnbull et al. showed that early ligation increased survival rates.3 A randomised trial suggested favourable patient-free survival and overall survival rate however without significant difference.4 So far, the no-touch technique is not considered the standard surgical approach in open surgery. However in laparoscopic colorectal surgery the no-touch principles can optimally be represented with the preferable medial to lateral approach where the supplying vessels are ligated before the tumour is manipulated. If less CTC are detected in blood during laparoscopic surgery, it can be hypothesised that the same thing holds true for lymphatic flow with occult tumour cells (OTC) passing the lymphatic sinus. Based on the sentinel lymph node (SN) concept in which the lymphatic drainage from the primary tumour follows a specific order, we showed that OTC are preferentially found in the SN of patients with colorectal cancer.5 The objective of our study was to assess the effect of the surgical approach (i.e. open lateral to medial versus laparoscopic no touch medial to lateral approach) on the levels of OTC in sentinel lymph nodes of patients with stage I and II colorectal cancer.


Study population A prospective consecutive series of patients operated on between November 2006 and June 2009 was analysed. Only patients undergoing potentially curative resection for biopsy-proven colorectal cancer were eligible. Patients with solid organ metastases detectable by pre-operative radiological staging or intra-operative visualisation were excluded, as were patients with macrometastases on conventional H&E slides. In total, 107 patients with stage I and II colorectal cancer were included (Figure 1). In 87 patients the choice for open or laparoscopic approach depended of the surgeon to whom the patient was referred. Twenty patients also participated in a randomised trial comparing open and laparoscopic surgery (ISRCTN: 79588422).6 All procedures were performed by experienced colorectal surgeons performing more than 20 procedures each year. All laparoscopic procedures were performed by one surgeon expert in laparoscopic surgery. All these resections were performed

88 89

5

chapter 5 less isolated tumour cells after laparoscopic resection

Sentinel node procedure A SN procedure was performed in all patients. An ex vivo sentinel lymph node mapping was used, as described previously in detail.7 After resection 0.5-2 ml patent blue (depending on the volume of the tumour) was injected below the subserosa, around the tumour, with the colonic specimen left intact. The first one to four blue lymph nodes were identified as sentinel nodes and either dissected or marked with a suture.

Pathological examination The surgical resection specimens were analysed immediately at the department of pathology using a standardised protocol. Tumour stage and grading were classified according to the 6th edition of the AJCC TNM classification.8 All lymph nodes (SNs and non-SNs) were stained with hematoxylin and eosin and evaluated for tumour involvement. For histologically proven N0 patients, three serial sections at 500μm intervals were immunohistochemically stained with three different monoclonal antibodies to reveal OTC. The anti-epithelial cell antibody Ber-EP4, directed against membrane glycoproteins (DAKO, The Netherlands), was combined with two anticytokeratin antibodies: CK20 with its expression limited to gastrointestinal epithelial cells (Euro Diagnostica, Arnhem, The Netherlands) and Cam5.2, directed against cytokeratin 7 and 8 expressed in all epithelial cells (Becton and Dickinson, Alphen aan den Rijn, The Netherlands). Immunohistochemically detected cells with any of the three antibodies were considered as OTC only when they showed unequivocal morphological features of cancer cells. These tumour cells within lymph nodes were subdivided into two categories according to the AJCC revised guidelines: tumour cell deposits between 0.2mm and 2.0mm were referred to as micrometastases and those smaller than 0.2mm as isolated tumour cells (ITC).9

Statistical analysis Statistical calculations were performed using the Statistical Software Package version 14.0 (SPSS INC., Chicago, IL, USA). Categorical variables were compared using the chi-square test, while the t-test was used for continuous data with a normal distribution. Univariate logistic regression analysis was used to identify factors predictive for the detection of OTC in histologically proven N0 patients. Method of resection, T-stage, differentiation grade and tumour diameter were selected a priori as important cofactors in the identification of CTC. Since the limited number of events (from a statistical point of view) meant that only a restricted number of possible predictors could be included,10 variables with multiple categories were recoded into dichotomous variables by combining categories with comparable prognosis (T-stage I

according to the no-touch isolation technique (i.e. medial to lateral approach with early vessel ligation). A lateral to medial approach was performed during open resections. Patients who had early conversion e.g. because of dense adhesions were analysed in the open group if the vascular trunk was not ligated before mobilisation of the bowel and the lateral to medial approach was used. The study was done in accordance with the guidelines of the local ethics committee.

Figure 1 Patient stream by application of inclusion criteria.

All patients with colorectal cancer

N = 229

N = 107

N = 184

Patients with T4 tumour or distant metastases

N = 45

Patients with metastases in lymph nodes after H&E

N = 77

Open resection N = 59

Laparoscopic resection N = 48

Early conversion N = 3

Laparoscopic group N = 4 5

Open group N = 62

90 91

5


the SN was strongly related to lymphovascular invasion, irrespective of the surgical approach. However, this could not be demonstrated for the occurrence of ITC. The presence of these cells was only related to the method of resection with an odds ratio of 3.3 (1.1-9.6) for open lateral to medial resection (Table 3).

and II versus III and IV; differentiation grade: well and moderate versus poor). For tumour diameter, a cut-off of 35 mm was decided upon as suggested in a study of a specifically constructed receiver operating curve.11 All tests were performed two sided, and p < 0.05 was considered statistically significant. To determine the clinico-pathological characteristics predictive for the presence of OTC, logistic regression was performed.

Results

Characteristics of the study population During the study period, 107 patients with stage I and II colorectal cancer were included (Table 1). Sixty-two patients were analysed in the open group, whereas 45 patients were analysed in the laparoscopic group (Figure 1). There were 48 right-sided colectomies, 5 left-sided colectomies, 53 (recto) sigmoid resections and one subtotal colectomy. The patient and tumour characteristics were not different between the two patient groups, except for tumour size which was slightly larger in the open group (Table 1).

Immunohistochemical detection of occult tumour cells A total of 208 SNs were analysed. Ten patients had true micrometastases of more than 0.2 mm in one or more SNs (9.3%, pN1mi+), whereas ITC were found in 23 patients (21.5%, pN0itc+). The presence of micrometastases was equally distributed between the open and laparoscopic group (n=5 in both groups)(Table 2). In contrast, ITC were more frequently found in the conventional open resection group when compared to the no-touch laparoscopic group. Eighteen patients (29%) in the open group had ITC, compared to 5 patients (11%) in the laparoscopic group. This results in an absolute risk reduction for the incidence of ITC of 18% and a number needed to treat of 6 in favour of the laparoscopic group (Table 2). One of the five patients in the laparoscopic group with ITC, had a larger cluster of 0.18 mm with evidently signs of proliferation and stromal reaction, which could therefore be considered as a micrometastasis although we adhered to the quantitative character-istics of the AJCC. ITC were most frequently found in the sinuses of the lymph nodes as would be expected when these cells are considered as transiently shed cells with a limited life span.

Predictive factors for the presence of occult tumour cells A relation between the presence of OTC and larger tumours (both infiltration depth and tumour diameter) was found. The occurrence of true micrometastases in

Table 1 Clinicopathological characteristics of the included patients.

All Patients(n=107)

Open group (n= 62)

Laparoscopic group (n= 45)

p-value

Age (yrs)a 70±11 70±10 69±12 0.8

Gender Maleb 48 28 20 0.9

Female b 59 34 25

Body Mass Index c 26 (18-31) 24 (18-28) 26 (22-31) 0.7

Preoperative CEA level c 2.2(0.5-10) 2.4(0.9-10) 2.2(0.5-8.4) 0.2

Tumour location b 0.2

colon 87 53 34

rectum 20 9 11

Tumour diameter (cm) a 5.1±2 4.2±2.0 0.04

Number of resected lymph nodes a 15.5±7 15.5±7 15.6±7 1.0

Number of identified SN a 2.0±2 1.9±2.1 2.1±2.4 0.7

Depth of invasionb 0.7

pT1 8 3 5

pT2 29 17 12

pT3 68 41 27

pT4 2 1 1

Differentiation gradeb 0.2

well 9 5 4

moderate 80 43 37

poor 18 14 4

Lymphovascular invasionb 0.1

absent 97 54 43

present 10 8 2

a Values are mean (SD)b Values are absolute numbersc Values are median (min-max)

92 93

5


Discussion

Our study demonstrates that the incidence of ITC in lymph nodes of patients with stage I and II colorectal cancer is lower after no-touch laparoscopic resection when compared to lateral to medial open resection. These results are in line with other studies demonstrating reduced levels of CTC in peripheral and portal blood during laparoscopic resection.2, 12 This is the first study analysing tumour cells in lymph nodes using immunohistochemical techniques, hereby overcoming the historically controversial issues regarding the reliability and reproducibility in measuring CTC in blood.13 An advantage of analysing OTC in lymph nodes is that histology can be preserved, discriminating between OTC and falsely positive stained cells. Most studies sampling peripheral or portal blood for CTC during surgery use reverse transcriptase-poly-merase chain reaction (RT-PCR) or the quantitative real-time RT-PCR. However, histological confirmation is lost in these techniques, which harbours the danger of identifying shed debris or haematopoietic cells as tumour cells. The clinical relevance of our finding is hard to assess. Although the prognostic role of OTC seems to be established in metastasised breast, colon and prostate cancer patients14-16, the clinical relevance of per-operatively detected cells is still debated. So far, these cells are generally considered to be transiently shed cells without prognostic significance. The metastatic process is an extremely inefficient process which involves survival in circulation or lymphatics, arresting at a distant target organ, extravasation into the surrounding tissue and survival in the foreign microenvironment, followed by proliferation and induction of angiogenesis while evading apoptotic death or immunological response.17 Tumours can shed millions of cells into the bloodstream daily, but it has been demonstrated that only a small percentage of tumour cells (0.05%) can survive and initiate a metastatic focus.17

In contrast, data on patients with breast cancer show that OTC in sentinel lymph nodes are related to a reduced 5-year disease free survival.18 Cultured single cells from rib marrow of patients with oesophagogastric cancer were viable when inoculated subcutaneously in athymic nude mice.19 In addition, in patients with colon cancer improved five-year survival rates was reported combined with a reduced frequency of cancer cells in portal blood using the open no-touch isolation technique.3 The only randomised trial on no-touch technique in patients with colon cancer did not show a significant increase in disease-free survival, but data demonstrated better results in all analyses for the no-touch patient group (time to recurrence, incidence of liver metastases and disease free survival).4 So far, several randomised trials comparing laparoscopic to open surgery with long-term follow-up did not show survival differences.20, 21 This would be in line with our hypothesis that the decreased number of detected ITC is merely a mechanic

Table 2 Presence of occult tumour cells in the SN in patients with colorectal cancer; comparison between open and laparoscopic resection.

All patientsn= 107

Open groupn=62

Laparoscopic groupn=45

p valueDifference (%)

Occult tumour cells 33 23 10

Micrometastases 10 5 5 0.6

0.2-2.0 mm 8% 11% -3% (-15-8%)

Isolated tumour cells

23 18 5 0.03

< 0.2 mm 29% 11% 18% (3-33%)

Table 3 Logistic regression analyses of predictors for the presence of OTC, micrometastases (0.2-2.0mm) and ITC (<0.2mm) in histologically N0 patients with colorectal cancer.

Occult tumour cells Isolated tumour cells

Variable OR a (CIb) p-value OR a (CIb) p-value

Method of resection open vs laparoscopic 0.5 (0.20-1.16) 0.1 3.3 (1.11-9.63) 0.03

Depth of invasion pT3/4 vs pT1/2 2.5 (0.99-6.6) 0.06 2.2 (0.75-6.6) 0.2

Differentiation grade poor vs well and moderate 2.1 (0.75-5.78) 0.2 2.1 (0.70-6.45) 0.2

Tumour diameter >3.5cm vs <3.5cm 2.7 (1.05-7.08) 0.04 1.7 (0.61-4.86) 0.3

Micrometastases Isolated tumour cells

Lymphovascular invasion

present vs absent 18.4 (4.0-85.1) <0.001 0.38 (0.05-3.16) 0.4

a Odds ratiob 95% Confidence interval

94 95

5


References

1. Yamaguchi K, Takagi Y, Aoki S, Futamura M, Saji S. Significant detection of circulating cancer cells in the blood by

reverse transcriptase-polymerase chain reaction during colorectal cancer resection. Ann Surg 2000; 232:58-65.

2. Wind J, Tuynman JB, Tibbe AG, et al. Circulating tumour cells during laparoscopic and open surgery for primary

colonic cancer in portal and peripheral blood. Eur J Surg Oncol 2009; 35:942-50

3. Turnbull RB, Kenneth K, Watson FR, Spratt J. Cancer of the colon: the influence of the no-touch isolation technic

on survival rates. Ann Surg 1967; 166:420-7.

4. Wiggers T, Jeekel J, Arends JW, et al. No-touch isolation technique in colon cancer: a controlled prospective trial.

Br J Surg 1988; 75:409-15.

5. Van der Zaag ES, Kooij N, Peters HM, van de Vijver M, Bemelman WA, Buskens CJ. Diagnosing occult tumour cells

and their predictive value in sentinel nodes of histologically negative patients with colorectal cancer. Eur J Surg

Oncol 2010;36:350-7.

6. Wind J, Hofland J, Preckel B, et al. Perioperative strategy in colonic surgery; LAparoscopy and/or FAst track

multimodal management versus standard care (LAFA trial), BMC Surg 6 (2006), p. 16.

7. Van der Zaag ES, Buskens CJ, Kooij N, Akol H, Peters HM, Bouma WH, Bemelman WA. Improving staging accuracy

in colon and rectal cancer by sentinel lymph node mapping: a comparative study. Eur J Surg Oncol 2009;

35:1065-70.

8. TNM classification of malignant tumours. 6th ed. New-York: Wiley-Liss; 2002.



10. Harrell FE, Jr., Lee KL, Mark DB: Multivariable prognostic models: issues in developing models, evaluating

assumptions and adequacy, and measuring and reducing errors. Stat Med 1996;15:361-387.

11. Cahill RA, Bembenek A, Sirop S, et al. Sentinel node biopsy for the individualization of surgical strategy for cure

of early-stage colon cancer. Ann Surg Oncol 2009; 16:2170-80.

12. Bessa X, Castells A, Lacy AM, et al. Laparoscopic assisted vs. open colectomy for colorectal cancer: influence on

neoplastic cell mobilization. J Gastrointest Surg 2001; 5:66-73.

13. Ring AE, Zabaglo L, Ormerod MG, Smith IE, Dowsett M. Detection of circulating epithelial cells in the blood of

patients with breast cancer: Comparison of three techniques. Br J Cancer 2005; 92:906-12.

14. Cristofanilli M, Budd GT, Ellis MJ, et al. Circulating tumour cells, disease progression, and survival in metastatic

breast cancer. N Engl J Med 2004; 351:781-91.

15. Cohen SJ, Punt CJ, Iannotti N, et al. Prognostic significance of circulating tumour cells in patients with metastatic

colorectal cancer. Ann Oncol. 2009;20:1223-9

16. Moreno JG, Miller MC, Gross S, Allard WJ, Gomella LG, Terstappen LW. Circulating tumour cells predict survival in

patients with metastatic prostate cancer. Urology 2005; 65:713-8.

17. Hunter KW. Host genetics and tumour metastasis. Br J Cancer 2004; 90:752-55.

18. de Boer M, van Deurzen CH, van Dijck JA, et al. Micrometastases or isolated tumour cells and the outcome of

breast cancer. N Engl J Med. 2009;361:653-63.

19. O’sullivan GC, Sheehan D, Clarke A, et al. Micrometastases in esophagogastric cancer: High detection rate in

resected rib segments. Gastroenterology 1999;116:543-548.

20. Buunen M, Veldkamp R, Hop WC, et al. Survival after laparoscopic surgery versus open surgery for colon

cancer:long-term outcome of a randomised clinical trial. Lancet Oncol. 2009;10(1):44-52.

21. Jayne DG, Thorpe HC, Copeland J, et al. Five-year follow-up of the Medical Research Council CLASICC trial of

laparoscopically assisted versus open surgery for colorectal cancer. Br J Surg. 2010;97(11):1638-45.

22. Lacy AM, García-Valdecasas JC, Delgado S, Castells A, Taurá P, Piqué JM, Visa J. Laparoscopy-assisted colectomy

versus open colectomy for treatment of non-metastatic colon cancer: a randomised trial. Lancet. 2002;359:2224-9.

23. Weitz J, Kienle P, Lacroix J, et al. Dissemination of tumour cells in patients undergoing surgery for colorectal

cancer. Clin Cancer Res 1998; 4:343-8.

24. Lundy J. Anesthesia and surgery: a double-edge sword for the cancer patient. J Surg Oncol 1980; 14:61-5.

process depending on the surgical approach, rather than being viewed upon as a prognostic factor. A better disease free survival was demonstrated in one early trial comparing laparoscopic to open resection, but this improvement was only seen in stage III patients and therefore these results cannot be compared to our findings in histologically N0 patients.22

Previously, it has been suggested that the decreased detection rate of CTC could be attributed to early ligation of the lymphovascular trunk (medial to lateral approach).2 We have demonstrated in a previous study that OTC are preferentially found in peri-tumoral SNs.5 Since lymphatic drainage to peri-tumoural lymph nodes is not disturbed by early ligation, it is unlikely that solely a medial to lateral approach would result in decreased numbers of ITC. Therefore, our results suggest that predominantly the no-touch technique of the laparoscopic approach explains the lower incidence of ITC when compared to the conventional open resection. Even without demonstrated prognostic relevance, this would make laparoscopic no touch resection the treatment of choice for patients with stage I and II colorectal cancer. The development of all metastases will start with a single cell, and so far there are no techniques available to predict the possible survival of ITC. Although most cells will not survive circulation and immunological responses, activation of blood coagulation and relative immune suppression due to surgical stress might enhance the metastatic potential of these intra-operatively spilled cells.23,24 A major drawback of our study is the non-randomised design. However, the difference in ITC is not likely to be explained by patient selection since patient- and tumour characteristics, CEA levels and the number of resected lymph nodes were comparable in both treatment groups. There was a difference in tumour diameter with smaller tumours in the laparoscopic group. However, regression analysis showed no relation between the presence of ITC and tumour diameter. The only predictive parameter for the presence of ITC was the method of resection with an OR of 3.3 for open lateral to medial surgery.

Conclusion

Laparoscopic no touch surgery results in less isolated tumour cells in lymph nodes compared to open lateral to medial surgery in patients with stage I and II colorectal cancer. Although prospective survival analyses in these patients with ITC and randomised trials should be awaited, every attempt should be made to prevent worsening of the prognosis of patients during surgery. In that respect laparoscopic resection with the no touch technique may have benefits over open surgery in patients with stage I and II colorectal cancer.

6Implications of sentinel lymph node mapping on nodal staging and prognosis in colorectal cancer

We prospectively assessed the effect of sentinel lymph node mapping on nodal

staging and its prognostic impact in terms of disease recurrence in patients with

colorectal cancer.

99

6

implication of sentinel lymph node mapping on prognosis

Introduction

The presence and extent of regional nodal metastasis is pivotal in predicting the clinical outcome of patients with colorectal cancer (CRC). For the diagnosis, treatment and survival of these patients, the completeness of nodal resection and accuracy of staging is crucial. Variations in disease free survival among patients with node negative early stage disease may reflect inadequate nodal resection or inaccuracies of pathologic staging. The number of harvested lymph nodes has been demonstrated to be an independent predictor of clinical outcome.1, 2 Therefore, guidelines recommend that a minimum of 12 lymph nodes is needed for adequate staging when conventional techniques are used.3 Still, in patients with node-negative disease staged by current standard pathologic techniques, up to 30% experience a disease recurrence.4

To improve staging accuracy, sentinel node (SN) mapping provides a potential means of improving the ability to detect nodal metastasis. In contrast to melanoma and breast cancer patients, in whom SN mapping has been proved to be highly effective in correctly assessing nodal status, the current evidence for SN mapping in CRC is conflictive.5, 6 Some studies show a high predictive value of SN mapping for nodal status in colorectal cancer.7, 8 These authors hypothesise that the detection of small tumour deposits in lymph nodes may improve staging accuracy.8 Also, the SN procedure results in an increased yield of harvested lymph nodes.9 On the other hand, another study reports prohibitively high false negative rates.10 Furthermore, studies demonstrating the benefit of adjuvant therapy in patients with micrometastatic-only node positive disease are lacking. The principle aim of our study was therefore to prospectively assess the effect of SN mapping on nodal staging and its prognostic impact in terms of disease recurrence in patients with CRC.


Study population Between November 2005 and July 2009, a total of 424 consecutive patients were operated on for CRC. Fifty patients undergoing emergent laparotomy for obstructing or perforated tumour were excluded, as well as 38 patients who had concurrent distant metastases. Another five patients were excluded because they did not have an invasive carcinoma in their resection specimen. The 331 patients included in our study all underwent standard surgical resection of the tumour-bearing colon with en-bloc regional lymphadenectomy either via midline laparotomy (n=214) or laparoscopic approach (n=117). All patients were operated on by experienced gastro-

100 101

6

chapter 6 implication of sentinel lymph node mapping on prognosis

intestinal surgeons, all using standard surgical principles during resection as mentioned above. In 189 patients a SN procedure was performed (group A). The decision to perform SN mapping was depending on the availability of the principal investigator with adequate experience in SN mapping (more than 20 procedures per year). The remaining patients without a SN procedure (n=142) form the control group (group B). (Figure 1) Both patient groups were operated on by experienced gastro-intestinal surgeons. The study was done in accordance with the guidelines of the local ethics committee.

Sentinel node procedure The ex vivo technique was used for SN mapping since this technique is also suitable for rectal cancer and has been associated with more consistent results than the in vivo procedure.11 We previously described the method we used in detail.12 After resection, 0.5-2 ml patent blue (depending on the volume of the tumour) was injected subserosally around the tumour with the specimen left intact. The first blue lymph nodes were identified as SNs and either dissected or marked with a suture.

Pathological examination All surgical resection specimens were analysed immediately at the department of pathology using a standardised protocol. All lymph nodes (SNs and non-SNs) were stained with hematoxylin and eosin (H&E) and evaluated for tumour involvement. Tumour stage and grading were classified according to the sixth edition of the UICC TNM classification.14

In group A, serial sectioning of all SNs was performed at 500mm intervals and three serial sections per level were stained with Ber-EP4 (DAKO, The Netherlands), CK20 (Euro Diagnostica, Arnhem, The Netherlands), and Cam5.2 (Becton and Dickinson, Alphen aan den Rijn, The Netherlands) to reveal occult tumour cells (OTC). Non-SNs were not analysed immunohistochemically since we previously demonstrated that OTC are predominantly found in the SN.15

Immunohistochemically detected single cells or clusters of cells were considered as OTC only if they showed unequivocal morphological features of cancer cells. The OTC within lymph nodes were subdivided into two categories according to the AJCC revised guidelines: tumour cell deposits between 0.2 mm and 2.0 mm were referred to as micrometastases and considered as true upstaging (pN1mi+), whereas those smaller than 0.2 mm were classified as isolated tumour cells (ITC).16 Since ITC have not been demonstrated to be of prognostic significance these lymph nodes were still considered N0 (pN0itc).

Figure 1 Patient stream by application of inclusion criteria.

All patients

with resection for colorectal cancer

N = 424

N = 331

N = 374

Patients with emergeny laparotomy

N = 50

Patients with distant metastases

N = 38

Group A

Patients with SN procedure

N = 189

Group B

Patients without SN procedure

N = 142

Patients with benign disease

N = 5

N = 336

Group A

Survival analyses

N = 184

Group B

Survival analyses

N = 142

Lost to follow-up

N = 5

102 103

6


Outcome parameters Patients with nodal positive colonic cancer on conventional H&E staining were offered adjuvant chemotherapy. Additional immunohistochemistry results in N0 patients were regarded as experimental, therefore the upstaged patients with micro-metastasis (pN1mi+) did not receive adjuvant treatment. All patients were followed in the outpatient clinics of the department of surgery or at the oncology unit. Follow-up included regular clinical examination every three months with CEA check, ultrasound of the liver, chest X-ray at 6 month interval followed by thoraco-abdominal CT-scan in case a lesion was detected. Colonoscopy was performed after 12 months. The development of new local recurrent or metastatic disease after the operation was defined as post-operative disease recurrence. Five patients in the SN group were lost to follow-up. Statistical analysis Statistical calculations were performed using SPSS version 15.0 (SPSS INC., Chicago, IL, USA). Categorical variables were compared using the Chi-squared test, while the Student’s t-test was used for continuous data with a normal distribution. The cumulative recurrence rates were calculated by the Kaplan-Meier method, and the differences in recurrence rates were analysed by the log-rank test. A univariate cox regression analysis was performed to identify candidate predictors of disease recurrence (age, gender, left or right sided colectomy, T stage, rectal cancer, complete resection and SN procedure). For every ten events one predictor could be entered into the multivariate analysis. Subsequently, factors that were most strongly associated with disease recurrence were analysed in a multivariate model. All tests were performed two sided, and a p value ≤0.05 was considered statistically significant.

Results

Clinicopathological characteristics of the study population The clinical and pathological data regarding the 331 patients are summarised in Table 1. We included 174 men (53%). Two hundred and fifteen patients (65%) had colonic cancer whereas 116 had rectal cancer. The patient group in which a SN procedure was performed had a higher amount of patients with colonic cancer and also right sided colectomy was performed more frequently. With regard to histo-pathologic results, there was an equal distribution of tumour classification, differen-tiation grade and completeness of resection between the SN group and control group.

Table 1 Clinicopathologic characteristics of 331 patients operated on for colorectal cancer.

Patient characteristics All Patients(n=331)

SN mapping(n=189)

No.

Conventional (n=142)

No.

p-value

Age (yrs, mean ± SD) 68.7±12 69.6±12 68.6±12 0.5

Sex ratio (M:F) 174:157 94:95 80 0.2

Method of resection open 214 118 96 0.3

laparoscopic 117 71 46

Type of resection

right sided 110 76 34 0.002

left sided 206 107 99

transverse 5 1 4

subtotal colectomy 10 8 2

Tumour characteristics

Tumour location colon 215 137 78 0.001

rectum 116 52 64

Dukes classificationa A 69 35 34 0.4

B 133 75 58

C 129 79 50

Complete resectionb R0 316 184 132 0.1

R1 13 5 8

R2 2 0 2

Differentiation grade well 18 9 9 0.6

moderate 227 134 93

poor 86 46 40

No of resected LNc overall 14.0±6.7 15.5±7.3 12.1±5.2 <0.0001

a A = T1-2, N0, M0; B = T3-4, N0, M0; C = T1-4, N1-2.b R0= complete resection; R1= microscopically incomplete resection,

R2= macroscopically incomplete resectionc LN= lymph nodes

104 105

6


Prognostic outcome In group A 184 patients and all 142 patients in group B were followed for a minimum of 12 months with a mean of 27.9 ± 14.0 months. The post-operative recurrence rate was 20%, which was not significantly different between the groups (18.5% in group A versus 21.1% in group B, p=0.3, log-rank test). (Figure 2) When the recurrence rate was analysed for 192 N0 patients, significantly less recurrences were seen in N0 patients after immunohistochemical analysis of the SN (4% in group A versus 15% in group B, p=0.04, log rank test). This resulted in a significantly longer median disease free survival in these patients of 55 months (53-57 months) compared to 48 months (44-52 months) for N0 patients without SN mapping. (Figure 3) Univariate analyses for N0 patients showed that the SN procedure was the strongest predictor for disease free survival. Multivariate analyses showed that N0 patients without a SN procedure had a higher risk for local recurrence compared to patients in whom a SN was performed with a hazard ratio of 4.1 (1.3-12.6). (Table 4)

A mean number of 14.0 lymph nodes (range 2-47) was examined per tumour specimen. This was significantly different between the two groups: 15.5 ± 7.3 lymph nodes in group A versus 12.1 ± 5.2 in group B (p<0.0001). In 137 (73%) of the patients in group A, the minimally required 12 nodes were resected, whereas this occurred in only 87 (61%) of patients in group B (p=0.007). Both in patients with colonic and rectal cancer, and patients with a right sided or left sided colectomy, significantly more lymph nodes were found after a SN procedure (Table 2).

Nodal staging after SN mapping The ex vivo mapping procedure identified at least one SN in 167 of the 189 patients (88%) with an average of 2.0 ± 1.4 SNs per patient. Overall, the SN correctly predicted lymph node status in 147 of the 167 patients (accuracy 88%), but in 20 of the 70 patients with N1 disease, the SN was false negative (29%). In the SN group, 81 patients (43%) were lymph node positive by conventional histopathology which was significantly higher than the 50 patients (35%) N1 patients in group B (p=0.03) (Table 3). After additional immunohistochemistry eight patients had micrometastases in there SN and were upstaged to pN1mi+ and thereby the percentage N1 patients increased to 47%. An additional 23 N0 patients in group A were found to have ITC in their SNs, and they were staged as pN0itc. Lymph node metastases were predominately seen in tumours invading through the muscularis propria. All patients who were upstaged by sentinel node mapping had T2 or T3 tumours (Table 3).

Table 2 The number of resected lymph nodes of 331 patients operated on for colorectal cancer.

Overall SN procedure Conventional p-value

(n=331) (n=189) (n=142)

Tumour colon 15.0±6.9 16.1±7.4 13.1±5.2 0.001

rectum 12.2±6.1 13.8±6.9 10.8±5.0 0.009

Type of resection right-sided 16.3±7.6 17.4±8.5 14.1±4.5 0.04

left-sided 12.8±5.8 13.8±5.8 11.3±5.2 0.001

T stage T1 and T2 12.6±6.1 13.7±6.3 11.5±5.8 0.06

T3 and T4 14.5±6.7 16.0±7.3 12.7±5.6 0.0001

Data presented as median (SD)

Table 3 Comparative analysis of nodal staging in CRC: SN mapping with additional immunohistochemical analysis versus conventional pathologic staging.

T-stageSN mapping

(no of pts)SN mapping

(%N1)SN + IHC†

(%N1)Conventional

(no of pts)Conventional

(%N1)

T1 10 1 (10) 1 (10) 10 1 (10)

T2 42 11 (26) 13 (31) 40 14 (35)

T3 124 60 (48) 66 (53) 77 30 (39)

T4 13 9 (69) 9 (69) 15 5 (33)

Total* 189 81 (43) 89 (47) 142 50 (35)

†IHC = immunohistochemistry. *p=0.002

106 107

6


Figure 3 Kaplan-Meier curves of 192 N0 patients with CRC. Patients with or without SN mapping.

Number op patients at risk

SN – 92 73 63 53 43 33 23

SN + 100 87 77 67 57 47 37

Figure 2 Kaplan-Meier curves of 326 patients with colorectal cancer. Patients with or without SN mapping.

Number op patients at risk

SN – 142 132 122 112 102 92 82

SN + 184 171 161 151 141 131 121

108 109

6


colectomies, a higher number of lymph nodes was found after the SN procedure. The relation between the number of lymph nodes analysed and nodal status has been demonstrated previously. When the number of analysed nodes per resection specimen increased, the percentage node positive patients rose from 26% to 33%.18 Different cut off values have been reported. For correctly classifying a colorectal tumour as node positive, 17 lymph nodes need to be examined, and high risk groups for recurrence are identified in Dukes A and B patients if less than 14 lymph nodes are harvested. 19 Recently, an optimal prognostic cut off value for the number of lymph nodes examined after curative resection was determined at 18.20 The reported five-year disease free survival of N0 patients was 97% with 18 or more analysed lymph nodes versus 68% with less than 18 lymph nodes. 20 Although we analysed on average 14 lymph nodes, and the time of follow-up was less than five years, we saw comparable differences in recurrence rates of N0 patients between the SN group and conventional staging group (4% versus 15%). When immunohistochemical analysis of the SN was added to conventional analysis of the SN, 8 N0 patients (7.4%) were upstaged to N1. A recent study on im-munohistochemical analysis of the SN found a higher increase in N1 patients when compared to conventional analysis (from 39% to 57%).8 However, in that report all encountered tumour cells, both ITC and micrometastases, were included in upstaging the nodal status. We strictly adhered to the definition of upstaging as recommended by the AJCC staging manual (pN1mi+) which states that ITC, in contrast to microme-tastases, do not seem to have prognostic significance.21 Reports with the same definitions used for micrometastases described comparable upstaging percentage of 5 to 8% pN1mi+.24,25 Although 7.4% is a relatively low percentage for upstaging, identification of these otherwise understaged patients is particularly important since patients with micrometastases are suspected to have the same prognosis as pN1 patients and may in part represent a subgroup of the pN0 patients that experience disease recurrence after a curative resection.21-25 In our study eight patients revealed micrometastases (pN1mi+) and 23 patients had ITC (pN0itc+) in their SNs. Disease recurrence occurred only once per group. Because of this low incidence in our study it is difficult to assess survival outcome of these patients in more detail. There is much debate about the clinical value of SN mapping in CRC due to the high false-negative rates reported in several multicentre studies.9, 26 A systematic review of 33 studies reported an overall identification rate of >90 % (range 66-100%) and a false negative rate of 26% (range 0-82%).27 Although these results seem acceptable, the fact that 13 of the 33 SN studies report a false negative rate of over 33% is worrisome. We performed an ex vivo SN mapping with an identification rate of 88% and a sensitivity of 71% which demonstrates that even after four years of experience the number of false negative procedures is still considerable. However, it has to be emphasised that SN mapping in colorectal patients is not intended to alter

Discussion

Performing a SN procedure with additional immunohistochemical testing significantly increases the percentage of node positive patients when compared to conventional pathological staging (47% versus 35% respectively), with a corresponding increase in recurrence free survival of N0 patients in the SN group. Since none of the N0 patients received adjuvant therapy, this finding suggests a better prognostication and more accurate nodal staging with SN mapping. We describe a non-randomised comparison of both patient groups. Patient and tumour characteristics were comparable between groups apart from the number of patients with rectal cancer and the amount of right sided colectomies. Nevertheless, multivariate analyses showed that a SN procedure remained the strongest predictor for disease recurrence. Interestingly, the increase in N1 patients was only partly due to the upstaging of patients with additional immunohistochemical staining. Even with conventional H&E staining, there was an almost 8% difference in N1 patients between the two groups. This is probably related to the increased number of harvested lymph nodes after SN mapping. Since the blue SNs are visible during subsequent pathological examination of the specimen, lymph node harvest might be facilitated potentially leading to a higher LN yield.17 In both colonic and rectal cancer, and right and left sided

Table 4 Uni- and multivariate analyses for disease recurrence in all 192 N0 patients.

Univariate Hazard ratio (Confidence interval) p value

Age 1.0 (0.9-1.0) 0.8

Gender 0.9 (0.3-2.0) 0.8

Right or left sided 0.5 (0.6-1.5) 0.2

T stage 2.4 (1.2-5.1) 0.013

Rectal cancer 2.8 (1.1-7.1) 0.03

Complete resection 3.2 (1.3-7.9) 0.009

SN procedure 4.5 (1.5-14) 0.008

Multivariate

Complete resection 2.4 (1.0-6.0) 0.06

SN procedure 4.1 (1.3-12.6) 0.015

110 111

6


References

1. Tsai HL, Lu CY, Hsieh JS, et al. The prognostic significance of total lymph node harvest in patients with

T2-4N0M0 CRC. J Gastrointest Surg 2007; 11:660-5.

2. Swanson RS, Compton CC, Stewart AK, Bland KI. The prognosis of T3N0 colon cancer is dependent on the

number of lymph nodes examined. Ann Surg Oncol 2003; 10:65-71.

3. National quality forum endorses consensus standard for diagnosis and treatment of breast and colorectal

cancer: National Quality Forum; first posted April 12, 2007, web based press release. Available at : http//

www.facs.org/cancer/qualitymeasures.html.



morbidity in MSLT-I, an international multicenter trial. Ann Surg 2005; 242:311-13


axillary dissection in breast cancer. N Engl J Med 2003; 349: 546-53.

7. Nissan A, Protic M, Bilchik A, Eberhardt J, Peoples GE, Stojadinovic A. Predictive model of outcome of

targeted nodal assessment in colorectal cancer. Ann Surg 2010; 251:265-74.



States Military Cancer Institute Clinical Trials Group Study GI-01. Ann Surg 2007; 245:846-57.

9. Saha S, Sehgal R, Patel M, et al. A multicenter trial of sentinel lymph node mapping in colorectal cancer:

prognostic implications for nodal staging and recurrence. Am J Surg 2006; 191:305-310


metastasis in colonic cancer. Dis Colon Rectum 2008; 51: 891-6.

11. Park JS, Chang IT, Park SJ, et al. Comparison of ex vivo and in vivo injection of blue dye in sentinel lymph

node mapping for colorectal cancer. World J Surg 2009; 33:539-46.

12. Van der Zaag ES, Buskens CJ, Kooij N, et al. Improving staging accuracy in colon and rectal cancer by

sentinel lymph node mapping: a comparative study. Eur J Surg Oncol 2009; 35:1065-70.

13. Spatz H, Probst A, Oruzio D, et al. Sentinel lymph node mapping as a side-effect of colonoscopic tattooing.

Surg Endosc 2010; 24 :589-93.

14. Sobin LH, Wittekind C. International Union Against Cancer classification of malignant tumours (ed 6). New

York, Wiley-Liss, 2002.

15. Van der Zaag ES, Kooij N, Van de Vijver MJ, et al. Diagnosing occult tumour cells and their predictive value

in sentinel nodes of histologically negative patients with colorectal cancer. Eur J Surg Oncol 2010; 36:350-7



17. Dawson K, Wiebusch A, Thirlby RC. Preoperative tattooing and improved lymph node retrieval rates from

colectomy specimens in patients with colorectal cancers. Arch Surg 2010;145:826-830

18. Wong JH, Severino M, Honnebier MB, et al. Number of nodes examined and staging accuracy in colorectal

carcinoom. J Clin Oncol 1999; 17:2896-2900.

19. Goldstein NS, Sanford W, Coffey M, Layfield LJ. Lymph node recovery from colorectal resection specimens

removed from adenocarcinoma: Trends over time and a recommendation for a minimum number of lymph

nodes to be recovered. Am J Clin Pathol 1996; 106:209-16.

20. Hashiguchi Y, Hase K, Ueno H, et al. Prognostic significance of the number of lymph nodes examined in

colon cancer surgery. Ann Surg 2010; 251:872-81.



22. Iddings D, Ahmad A, Elashoff D, Bilchik A. The prognostic effect of micrometastases in previously staged

lymph node negative (N0) colorectal carcinoma: a meta-analysis. Ann Surg Oncol 2006;13:1386-1392.

23. Faerden AE, Sjo OH, Bukholm IR et al. Lymph node micrometastases and isolated tumor cells influence

survival in stage I and II colon cancer. Dis Colon Rectum 2011;54:200-206.

the extent of the operation and failures are therefore unlikely to result in under-treat-ment. It has been suggested that with a stricter patient selection and more adequate technical experience in SN mapping, results might become more accurate. We previously demonstrated that SN procedures are more accurate in colonic than in rectal cancer12 and in a recent report the number of SNs per specimen and the presence of tumour replaced nodes were identified as independent predictors.7 In addition, it was demonstrated that better results were obtained in early carcinomas.28 In 225 T1/2 carcinoma of two prospective databases, they found a significantly higher sensitivity and negative predictive value (89% and 97% respectively) when compared to more advanced carcinomas.

Conclusion

We demonstrate that SN mapping with immunohistochemical analysis improves nodal staging of patients with CRC. The SN procedure results in more harvested lymph nodes with a higher proportion of N1 patients. The more accurate staging of SN mapping is reflected in a better prognosis of true N0 patients after resection. Therefore, performing a SN procedure leads to better identification of patients possibly benefiting from adjuvant therapy. Prospective analysis should be performed to focus on survival in patients with micrometastases and isolated tumour cells in their SN. Randomised trials should be conducted to assess whether the administration of adjuvant therapy to these patients results in a better prognosis.

112 113

6


24. Bilchik A, Nissan A, Wainberg Z, et al. Surgical quality and nodal ultrastaging is associated with long-term

disease-free survival in early CRC: an analysis of 2 international multicenter prospective trials.Ann Surg.

2010;252(3):467-74


a prospective multicenter trial. Ann Surg 2007;246(4):568-75

26. Nicholl M, Bilchik AJ. Is routine use of sentinel node biopsy justified in colon cancer? Ann Surg Oncol 2008;

15:1-3.

27. Des Guetz G, Uzzan B, Nicolas P, et al. Is sentinel lymph node mapping in CRC a future prognostic factor? A

meta-analysis. World J Surg 2007; 31:1304-12.

28. Cahill RA, Bembenek A, Sirop S, et al. Sentinel node biopsy for the individualization of surgical strategy for

cure of early-stage colon cancer. Ann Surg Oncol. 2009;16(8):2170-80

7The effect of colonoscopic tattooing on lymph node retrieval and sentinel node mapping

We determined if colonoscopic tattooing contributes to staging accuracy

by analysing whether it leads to a higher lymph node yield and to determine its

accuracy as a sentinel lymph node procedure.

117 7

the effect of colonoscopic tattooing on lymph node retrieval and sentinel node mapping

Introduction

In colorectal cancer (CRC) colonoscopic tattooing with Indian ink is performed to mark the tumour site before surgery. Colonoscopic tattooing is necessary to localize the tumour during surgery, particularly in laparoscopic resection where palpation is not possible, in mucosal lesions and in polypectomy sites.1, 2 In addition to tumour localisation, colonoscopic tattooing may serve more purposes. The carbon particles in the Indian ink are engulfed by macrophages in the lymph nodes (LN) which are therefore marked permanently.3 These marked LNs often become apparent during surgery thereby demonstrating the lymphatic drainage of the tumour. When these nodes are outside the planned resection, surgery can be adjusted. Since these blue nodes are still visible during subsequent pathological examination of the specimen, LN harvest might be facilitated potentially leading to a higher LN yield.4 The extent of nodal resection has important implications in terms of diagnosis, treatment and survival in patients with CRC. Several studies suggested that long-term survival is correlated with the number of LNs retrieved.4-7 In the American Joint Committee on Cancer (AJCC) staging manual a minimum of 12 LNs is recommended for optimal CRC staging in stage I and II patients not receiving adjuvant therapy.8 Colonoscopic tattooing may also serve an additional role as sentinel lymph node (SN) mapping, although it will not discriminate between first and second echelon nodes.9 Both in vivo and ex vivo procedures have been frequently studied, but SN mapping is still not part of daily practice. This is probably due to the conflicting results. SN mapping in CRC does not aim to reduce the extent of the resection, as is common practice in breast cancer surgery, but it aims to refine staging by marking the LNs most likely to contain metastases. Ultrastaging methods such as multi- sectioning and immunohistochemical analysis could then be confined to the SN, thereby limiting the labour-intensive techniques but still retain accurate staging.10, 11 Aim of our study was to determine if colonoscopic tattooing can contribute to staging accuracy by analysing whether it leads to a higher lymph node yield per specimen and to determine its accuracy as a SN procedure.

Patients & methods

All patients who underwent an elective resection for CRC after colonoscopic tattooing between July 2005 and September 2009 were identified from the prospective surgical databases of the two participating hospitals; a university hospital and a teaching hospital. The control group consisted of a series of 210 patients who had an elective resection for CRC in the same period of time and in the same hospitals,

118 119 7

chapter 7 the effect of colonoscopic tattooing on lymph node retrieval and sentinel node mapping

If the endoscopic tattoo was originally placed to mark the site of a possibly incomplete polypectomy site, a single bleb of SPOT ink was deposited at 3 centimetres of the lesion, to prevent the ink causing fibrosis in the polypectomy scar.

Pathological examination Specimens were examined according to standardised CRC protocols, specific for resection of a rectum or colon carcinoma. The pathologist identified LNs by sight and palpation, location of the LNs in relation to the tumour was recorded and then all identified LNs were extracted. No special fat clearance techniques were used. Next, 4 μm thick haematoxylin and eosin (H&E) stained slides were taken per 5 mm of each LN. Staging and grading of the tumour was done according to the TNM-classification of the AJCC (6th edition). Identification of tattooed LNs was done retrospectively by microscopic detection of carbon dye in the H&E stained slides, since pathologic handling of the specimen will not change the histomorphology of carbon-particle containing LNs (in contrast to other blue dyes used for SN mapping).

Statistical analysis Continuous data were presented as means ± standard deviations or as medians and interquartile ranges according to distribution. Categorical data were presented as frequencies or percentages. For dichotomous outcomes, groups were compared by means of the Chi-square test or Fisher’s exact test. Independent t-test was used to compare means and Mann-Whitney U test was used for continuous, not normally distributed outcomes. Univariable and subsequent multivariable linear regression was used to determine independent predictive factors for a higher LN yield. Log-transformation of the outcome variable was performed, since it was not normally distributed. Variables with a p-value of >0.2 in univariable analysis were then entered into the multivariable analysis. A p-value <0.05 was considered to be statistically significant. Statistical analyses were performed with PASW Statistics for Windows (version 18).

Results

A total of 95 patients with CRC who underwent preoperative endoscopic tattooing were included in the study and 210 patients served as controls. The patient flow chart is shown in Figure 1. Baseline characteristics for the tattooed and non-tattooed patients are presented in Table 1. The tumour was located in the colon in 229 patients and in the rectum in 76 patients. Of all patients with rectal cancer, 59 patients (78%) underwent preoperative radiotherapy, of whom 47 patients received 5x5 Gray and 12 patients were treated with neoadjuvant chemoradiotherapy.

without having had a pre-operative tattoo. To be included in the study, histopathology of the specimen had to show an adenocarcinoma of the colon or rectum. Patient data were retrieved from the surgical database, the operative notes and the pathology reports. One hundred three patients (96 in the university hospital and seven in the teaching hospital) participated in a randomised clinical trial in which laparoscopic and open (segmental) colectomy for CRC were compared within a ‘fast track’ or ‘standard care’ perioperative program (LAFA-trial NTR222).12

Outcome measures Primary outcome measures were the total number of LNs retrieved per specimen, the detection rate with the total number of blue nodes, and the sensitivity of preoperative endoscopic tattooing as a SN procedure. Several factors are known to influence the LN retrieval. Earlier studies have reported that LN retrieval is influenced by site of the tumour, sex, age, TNM staging of the tumour, and preoperative (chemo)radiotherapy.13-15 To address all these possible confounding factors, a multivariable analysis was performed to determine if patients who underwent preoperative endoscopic tattooing had a higher LN yield. Detection rate was defined as the percentage of all tattooed patients in which carbon was observed in one of the LNs. Accuracy parameters are based on all the patients in whom tattooing was performed successfully, i.e. patients with carbon present in the LNs. Sensitivity refers to the number of times the carbon positive node reflects the fact that nodal disease is present. The pathologist determined if a LN was tumour positive, based on standard histopathological evaluation (further described below) without performing ultrastaging. A true positive LN was defined as the presence of carbon in a tumour positive LN. False negatives were defined as the presence of tumour negative carbon LNs whereas other LNs in the same specimen were tumour positive. In SN clinical practice, the false positive rate is zero per definition.

Colonoscopic tattooing Endoscopic tattooing was performed by a gastroenterologist according to protocol, preferably a maximum of 30 days before surgery. SPOT ink was used, a suspension containing highly purified, very fine carbon particles (SPOT, GI Supply, Camp Hill, PA, USA).16 The tattooing agent was delivered through an injection needle passed through the working channel of the endoscope. If the lesion was located in the rectosigmoid, the exact distance to the anus was measured and tattoos were placed 1-2 centimetres distally from the lesion. If the lesion was higher up in the colon, tattoos were placed both distally and proximally on both sides of the colon (i.e. at 3 and 9 o’clock).

120 121 7


tumour located in the colon, higher T-staging and higher N-staging. Preoperative treatment with (chemo)radiotherapy was not entered as a separate variable into the multivariable analysis, since it is directly related to the location of the tumour.

Of the 95 tattooed patients, carbon was observed in the LNs in 67 patients, resulting in a 71% detection rate with an average of five blue nodes per patient. The patients with successful and unsuccessful tattoos were compared for baseline characteristics, but no significant differences were found. The only remarkable finding was that all three patients with a T4 tumour were tattooed unsuccessfully as was the one patient undergoing neoadjuvant chemoradiotherapy. Per-operatively identified blue nodes led to adjustment of the planned resection in two patients. In one patient, a blue node was found outside the regional mesentery (aberrant drainage) and in one patient a more extensive resection was performed since blue nodes were seen along the central vasculature. Clinical, pathological and histological appearances of carbon containing LNs are shown in Figures 2, 3 and 4. Significantly more LNs were retrieved in the patients with carbon present in the LNs compared to the patients with no carbon present; median 15 LNs [IQR 10-20] versus 12 LNs [IQR 9-16] (p=0.014), respectively. In the multivariable analysis (Table 2) the presence of carbon containing LNs showed to be a significant independent predictive factor for a higher LN yield (p=0.005), together with the following variables:

Figure 1 Patient stream by application of inclusion criteria. Table 1 Baseline characteristics for tattoo versus non-tattoo.

Tattoo n=95 No tattoo n=210 p-value

Male gender 49 (51.6) 118 (56.2) 0.454

Age 68 [61 - 76] 69.0 [58 - 76] 0.798

Tumour location 0.056

Colon 78 (82.1) 151 (71.9)

Rectum 17 (17.9) 59 (28.1)

Radiotherapy

5x5 Gray 9 (52.9) 38 (64.4) 0.078

Chemoradiotherapy 1 (5.9) 11 (18.6)

Laparoscopic resection 49 (51.6) 79 (37.6) 0.022

T stage 0.004

T1 19 (20.0) 15 (7.1)

T2 19 (20.0) 50 (23.8)

T3 54 (56.8) 126 (60.0)

T4 3 (3.2) 19 (9.0)

N stage 0.248

N0 66 (69.5) 131 (62.4)

N1 17 (17.9) 44 (21.0)

N2 12 (12.6) 35 (16.7)

M stage 1.000

M0 86 (90.5) 179 (85.2)

M1 3 (3.2) 8 (3.8)

MX 6 (6.3) 23 (11.0)

Tumour differentiation 0.027

well 6 (6.3) 10 (4.8)

moderate 70 (73.7) 149 (70.9)

poor / undifferentiated 10 (10.5) 45 (21.5)

unknown 9 (9.5) 6 (2.9)

Data presented as n (%) or median [IQR].

122 123 7


If preoperative tattooing was analysed for SN mapping accuracy parameters, LN involvement in carbon-containing nodes was found in 20 patients of the 24 successfully tattooed N1 patients. This resulted in a sensitivity of 83% with an accompanying false negative rate of 17%.( Table 3) The overall accuracy of predicting LN status was 94% (63/67).

Figure 2 Coloured lymph nodes visible during laparoscopy after tattooing.

Figure 4 Ink patches in H&E stained lymph nodes slides (magnification 4x).

Figure 3 Black lymph nodes visible in resected specimen.

Table 2 Linear regression of total number of lymph nodes retrieved.

VariableUnivariable

p-valueMultivariable

p-valueOR (95% CI)

Sex 0.862 - -

Age 0.229 - -

Laparoscopic vs. open surgery 0.755 - -

Colon vs. rectal cancer 0.007 0.005 1.24 (1.07 - 1.44)

(Chemo)radiotherapy in rectal cancer 0.058 - -

T-stage 0.004 0.004 1.14 (1.04 - 1.24)

N-stage 0.011 0.038 1.10 (1.01 - 1.20)

M-stage 0.206 - -

Tumour differentiation 0.502 - -

Carbon present 0.012 0.005 1.26 (1.07 - 1.47)

Analysis were carried out on log-transformed data. CI=confidence interval

Table 3 Accuracy of endoscopic tattooing as a sentinel node procedure.

Detection rate 67/95 (70.5)

True positives 20

False negatives 4

Sensitivity rate 20/24 (83.3)

Accuracy rate 63/67 (94)

Median no. of CcLN per specimen 5.0 [4.0 - 6.0]

Percentage of CcLN per specimen 33.3 [20.8 - 55.6]

Data presented as n (%) or median [IQR]. CcLN = carbon-containing lymph node.

124 125 7


will improve survival in these patients is currently recruiting patients (EnROUTE; NCT01097265). If ultrastaging becomes part of daily practice, preoperative tattooing could be used to limit expensive and time-consuming methods to the carbon containing LNs, on average 33% of total LNs in this series. As with all observational studies, conclusions have to be drawn with caution. Baseline differences of the studied patients were different for several variables, but were corrected for by means of a multivariable analysis. In our study the tattooed patients relatively often had a T1 or T2 tumours, and nearly 70% had no LN metastasis. However, this is the exact group of patients for which accuracy of the LN staging is most important. If pN0 patients are actually understaged pN1 patients, they miss out on adjuvant chemotherapy, thereby significantly decreasing their survival.10, 11 Moreover, with the increasing implementation of CRC screening programs a considerable rise in detection of T1 tumours and polypectomies is to be expected. Endoscopic tattooing is safe and long-lasting when injected tangentially in small aliquots as described in a recent published review.18 SN detection after endoscopic tattooing is appealing because it is a side-effect of a procedure that is often performed in daily practice, not only in specialised centres. Thus, no extra efforts concerning facilities, cost, labour and operation time are needed, contrary to regular SN techniques.

Conclusion

Since our study demonstrates several advantages of pre-operative endoscopic tattooing for LN staging in addition to tumour marking, we recommend this procedure as standard routine work-up in CRC patients

Discussion

We performed this study to determine if preoperative endoscopic tattooing can contribute to improve staging. The blue nodes can be used peroperatively to guide surgery and regional lymphadenectomy. In addition, a significantly higher median number of LNs was retrieved per specimen after successful endoscopic tattooing, even after correction for possible confounding factors. As a SN procedure, endoscopic tattooing had an acceptable sensitivity of 83% and an accuracy of 94%. Detection rate, however, was slightly disappointing (71%). The increased LN yield can be attributed to the fact that the nodes become visibly black and are therefore more easily identified by both the surgeon and the pathologist.4 Studies have demonstrated that an increased number of harvested LNs leads to better prognostication with a higher proportion of detected pN1 patients benefitting from adjuvant chemotherapy and an accompanying better prognosis in the pN0 group.6, 8 A major drawback of our study is the non-randomised design. However, except for T stage baseline characteristics were comparable. This is most likely due to the fact that patients with small tumours and incomplete polypectomies were more often tattooed to facilitate detection during surgery. So far, there is not much experience with endoscopic tattooing as SN mapping. There is no definite proof yet that the carbon containing LNs are indeed the first echelon nodes, but it does seem very plausible that the greatest degree of ink is deposited in the first order draining nodes.3 However, secondary echelon nodes cannot be discriminated from first echelon nodes with this technique. This will probably result in a higher number of detected SNs. In this study a high number of blue nodes was found, when compared to the mean number of SNs described in a recent systematic review.10 Nissan et al. identified the number of SNs as an independent predictive factor for accuracy parameters. Indeed, the sensitivity found in this study is higher than other reported sensitivity rates in CRC SN mapping. In contrast, our endoscopic tattooing detection rate was lower than in comparable studies9, 17, and the mean 94% detection rate found in the meta-analysis combing in vivo and ex vivo SN mapping results.10 Although a standard protocol was used, tattooing before CRC resection is not yet done on a routine base in either centre. This could explain the relatively high technical failure rate. With endoscopic tattooing becoming part of quality standards developments a considerable increase of preoperative tattooing in the near future is expected, thereby improving skills and thereupon detection rates. No additional ultrastaging methods were applied in this study, so upstaging rates could not be determined. So far the prognostic value of occult tumour cells is still debated and serial sectioning with immunohistochemistry is not part of routine pathological staging. A large randomised trial to prove the clinical significance of mi-crometastases and occult tumour cells in LN and whether adjuvant chemotherapy

126 127 7


References

1. Yeung JM, Maxwell-Armstrong C, Acheson AG. Colonic tattooing in laparoscopic surgery - making the

mark? Colorectal Dis 2009;11:527-530.

2. Cho YB, Lee WY, Yun HR, Lee WS, Yun SH, Chun HK. Tumor localization for laparoscopic colorectal surgery.

World J Surg 2007;31:1491-1495.

3. Cahill RA, Lindsey I, Cunningham C. Sentinel node mapping by colonic tattooing. Surg Endosc 2010;24:2365-

2366.

4. Dawson K, Wiebusch A, Thirlby RC. Preoperative tattooing and improved lymph node retrieval rates from

colectomy specimens in patients with colorectal cancers. Arch Surg 2010;145:826-830.

5. Tsai HL, Lu CY, Hsieh JS, Wu DC, Jan CM, Chai CY, Chu KS, Chan HM, Wang JY. The prognostic significance of

total lymph node harvest in patients with T2-4N0M0 colorectal cancer. J Gastrointest Surg 2007;11:660-665.

6. Chen SL, Bilchik AJ. More extensive nodal dissection improves survival for stages I to III of colon cancer: a

population-based study. Ann Surg 2006;244:602-610.

7. Swanson RS, Compton CC, Stewart AK, Bland KI. The prognosis of T3N0 colon cancer is dependent on the

number of lymph nodes examined. Ann Surg Oncol 2003;10:65-71.

8. Edge S.B., Byrd D.R., Compton C.C., Fritz A.G., Greene F.L., Trotti A. AJCC Cancer Staging Manual. New York,

NY: Springer, 2010.

9. Spatz H, Probst A, Oruzio D, Anthuber M, Messmann H, Arnholdt HM, Markl B. Sentinel lymph node mapping

as a side-effect of colonoscopic tattooing. Surg Endosc 2010;24:589-593.

10. van der Pas MH, Meijer S, Hoekstra OS, Riphagen II, de Vet HC, Knol DL, van Grieken NC, Meijerink WJ. Senti-

nel-lymph-node procedure in colon and rectal cancer: a systematic review and meta-analysis. Lancet

Oncol 2011;12:540-550.

11. van der Zaag ES, Buskens CJ, Kooij N, Akol H, Peters HM, Bouma WH, Bemelman WA. Improving staging


Oncol 2009;35:1065-1070.

12. Vlug MS, Wind J, Hollmann MW, Ubbink DT, Cense HA, Engel AF, Gerhards MF, van Wagensveld BA, van der

Zaag ES, van Geloven AA, Sprangers MA, Cuesta MA, Bemelman WA. Laparoscopy in Combination with Fast

Track Multimodal Management is the Best Perioperative Strategy in Patients Undergoing Colonic Surgery:

A Randomized Clinical Trial (LAFA-study). Ann Surg 2011.

13. Field K, Platell C, Rieger N, Skinner I, Wattchow D, Jones I, Chen F, Kosmider S, Wohlers T, Hibbert M, Gibbs P.

Lymph node yield following colorectal cancer surgery. ANZ J Surg 2011;81:266-271.

14. Ha YH, Jeong SY, Lim SB, Choi HS, Hong YS, Chang HJ, Kim DY, Jung KH, Park JG. Influence of preoperative

chemoradiotherapy on the number of lymph nodes retrieved in rectal cancer. Ann Surg 2010;252:336-340.

15. Chou JF, Row D, Gonen M, et.al. Clinical and pathologic factors that predict lymph node yield from surgical

specimens in colorectal cancer: a population-based study. Cancer 2010;116:2560-2570.

16. Askin MP, Waye JD, Fiedler L, Harpaz N. Tattoo of colonic neoplasms in 113 patients with a new sterile

carbon compound. Gastrointest Endosc 2002;56:339-342.

17. Feingold DL, Addona T, Forde KA, Arnell TD, Carter JJ, Huang EH, Whelan RL. Safety and reliability of

tattooing colorectal neoplasms prior to laparoscopic resection. J Gastrointest Surg 2004;8:543-546.

18. Kethu SR, Banerjee S, Desilets D, Diehl DL, Farraye FA, Kaul V, Kwon RS, Mamula P, Pedrosa MC, Rodriguez SA,

Wong Kee Song LM, Tierney WM. Endoscopic tattooing. Gastrointest Endosc 2010;72:681-685.

Summary, conclusions and future perspectives

Samenvatting

Dankwoord

Curriculum vitae

131

summary, conclusions and future perspectives

Summary

In this thesis the role of the sentinel lymph node (SN) procedure in colorectal surgery is described. The aim of our research was to assess the accuracy of SN mapping in staging patients with colorectal cancer (CRC). In Chapter 1 we determined the accuracy of this procedure from published data and identified factors that contribute to the conflicting reports in the literature. A systematic search of the Medline, Embase and Cochrane databases from 1999 to July 2011 revealed 98 potentially eligible studies, of which 57 were analysed including 3934 patients. The pooled SN identification rate was 91% with a higher identification rate in studies including more than 100 patients or studies using the ex vivo SN technique. The pooled sensitivity of the SN procedure was 70% with an overall accuracy of 88%. Subgroups with higher sensitivity were identified: ³4 SNs versus <4 SNs (85% versus 66%, p=0.003), colon versus rectal cancer (78% versus 66%, p=0.04), and T1/T2 versus advanced T3/T4 carcinomas (93% versus 59%, p=0.01). Serial sectioning and immuno-histochemical analysis resulted in a mean upstaging of 19%. The rate of upstaging, defined as a micrometastasis (pN1mi+) rather than isolated tumour cells (ITC) (pN0itc+) was 8%. If one included the immunohistochemical findings, the pooled sensitivity of the SN procedure increased to 80%. Our meta-analysis demonstrates an overall disappointing sensitivity of SN mapping in colorectal patients. However, in early staged colon cancer the SN procedure has acceptable accuracy rates and refines staging. To compare the predictive value of SN mapping between patients with colon and rectal cancer a prospective comparison was performed, described in Chapter 2. An ex vivo SN procedure was performed in 100 patients with colon and 32 patients with rectal cancer. If the SN was negative, immunohistochemical analyses using two different antibodies against cytokeratins (Cam5.2, and CK 20) and one antibody against epithelial cells (BerEp-4) were performed to detect occult tumour cells (OTC). According to the America Joint Committee on Cancer (AJCC) ITC (< 0.2mm) were discriminated from micrometastases (0.2-2mm). A SN was identified in 117 patients and it accurately predicted nodal status in 106 patients (accuracy 91%). Both sensitivity and negative predictive value were higher in colon carcinomas than in rectal carcinomas (83% versus 57%, p=0.06 and 93% versus 65%, p=0.002 respectively). In patients with extensive lymph node metastases the SN procedure was unsuccessful. Eleven of the 13 unsuccessful SN procedures occurred in patients with rectal cancer who had had pre-operative radiotherapy. After immunohistochemical analysis, 21 of the 73 N0 patients had OTC in their SN; eight patients had micrometastases and 13 patients had ITC. The SN mapping accurately predicted nodal status in patients with colon cancer. Immunohistochemical analysis demonstrated micrometastatic disease in eight out of 73 N0 patients, with a true upstaging rate of 11%. However, SN mapping was less reliable in patients with rectal cancer after pre-operative radiotherapy.

132 133

summary, conclusions and future perspectives summary, conclusions and future perspectives

patients operated on for CRC was analysed. After conventional H&E staining 107 patients without lymphatic metastases were included; 59 patients had open surgery, 48 patients underwent laparoscopic resection. Patients in the laparoscopic group underwent a no touch medial to lateral approach whereas the conventional lateral to medial approach was applied in open surgery. A SN procedure was performed in all patients. The SNs were analysed for the presence of OTC. These tumour cells were divided into micrometastases (0.2-2mm) or ITC (<0.2mm). In ten patients micro-metastases were found equally distributed over the groups. However, ITC were more often found after open surgery (18 versus 5 patients, p=0.03). The presence of OTC was related to the depth of tumour invasion and tumour diameter (>3.5cm). Logistic regression analysis identified lymphangio invasion as a predictor for micrometastases (OR 18.4) whereas open resection was predictive for the presence of ITC (OR 3.3). The no touch medial to lateral laparoscopic surgery resulted in less isolated tumour cells in lymph nodes compared to open surgery in patients with stage I and II colorectal cancer. Every attempt should be made to prevent worsening of the prognosis of patients during surgery. In that respect laparoscopic resection with the no touch technique may have benefits over open surgery in patients with stage I and II colorectal cancer. SN mapping for staging in patients with CRC remains controversial because survival analyses are lacking. In Chapter 6 we prospectively assessed the effect of SN mapping on nodal staging and its implication on survival in these patients. Between November 2005 and July 2009, 331 patients underwent a resection. In 189 patients (group A) an SN procedure was performed with immunohistochemical analysis of the SN. Tumour cell deposits between 0.2 mm and 2.0 mm were referred to as micro-metastases (pN1mi+). The remaining patients (n=142, group B) had standard nodal staging. Multivariate cox regression analysis was performed to identify prognostic factors for disease recurrence. The average number of harvested lymph nodes was higher in group A than in group B (16 versus 12, p<0.0001). After conventional staging, 81 patients (43%) had nodal metastasis in group A. This increased to 89 patients (47%) when immunohistochemically detected micrometastases were included. In group B 50 patients (35%) had nodal metastasis. During follow-up, a lower recurrence rate was seen in true pN0 patients after SN mapping revealed no tumour cells when compared to the conventional staging group (4% versus 15%, p=0.04). The SN procedure (hazard ratio: 4.1) was an independent strong predictor of disease recurrence. The SN procedure results in a more accurate staging of patients with CRC. This is reflected by a better prognosis of pN0 patients after SN mapping. In patients with CRC colonoscopic tattooing is performed to mark the tumour site before surgery. Possibly this pre-operative tattooing can also serve as a SN procedure. The aim of our study described in Chapter 7 was to determine if colonoscopic tattooing can contribute to staging accuracy by increasing the lymph node (LN) yield

Most studies on the SN procedure in patients with CRC include immunohisto-chemical analysis of the SN only. To evaluate the real diagnostic accuracy of the SN procedure with immunohistochemical analysis, the presence of OTC in all histologically negative lymph nodes was compared to the presence of these cells in SNs. Also the reproducibility of diagnosing OTC and the sensitivity of three different antibodies was assessed and described in Chapter 3. Between November 2006 en July 2007, an SN procedure was performed in 58 histologically N0 patients with CRC. All lymph nodes (n=908, mean 16) were step-sectioned and immunohistochemistry was performed. Tumour cells were identified in 19 of 58 patients, with micrometastases (0.2-2mm) in seven and ITC (< 0.2mm) in 12 patients. The overall agreement in diagnosing OTC between two independent pathologists was 86%. A SN was identified in 53 of 58 patients. All micrometastases were found in the SNs. In two patients with negative SNs, ITC were demonstrated in non-SNs (sensitivity 88%, overall accuracy 96%). Additional immunohistochemical analysis of histologically negative lymph nodes demonstrated OTC in 33% of the patients resulting in an upstaging rate of 12%. If occult tumour cells are present they can predominantly be found in the SN of patients with colorectal cancer. According to the latest version of the TNM-classification OTC are divided in micro-metastases (pN1mi+) and ITC (pN0itc+). The question is whether this division can be made reliably. In Chapter 4 we assessed the interobserver agreement between pathologists in judging photographs of OTC in lymph nodes of patients with colon cancer. All lymph nodes of 82 pN0 patients with colon cancer were analysed immuno-histochemically with three different monoclonal antibodies against epithelial cells. Digital pictures of the 37 detected lesions were placed on a secured website. Forty randomly selected pathologists were asked to categorise the lesions into ‘micro-metastases’, ‘isolated tumour cells’ or ‘different’ and they had the opportunity to comment. The degree of agreement was calculated by the Kendall W coefficient. Thirty-five pathologists (88%) categorised the 37 lesions. Five lesions (14%) were categorised unanimously as micrometastases or ITC. In 26 pictures (70%) the agreement was poor to moderate. When the analysis was performed only on those diagnoses of which the pathologists were confident about their judgment, the percentage of lesions with good agreement rose to 49%. Differences in agreement were principally associated with multifocal lesions, clusters of tumour cells < 0.2 mm with proliferation characteristics in the parenchyma of the lymph node and lymphangio invasion. The differentiation between micrometastases and ITC in lymph nodes of patients with colon cancer is not uniform. Laparoscopic surgery has the potential of less tumour cell spread because of the no touch isolation technique. In Chapter 5 we assessed the effect of the surgical approach (open versus no touch laparoscopic) on the presence of tumour cells in SNs of patients with stage I and II CRC. A single centre consecutive prospective series of

134 135

summary, conclusions and future perspectives summary, conclusions and future perspectives

performed into the relation of the nature and behaviour of these cells compared to cancer stem cells. It is also, very interesting to assess the relation between OTC detected by immunohistochemical analysis, tumour cell characteristics and genetic tumour cell profile on survival of these patients. If repeated survival analyses show a decreased disease free survival in patients with micrometastases randomised trials should be performed to assess if survival of these patients improves after administration of systemic chemotherapy. In the Netherlands such a randomised trial is currently recruiting patients (EnROUTE; NCT01097265).

per specimen and to determine its accuracy as a SN procedure. In a consecutive series of 95 tattooed patients between 2005 and 2009, all LNs were microscopically examined for the presence of carbon-particles and metastases. A series of 210 patients operated on for CRC in the same time period without tattooing served as controls. A higher LN yield was observed in patients with pre-operative tattooing, median 15 versus 12 (p=0.014). In multivariable analysis the presence of carbon containing LNs showed to be an independent predictive factor for higher LN yield (p=0.005), together with higher T- and N-stage, and tumour location in the colon. Carbon-particles were detected in 71% of tattooed patients with an average of five carbon positive LNs per case. If pre-operative tattooing was used for SN mapping, the overall accuracy of predicting the LN status was 94%. In the 24 patients with lymphatic metastases there were four false negative procedures (sensitivity 83%). After tattooing of colon carcinomas, the LN yield was higher than in a control group, and it can be used as a SN procedure with acceptable accuracy rates. Since we demonstrate several advantages of pre-operative endoscopic tattooing for LN staging in addition to tumour marking, we recommend this procedure as standard routine work-up in patients with colorectal cancer.

Conclusions and future perspectives

The SN procedure accurately predicts nodal status in patients with early staged colon cancer. However, SN mapping is less reliable in patients with rectal cancer after pre-operative radiotherapy. If occult tumour cells are present they are predominantly found in the SN. Additional immunohistochemical analysis of the SN of patients with histologically negative lymph nodes demonstrates OTC in 33% of these patients, resulting in an upstaging rate of 12%. Therefore, the SN procedure results in a more accurate staging of patients and is an independent predictor of disease-free survival. This is reflected by a better prognosis of pN0 patients after SN mapping. Since the ex-vivo SN mapping is an easy and safe procedure we recommend that the SN procedure should always be considered in addition to conventional resection in patients with colon cancer. However, the differentiation between micrometastases and ITC in lymph nodes of patients with colon cancer is not uniform. If patients with OTC in their SN are treated with systemic chemotherapy, better definitions are needed. Especially, multifocal lesions and clusters of tumour cells < 0.2 mm with malignant proliferation character-istics in the parenchyma of the lymph node should be discussed and uniform criteria for these items should be made. It is very important that the prognostic clinical relevance of these OTC will be assessed. Further survival analyses of patients with CRC are needed but with a clear division of micrometastases and ITC. Research should be

137

samenvatting

Samenvatting

In dit proefschrift wordt de rol van de schildwachtklier (SWK) procedure bij colorectale chirurgie beschreven. Het doel van ons onderzoek was de diagnostische en prognostische waarde van deze procedure te beoordelen in de behandeling van patiënten met darmkanker. In Hoofdstuk 1 hebben wij een literatuuronderzoek verricht naar de diagnostische waarde van deze procedure en factoren geïdentificeerd die bijdragen aan de tegenstrijdige rapportage in de literatuur. Een systematische zoektocht van de Medline, Embase en Cochrane databases van 1999 tot juli 2011 toonde 98 onderzoeken, waarvan 57 onderzoeken voldeden aan de kwaliteitscriteria bestaande uit 3.934 patiënten. De SWK werd vaker succesvol geïdentificeerd in onderzoek met meer dan 100 patiënten of wanneer de ex vivo SWK procedure werd gebruikt. In het algemeen was de sensitiviteit van de SWK procedure 70% met een juist voorspelde lymfklierstatus in 88% van de patiënten. De diagnostische waarde was hoger in de volgende subgroepen: wanneer er meer dan vier SWK per patiënt werden gevonden (85% versus 66%), darm versus endel-darmkanker (78% versus 66%), en T1/T2 tumoren versus T3/T4 tumoren (93% versus 59%). Aanvullende stap-serie immunohistochemische analyse van de SWK zorgde voor een herstadiëring in 19% van de patiënten. Echter, herstadiëring gedefinieerd als micrometas-tasen (pN1mi +) in plaats van losse tumorcellen (pN0itc +) was in 8% van de patiënten het geval. Werden deze aanvullende immunohistochemische bevindingen meegenomen dan verbeterde de sensitiviteit van de SWK procedure naar 80%. Deze meta-analyse toont een teleurstellende sensitiviteit van de SWK procedure bij patiënten met darmkanker. Echter, in een vroeg stadium van darmkanker heeft de SWK procedure een betere nauwkeurigheid en kan het de stadiëring verbeteren. De diagnostische waarde van de SWK procedure bij patiënten met darmkanker wordt in Hoofdstuk 2 vergeleken met patiënten met endeldarm kanker. Een SWK procedure werd uitgevoerd bij 100 patiënten met dikke darmkanker en 32 patiënten met endeldarm kanker. Aanvullend immunohistochemisch kleuringen van de SWK, met behulp van drie verschillende monoklonale antilichamen: twee tegen cytokeratines (Cam5.2, en CK 20) en een antilichaam tegen epitheliale cellen (BerEp-4), werden uitgevoerd om eventueel gemiste tumorcellen aan te tonen. Er werd een onderverdeling gemaakt in losse tumorcellen (<0,2 mm) en micrometastasen (0,2-2mm). In 117 patiënten kon een SWK worden gevonden, en bij 106 patiënten was er sprake van een juist voorspellende waarde voor lymfklier metastasering door de SWK (nauwkeurigheid 91%). Zowel de sensitiviteit als de negatief voorspellende waarde was hoger bij dikke darmkanker vergeleken met endeldarm kanker (respectievelijk 83% versus 57% en 93% versus 65%). Bij patiënten met uitgebreide lymfklier metastasen was de SWK procedure niet betrouwbaar. Elf van de 13 mislukte SWK procedures werden uitgevoerd bij patiënten met endeldarm kanker die pre-operatieve radiotherapie ondergingen. Na aanvullende immunohistochemische

138 139

samenvatting samenvatting

de 37 afwijkingen. Bij vijf foto’s was er unanieme overeenstemming over de diagnose ‘micrometastase’ of ‘losse tumorcellen’. Bij 26 foto’s (70%) was de overeenstemming slecht tot redelijk. Indien de analyse werd uitgevoerd bij diagnosen waarvan de pathologen aangaven dat het een zekere diagnose betrof, steeg het percentage foto’s met een goede overeenstemming naar 49%. Verschil in overeenstemming was voornamelijk geassocieerd met multifocale afwijkingen, tumorcel clusters kleiner dan 0,2 mm met proliferatieken-merken in het lymfklierparenchym en lokalisatie van de tumorcellen in vaten. Het onderscheid tussen micrometastasen en losse tumorcellen in lymfklieren bij patiënten met darmkanker blijkt niet eenduidig. Indien deze indeling klinische consequenties krijgt bij de behandeling van patiënten met darmkanker, is het noodzakelijk om betere definities te maken. Tijdens laparoscopische resectie van darmkanker wordt de darm verwijderd met behulp van de no-touch techniek waarbij de darm zo min mogelijk wordt aangeraakt. Door deze methode worden er mogelijk minder tumorcellen verspreid. In Hoofdstuk 5 onderzochten wij het effect van de chirurgische benadering (conventioneel open versus no-touch laparoscopische) op de aanwezigheid van tumorcellen in de SWK van patiënten met stadium I en II darmkanker. Na conventioneel pathologisch onderzoek werden 107 patiënten zonder lymfklier metastase geanalyseerd, 59 patiënten hadden een open operatie ondergaan en 48 patiënten ondergingen een laparoscopische resectie. Patiënten in de laparoscopische groep werden geopereerd van mediaal naar lateraal (no-touch benadering) terwijl in de open chirurgie groep de conventionele benadering van lateraal naar mediaal werd toegepast. De SWK werd immunohistochemisch geanalyseerd op de aanwezigheid van tumorcellen. Deze tumorcellen werden verdeeld in micrometastasen (0,2-2mm) of losse tumorcellen (<0,2 mm). In totaal werden bij tien patiënten micrometas-tasen gevonden gelijk verdeeld over beide groepen. Maar losse tumorcellen werden vaker aangetroffen na een open operatie (18 versus 5 patiënten). De aanwezigheid van tumorcellen was gerelateerd aan de diepte van de tumor ingroei in de darmwand en tumor diameter (> 3,5 cm). De aanwezigheid van lymfangio invasie was voorspellend voor micrometastasen (OR 18,4), terwijl het type operatie (open resectie) voorspellend was voor de aanwezigheid van losse tumorcellen (OR 3,3). De no-touch laparoscopische benadering resulteert, in vergelijking met open chirurgie, in minder losse tumorcellen in lymfklieren bij patiënten met stadium I en II darmkanker. De SWK procedure bij darmkanker blijft controversieel omdat overlevings analyse vooralsnog ontbreken. In hoofdstuk 6 hebben wij het effect van de SWK procedure op de overleving bij patiënten met darmkanker onderzocht. Tussen november 2005 en juli 2009 ondergingen 331 patiënten een resectie. Bij 189 patiënten (groep A) werd een SWK procedure uitgevoerd met aanvullende immunohistochemische analyse. Tumorcel deposities tussen 0,2 mm en 2,0 mm werden aangeduid als micrometastasen (pN1mi +). Bij de overige 142 patiënten (groep B) werd conventioneel pathologisch onderzoek naar de lymfklieren verricht. Multivariate Cox regressie analyse werd uitgevoerd om

kleuringen werden bij 21 van de 73 patiënten zonder lymfklier metastase (N0) alsnog tumorcellen in hun SWK gevonden; acht patiënten hadden micrometastasen en 13 patiënten hadden losse tumorcellen. De SWK procedure kan dus bij patiënten met dikke darmkanker de lymfklier status nauwkeurig voorspellen. Immunohistochemisch onderzoek van de SWK toont bij acht van de 73 N0 patiënten alsnog micrometastase, resulterend in een herstadiëring in 11% van deze patiënten. Echter, bij patiënten met endeldarm kanker na pre-operatieve radiotherapie is de SWK procedure minder betrouwbaar. Het meeste onderzoek naar de SWK procedure bij darmkanker omvat alleen immuno-histochemisch onderzoek naar occulte tumorcellen (OTC) van de SWK zelf. Evaluatie van de echte diagnostische nauwkeurigheid van de SWK procedure is een vergelijking van de aanwezigheid van OTC in de SWK met de aanwezigheid van deze cellen in alle histologisch negatieve lymfklieren waarbij alle klieren aanvullend immunohistochemisch gekleurd worden. In Hoofdstuk 3 beschrijven wij deze vergelijking en beoordelen ook de reprodu-ceerbaarheid van de diagnose van OTC. Tevens vergeleken wij de gevoeligheid van de drie verschillende antilichamen voor het aantonen van OTC. Tussen november 2006 en juli 2007, werd de SWK procedure uitgevoerd in 58 histologisch N0 patiënten met darmkanker. Alle lymfklieren (n = 908, gemiddelde 16) werden stap-serie immunohistochemisch onderzocht met behulp van drie verschillende monoklonale antilichamen. OTC werden bij 19 van de 58 patiënten gevonden met micrometastasen (0,2-2mm) bij zeven patiënten en losse tumorcellen (<0,2 mm) bij 12 patiënten. De overeenkomst in de diagnose van OTC tussen twee onafhankelijke pathologen was 86%. Een SWK werd gevonden bij 53 van 58 patiënten. Alle micrometastasen werden uitsluitend gevonden in de SWK. Bij twee patiënten werden buiten de SWK, losse tumorcellen aangetoond in lymfklieren (sensitiviteit 88% en accuratesse 96%). Extra immunohistochemische kleuringen van histologisch negatieve lymfklieren toont dus tumorcellen in 33% van de patiënten, wat leidt tot een herstadiëring in 12% van de gevallen. Indien OTC aanwezig zijn dan zijn deze voornamelijk te vinden in de SWK. Daarom kan de SWK de lymfklier status betrouwbaar in kaart brengen en verbetert het de stadiëring van patiënten met darmkanker. Volgens de meest recente versie van de TNM-classificatie worden OTC onderverdeeld in micrometastasen (pN1mi +) en geïsoleerde tumorcellen (pN0itc +). De vraag is of dit onderscheid betrouwbaar kan worden uitgevoerd. Wij hebben, in Hoofdstuk 4, de overeenkomst onderzocht tussen pathologen in het beoordelen van foto’s van deze OTC in lymfklieren van patiënten met darmkanker. Van 82 pN0 patiënten werden alle lymfklieren geanalyseerd met drie monoklonale antilichamen. Van de 37 gedetecteerde afwijkingen werden digitale foto’s gemaakt en op een beschermde website geplaatst. Veertig willekeurig geselecteerde pathologen werd gevraagd deze foto’s te bekijken en de afwijkingen te beoordelen als ‘micrometastase’, ‘losse tumorcellen’ of ‘anders’ met de mogelijkheid commentaar te geven. De mate van overeenstemming werd bepaald door berekening van de Kendall W-coëfficiënt. Vijfendertig pathologen (88%) beoordeelden

140 141

samenvatting list of abbreviations

List of abbreviations / Afkortingen

AJCC American Joint Committee on CancerUICC The International Union Against CancerCRC Colorectal cancer / Colorectaal carcinoomLN(s) Lymph node(s)H&E Haematoxyline and Eosin stainingCTC Circulating tumour cellsOTC Occult tumour cells / Occulte tumorcellen; tumorcellen niet ontdekt tijdens

routine pathologisch onderzoekITC Isolated tumour cells; tumour cells smaller than 0.2 mmRT-PCR Real-time polymerase chain reactionSN Sentinel lymph nodeSWK SchildwachtklierpN0 Geen lymfklier uitzaaiingen na conventioneel pathologisch onderzoekpN1 Lymfklier uitzaaiingen aangetoond door pathologisch onderzoek

prognostische factoren te identificeren voorspellend voor ziekte vrije overleving. Het gemiddeld aantal lymfklieren was hoger in groep A dan in groep B (16 versus 12). Na conventioneel pathologisch onderzoek werd bij 81 patiënten (43 procent) lymfklier metastase aangetoond (groep A). Dit steeg tot 89 patiënten (47 procent) na toevoeging van immunohistochemisch onderzoek naar micrometastasen in de SWK. In groep B hadden 50 patiënten (35 procent) lymfklier metastase. Tijdens de follow-up werd in groep A een lager recidief percentage gezien bij de ware pN0 patiënten zonder OTC dan in groep B (4 procent ten opzichte van 15 procent). De SWK procedure (hazard ratio: 4,1) was een onafhankelijke voorspeller voor ziekte vrije overleving. Deze procedure resulteert in een meer nauwkeurige stadiëring van patiënten met darmkanker. Dit komt tot uiting door een betere prognose van de ware pN0 patiënten na de SWK procedure. Wij raden daarom aan dat de SWK procedure verricht wordt tijdens de chirurgische resectie van darmkanker. Tijdens de SWK procedure wordt patent blauw rondom de tumor gespoten; hierdoor kleurden de lymfklieren die met de tumor verbonden zijn blauw. Echter, voorafgaand aan de operatie wordt bij patiënten met darmkanker met behulp van colonoscopie de plaats van de tumor gemarkeerd met inkt (tatoeage). Deze methode zou ook kunnen dienen als SWK procedure. Het doel van ons onderzoek, beschreven in Hoofdstuk 7, was te bepalen of deze markering de stadiëring kan verbeteren door het verhogen van het aantal lymfklieren per preparaat en de sensitiviteit ervan te bepalen als SWK procedure. Tussen 2005 en 2009 werden in een opeenvolgende reeks van 95 getatoeëerde patiënten alle lymfklieren microscopisch onderzocht op de aanwezigheid van inkt deeltjes en tumorcellen. De controle groep bestond uit een reeks van 210 geopereerde patiënten zonder deze tatoeage. Bij patiënten met pre-operatieve tatoeage van de tumor werd een hoger aantal lymfklieren waargenomen (mediaan 15 versus 12). In een multivariate analyse bleek de aanwezigheid van inkt houdende lymfklieren een voorspellende factor voor een hoger aantal lymfklieren. Andere voorspellende factoren waren een verder gevorderd stadium van darmkanker en de locatie van de tumor in de darm. Inktdeeltjes werden aangetoond in 71% van de getatoeëerde patiënten met een gemiddelde van vijf inkt houdende lymfklieren per patiënt. Indien het pre-operatief tatoeëren gebruikt zou worden als SWK procedure, dan is de juist voorspellende waarde van lymfkliermetastase 94%. In 24 patiënten met lymfklier metastase waren er vier fout negatieve SWK procedures(sensitiviteit 83%). Na colonoscopisch tatoeëren van patiënten met darm - kanker is de lymfklieropbrengst na de resectie hoger en zou mogelijk deze methode gebruikt kunnen worden als SWK procedure. Daarom adviseren wij het pre-operatief colonoscopisch tatoeëren altijd toe te passen bij patiënten met darmkanker.

143

list of publications

List of publications

1. Decreased incidence of isolated tumour cells in lymph nodes after laparoscopic resection for colorectal cancer.

van der Zaag ES, Buskens CJ, Vlug MS, Peters HM, Bouma WH, Bemelman WA. Surg Endosc. 2011 Jun 24.[Epub]

2. Laparoscopy in Combination with Fast Track Multimodal Management is the Best Perioperative Strategy in Patients Undergoing Colonic Surgery: A Randomized Clinical Trial (LAFA-study).

Vlug MS, Wind J, Hollmann MW, Ubbink DT, Cense HA, Engel AF, Gerhards MF, van Wagensveld BA, van der Zaag ES, van Geloven AA, Sprangers MA, Cuesta MA, Bemelman WA; on behalf of the collaborative LAFA study group. Ann Surg. 2011 May 18.

3. Categorization of occult tumour cells in lymph nodes in patients with colon cancer not reliable enough.

van der Zaag ES, Welling L, Peters HM, van de Vijver MJ, Bemelman WA, Buskens CJ. Ned Tijdschr Geneeskd. 2011;155(7):A2697.

4. The persisting presacral sinus after anastomotic leakage following anterior resection or restorative proctocolectomy.

van Koperen PJ, van der Zaag ES, Omloo JM, Slors JF, Bemelman WA. Colorectal Dis. 2011 Jan;13(1).

5. A multicenter randomized clinical trial investigating the cost-effectiveness of treatment strategies with or without antibiotics for uncomplicated acute diverticulitis (DIABOLO trial).

Unlü C, de Korte N, Daniels L, Consten EC, Cuesta MA, Gerhards MF, van Geloven AA, van der Zaag ES, van der Hoeven JA, Klicks R, Cense HA, Roumen RM, Eijsbouts QA, Lange JF, Fockens P, de Borgie CA, Bemelman WA, Reitsma JB, Stockmann HB, Vrouenraets BC, Boermeester MA; Dutch Diverticular Disease 3D Collaborative Study Group. BMC Surg. 2010 Jul 20;10:23.

6. Diagnosing occult tumour cells and their predictive value in sentinel nodes of histologically negative patients with colorectal cancer.

van der Zaag ES, Kooij N, van de Vijver MJ, Bemelman WA, Peters HM, Buskens CJ. Eur J Surg Oncol. 2010 Apr;36(4):350-7.

7. Improving staging accuracy in colon and rectal cancer by sentinel lymph node mapping: a comparative study.

van der Zaag ES, Buskens CJ, Kooij N, Akol H, Peters HM, Bouma WH, Bemelman WA. Eur J Surg Oncol. 2009 Oct;35(10):1065-70.

8. Systematic review of laparoscopic vs open colonic surgery within an enhanced recovery programme.

Vlug MS, Wind J, van der Zaag ES, Ubbink DT, Cense HA, Bemelman WA. Colorectal Dis. 2009 May;11(4):335-43.

144 145

list of publications list of publications

9. Laparoscopic ileocolic resection versus infliximab treatment of distal ileitis in Crohn’s disease: a randomized multicenter trial

Eshuis EJ, Bemelman WA, van Bodegraven AA, Sprangers MA, Bossuyt PM, van Milligen de Wit AW, Crolla RM, Cahen DL, Oostenbrug LE, Sosef MN, Voorburg AM,

Davids PH, van der Woude CJ, Lange J, Mallant RC, Boom MJ, Lieverse RJ, van der Zaag ES, Houben MH, Vecht J, Pierik RE, van Ditzhuijsen TJ, Prins HA, Marsman WA, Stockmann HB, Brink MA, Consten EC, van der Werf SD, Marinelli AW, Jansen JM, Gerhards MF, Bolwerk CJ, Stassen LP, Spanier BW, Bilgen EJ, van Berkel AM, Cense HA, van Heukelem HA, van de Laar A, Slot WB, Eijsbouts QA, van Ooteghem NA, van Wagensveld B, van den Brande JM, van Geloven AA, Bruin KF, Maring JK, Oldenburg B, van Hillegersberg R, de Jong DJ, Bleichrodt R, van der Peet DL, Dekkers PE, Goei TH, Stokkers PC. BMC Surg. 2010 Jul 20;10:23.

10. Angioplasty or bypass for superficial femoral artery disease? A randomised controlled trial.

van der Zaag ES, Legemate DA, Prins MH, Reekers JA, Jacobs MJ. Eur J Vasc Endovasc Surg. 2004 Aug;28(2):132-7.

11. Prevention of occlusions after infra-inguinal bypass surgery with oral anticoagulants or acetylsalicylic acid; a randomized comparison

van der Zaag ES, Legemate DA. Ned Tijdschr Geneeskd. 2001 Sep 8;145(36):1765.12. Aortoiliac reconstructive surgery based upon the results of duplex scanning. van der Zaag ES, Legemate DA, Nguyen T, Balm R, Jacobs MJ. Eur J Vasc & Endovasc

Surg. 1998 Nov;16(5):383-9.13. Treatment of intermittent claudication; prospective randomized study in the

BAESIC-Trial (bypass, angioplasty or endarterectomy patients with severe inter mittent claudication)

van der Zaag ES, Prins MH, Jacobs MJ. Ned Tijds Geneeskd. 1996 Apr 6;140(14):787-8.14. Effects of bile salt and phospholipid hydrophobicity on lithogenicity of human

gallbladder bile. van Erpecum KJ, Portincasa P, Stolk MF, van de Heijning BJ, van der Zaag ES, van

den Broek AM, van Berge Henegouwen GP, Renooij W. Eur J Clin Invest. 1994 Nov;24(11): 744-50.

147

dankwoord

Dankwoord

Dit proefschrift is ontstaan door een uitstekende samenwerking tussen vele afdelingen die alle op een eigen manier meewerkten aan de behandeling van patiënten met darmkanker. Maar eerst wil ik de patiënten bedanken die belangeloos aan het onderzoek hebben meegewerkt.

Mijn promotor prof. dr. W.A. Bemelman, beste Willem, ik ben ongelofelijk blij dat jij mij hebt begeleid bij dit proefschrift. Ik zie ons nog om de tafel zitten terwijl wij dit plan bespraken en zie hier nu al het resultaat! Ik ben erg onder de indruk van je enthousiasme. Tot mijn verbazing kreeg ik steeds per kerende post mijn artikelen terug die je tot diep in de nacht had voorzien van nuttige toevoegingen. Ik hoop dat wij deze (wetenschappelijk) band tussen Amsterdam en Apeldoorn nog lang in stand kunnen houden!

Mijn copromotor dr. W.H. Bouma, beste Wim, in 2001 kwam ik in Apeldoorn voor de laatste drie jaren van mijn opleiding tot chirurg. Altijd heb ik met veel plezier met jou samengewerkt en veel van je geleerd. Daarom was ik erg blij dat ik in 2006 terug kon komen en mocht toetreden tot een hechte maatschap. Ik ben je erg dankbaar dat jij dit promotieonderzoek voor mij mogelijk hebt gemaakt.

Tsja, en toen was daar dr. C.J. Buskens, beste Christianne, Chris. Wat moet ik zeggen, zonder jou was dit proefschrift er nooit gekomen! Ik ben onder de indruk van jouw tomeloze inzet en enthousiasme. Het ene wetenschappelijke idee was nog niet geboren of er kwam alweer een volgende, niet bij te houden. Ik heb erg veel geleerd van jouw doorzettingsvermogen en inzet. Jij bent nu een zeer veelbelovend staflid in het AMC geworden maar gelukkig houdt het hier niet op. Ik kijk uit naar onze toekomstige samenwerking want jij zit nog steeds vol met goede ideeën.

De overige leden van de commissie wil ik bedanken dat zij zitting wilden nemen in mijn promotiecommissie; prof. dr. J.F. Lange, prof. dr. J.P. Medema, prof. dr. D.J. Richel, dr. P.J. Tanis en prof dr. M.J. van de Vijver.

De afdelingen pathologie in het Gelre Ziekenhuis en in het AMC waren onmisbaar voor dit onderzoek. Herman Peters en Evert Weltevreden heel veel dank voor jullie support. Nina Kooij en Janny van Benthem dank voor jullie inzet bij alle kleuringen. Marc van de Vijver dank dat jij alle coupes wilde reviseren.

Dank aan alle co-auteurs en medewerkers van de verschillende artikelen: Halil Akol, Sanne Bartels, Evelien Dekker, Kemal Dogan, Dirk Ubbink, Malaika Vlug en Lieke Welling.

148 149

dankwoord dankwoord

Dé maatschap chirurgie Apeldoorn, dank voor jullie steun op velerlei gebied! Wij moeten maar snel weer eens met z’n allen gaan fietsen of wa(d)t dan ook.

De verpleegafdelingen chirurgie en met name afdeling A6, dames van de polikliniek chirurgie en al het personeel van het operatiekamercomplex dank voor jullie vak kundige inzet bij alle (colorectale) patiënten.

Assistenten chirurgie ook jullie wil ik bedanken voor jullie inzet dag en nacht! En dank voor alle handige tips & trics hoe je nou zo’n boekje eigenlijk in elkaar moet zetten. Ik hoop dat ik jullie nog vaak iets één keertje mag voordoen.

Mijn paranimfen, Hessel Buscher en Peter van Duijvendijk, dank voor jullie hulp. Het wordt een groot feest. Ik kijk uit naar de jaren dat wij met heel veel plezier met elkaar zullen samenwerken!

Mijn ouders, eindelijk na negen jaar weer een promotie in de familie! Jullie hebben mij altijd de vrijheid gegeven en gestimuleerd mijn eigen weg te gaan. En kijk eens tot hoever die weg kan komen, dank voor jullie steun en kritische noten.

José Loonen, Lieve José, ik ben jouw voor twee dingen ongelofelijk dankbaar. Dat jij “ja” zei toen ik naar de hand van jouw dochter vroeg en de onvoorwaardelijke steun die jij ons jonge gezin geeft! Zonder jou was dit allemaal niet mogelijk geweest, al die avonden dat jij de kinderen liefdevol opving en naar bed bracht en het eten klaar stond als wij thuis kwamen zijn onvergetelijk!

Sophie, Pelle en Annemein het lap-tobben is afgelopen nu de spreekbeurt nog en dan is het echt klaar! Wat ben ik een gelukkige vader als ik ieder dag thuis kom en weer struikel over jullie enthousiasme, geluk, speelgoed, pop, driewieler en terecht kom in jullie onvoorwaardelijke liefde. Ik hoop dat ik nog heel vaak mag struikelen.

En dan mijn allerliefste Hester, ook de prinses op de witte fiets bestaat! De meeste dromen, zijn géén bedrog...

151

curriculum vitae

Curriculum vitae

On a cloudy morning, Spring 1968, at 6 a.m., Edwin van der Zaag was born in Amsterdam. From that moment on his days would always start at 6 a.m., leading to his nickname “de wekker”. He was born a stone’s throw from the Jaap Eden ice-skating rink and football stadium de Meer, which explains his desperation to become a speedskater and an Ajax fan for life. The family moved to a village near the city of light Eindhoven, and he spent much of his time on his bicycle cycling to the Strabrechtse heide or the Achelse kluis. Also, he tried to go as fast as he could on his speedskates, closely followed by Leen Pfrommer. In his spare time he attended the Athenaeum at the Eindhovens Protestants Lyceum. In 1987 he continued his efforts to become a speedskater at the Vechtsebanen and picked-up his field-hockey career at Kampong. Because of all these activities his appetite was insatiable, which explained his second nickname “Kliko”. He received his medical degree in 1994 at the University of Utrecht and eventually started his residency in Surgery in 1998 at the department of Surgery of the Academic Medical Centre in Amsterdam (prof. dr. H. Obertop and prof. dr. D.J. Gouma). For six months he worked at the department of Surgery at the Antonie van Leeuwenhoek hospital in Amsterdam (prof. dr. B. Kroon), which made him realise he wanted to be involved in oncological care. In 2002 he married Hester in the most beautiful city of the Netherlands, and soon a girl with golden hair, Sophie, was born. In 2004 he finished his residency at the Gelre hospitals Apeldoorn where dr. W.H. Bouma and dr. E.J. Hesselink showed him all ins and outs of gastro-intestinal surgery. In 2005 the young family exchanged the Dom for the Martini tower and he specialised in laparoscopic and colorectal surgery as a fellow at the department of Surgery of the Medical Centre Leeuwarden (dr. W.J.H.J. Meijerink) and the University Medical Centre Groningen (prof. dr. R.J. Ploeg). Their son Pelle, with his contagious laugh, was born in this period. In 2006 he was happy to join the department of Surgery in Apeldoorn. They found the ”Lindenhof” in the picturesque village of Twello where Annemein, the girl with sparkling blue eyes, was born. And they lived happily ever after……..

UvA-DARE (Digital Academic Repository) Surgery for colorectal … · patients with CRC. In Chapter...

Documents

Transcript of UvA-DARE (Digital Academic Repository) Surgery for colorectal … · patients with CRC. In Chapter...