Research Article Prognostic Impact of mRNA Expression ...Research Article Prognostic Impact of mRNA...

Research ArticlePrognostic Impact of mRNA Expression Levels ofHER1ndash4 (ERBB1ndash4) in Patients with Locally AdvancedRectal Cancer

Melanie Kripp1 Kirsten Merx1 Ralph Markus Wirtz2 Timo Gaiser3

Sebastian Eidt2 Juliana Schwaab1 Stefan Post4 Frederik Wenz5 Andreas Hochhaus6

Ralf-Dieter Hofheinz1 and Philipp Erben7

1 III Medizinische Klinik Universitatsmedizin Mannheim 68167 Mannheim Germany2Stratifyer Molecular Pathology GmbH 50935 Koln Germany3Pathologisches Institut Universitatsmedizin Mannheim 68167 Mannheim Germany4Chirurgische Klinik Universitatsmedizin Mannheim 68167 Mannheim Germany5Klinik fur Strahlentherapie und Radioonkologie Universitatsmedizin Mannheim 68167 Mannheim Germany6Abteilung HamatologieOnkologie Universitatsklinikum Jena 07747 Jena Germany7Klinik fur Urologie Universitatsmedizin Mannheim 68167 Mannheim Germany

Correspondence should be addressed to Philipp Erben philipperbenmedmauni-heidelbergde

Received 27 May 2016 Accepted 19 July 2016

Academic Editor Haruhiko Sugimura

Copyright copy 2016 Melanie Kripp et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Background No predictive or prognostic biomarker is available for patients with locally advanced rectal cancer (LARC) undergoingperioperative chemoradiotherapy (CRT) Members of the human epidermal growth factor receptor (HER) family of receptortyrosine kinases EGFR (HER1 ERBB1) HER2 (ERBB2) HER3 (ERBB3) and HER4 (ERBB4) are therapeutic targets in severalcancers The analysis was performed to assess expression levels and study the potential prognostic impact for disease-free andoverall survival in patients with LARC Patients and Methods ERBB1ndash4 mRNA expression and tumor proliferation using Ki-67(MKI67) mRNA were evaluated using RT-quantitative PCR in paraffin-embedded tumor samples from 86 patients (median age63) treated with capecitabine or 5-fluorouracil-based CRTwithin a phase 3 clinical trialResults A positive correlation of HER4 andHER2HER3 andHER2 andHER4 andHER3with each otherwas observed Patientswith highmRNAexpression ofERBB1 (EGFRHER1) had significantly increased risk for recurrence and death Patients with high mRNA expression ofMKI67 had reduced riskfor relapse Conclusion This analysis suggests a prognostic impact of both ERBB1 and MKi67 mRNA expression in LARC patientstreated with capecitabine or fluorouracil-based chemoradiotherapy

1 Introduction

Rectal cancer is the fifth most common type of cancer inadults worldwide [1] Although surgery may be curative inlocally advanced disease local recurrence and metastasesoccur despite complete resection Until the late 1980s therate of local recurrence and distant metastases followingcurative surgery was about 30 [2] The outcome of locallyadvanced rectal cancer (LARC) has significantly improveddue to the combination of optimized surgical techniquesnotably total mesorectal excision (TME) with neoadjuvant

radio- and radiochemotherapy [3ndash5] By using this treatmentmodality 10-year cumulative local recurrence rate is generallybelow 10 On the other hand fluorouracil in conjunc-tion with neoadjuvant long-term radiotherapy reduces localrecurrences but does not prolong overall survival (OS) [6 7]

Nowadays distant metastases represent the most com-mon type of treatment failure in rectal cancer indicatingthe need for optimized systemic medical treatment such asmodifications of perioperative bolus 5-fluorouracil (5-FU)treatment However neither biomodulation of fluorouracilnor combinations with older cytostatic drugs have clearly

Hindawi Publishing CorporationGastroenterology Research and PracticeVolume 2016 Article ID 3481578 9 pageshttpdxdoiorg10115520163481578

2 Gastroenterology Research and Practice

proved advantage compared to bolus 5-FU [8] Only admin-istering 5-FU as continuous infusion during radiation led toimproved survival and increased time to relapse [9] Likewisethe addition of oxaliplatin to perioperative fluorouracil treat-ment has given diverging results in five clinical trials [10 11]

Altogether the identification of patients being at high riskfor distant metastases still represents a major challenge totailor the management of rectal cancer therapy For patientswith resected rectal cancermdashin addition to the adequacy ofsurgical excision (evaluated by the circumferential resectionmargins)mdashthe TNM classification is still the most reliableindicator of risk for systemic recurrence [12] However forpatients scheduled to undergo neoadjuvant chemoradiother-apy clinical stratification (eg withMRI scan) has limitationsespecially in view of the inability to exactly predict the nodalstatus Risk stratification based on molecular markers couldprovide better estimate of individual risk and tailored treat-ment In this regard especiallymolecularmarkers bearing thepotential to serve as therapeutic targets formedical treatmentare indispensable

Several drugs are licensed or in clinical evaluation forthe treatment of tumors expressing human epidermal growthfactor tyrosine kinase receptors (EGFR) Besides EGFRwhich is encoded by the ERBB1 gene the human epidermalgrowth factor receptor (HER) family of receptor tyrosinekinases includes three additional members HER2 (ERBB2)HER3 (ERBB3) and HER4 (ERBB4) Activation of the HERsignaling network has been shown to promote tumor inva-sion and metastasis both in vitro [13ndash16] and in vivo [17 18]

Mutations and overexpression of HER family genes arefrequently present in rectal cancer For instance a higher rateof HER2 amplification in high grade tumors was reportedand EGFR and HER3 mRNA expression was described to beassociated with the occurrence of metastases in patients withLARC [19]

Here we sought to evaluate (a) the mRNA expression ofERBB1ndash4 in pretreatment tumor tissue (b) the correlationof expression of each receptor along with tumor cell pro-liferation using MKI67 mRNA expression in patients withLARC treated with 5-FU-based chemoradiotherapy and (c)the impact of these markers on prognosis

2 Material and Methods

Patient Cohort Tumor tissue for this study stemmed frompatients participating in a phase III clinical trial conductedat the University Hospital of Mannheim between 2002 and2007 These patients had histologically confirmed LARC(adenocarcinoma cT3-4 any N or cT2 N+) This nonin-feriority trial compared 5-FU with the oral 5-FU prodrugcapecitabine for the perioperative treatment of LARCDetailsof the study protocol and the results have been publishedpreviously [20] In brief the trial began in 2002 as an adju-vant trial comparing capecitabine-based chemoradiotherapywith fluorouracil-based chemoradiotherapy Patients in thecapecitabine group were scheduled to receive two cycles ofcapecitabine followed by chemoradiotherapy (504Gy pluscapecitabine 1650mgm2) then three cycles of capecitabine

Patients in the fluorouracil group received two cycles ofbolus fluorouracil followed by chemoradiotherapy (504Gyplus infusional fluorouracil 225mgm2 daily) then two cyclesof bolus fluorouracil The protocol was amended in 2005to allow a neoadjuvant cohort in which patients in thecapecitabine group received chemoradiotherapy followed byTME surgery and five cycles of capecitabine and patients inthe fluorouracil group received chemoradiotherapy (504Gyplus infusional fluorouracil 1000mgm2 days 1ndash5 and 29ndash33)followed by surgery and four cycles of bolus fluorouracilPatients were randomly assigned to treatment groups ina 1 1 ratio The primary endpoint was overall survivalNoninferiority of capecitabine in terms of 5-year overallsurvival was tested 392 patients were evaluable with amedianfollow-up of 52 months Five-year overall survival in thecapecitabine group was noninferior to that in the fluorouracilgroup (76 [95CI 67ndash82] versus 67 [58ndash74] 119901 = 00004)Similar numbers of patients had local recurrences in eachgroup 12 [6] in the capecitabine group versus 14 [7] in thefluorouracil group 119901 = 067

All patients providing tumor tissue for the current anal-ysis were treated at a single center (University HospitalMannheim University of Heidelberg)

Both the clinical study protocol and the molecular inves-tigations reported here were approved by the local ethicscommittee

21 RNA Isolation from Formalin-Fixed Paraffin-Embedded(FFPE) Tissue and Quantitative Reverse Transcription-Poly-merase Chain Reaction (RT-qPCR) Assessment Fixation oftumor specimens followed standard protocols 10 bufferedformalin for at least 8 hours Hematoxylin-eosin (HE) stainedsections were evaluated for pathological stage accordingto the 2002 TNM classification of the American JointCommittee on Cancer classification For molecular analy-ses HE slides were reevaluated by an experienced pathol-ogist (SE) for confirmation of the presence of invasivetumor Each case was macrodissected for an invasive tumorareal comprising at least 30 of tumor cells One 10 120583msection was used for the isolation of RNA accordingto a fully automated high-throughput extraction work-flow which runs on an Xtract-XL liquid-handling robot(STRATIFYER Molecular Pathology GmbH Cologne Ger-many) The extraction solutions and chemicals are com-mercially available in Germany as part of the XTRAKTFFPE kit which is based on magnetic bead technology(STRATIFYER) [21] In brief FFPE sections were solubilizedand paraffin was melted by incubating with a lysis buffer ina Thermo-mixer Tissue was lyzed with Proteinase K Thelysates were then admixed with germanium-coated mag-netic particles in buffer-controlled conditions which enhancepreferential attachment of nucleic acid molecules to thesurface of the particles Purification was carried out bymeans of 3 consecutive washing cycles involving magneti-zation centrifugation washing and removal of the supernatant Nucleic acids were eluted with 100 120583L elution bufferand treated with DNase I The DNA-free RNA eluates werestored at minus80∘C until use

Gastroenterology Research and Practice 3

Table 1 Primerprobe sets used for amplification of the target and reference genes

Gene Oligo ID Oligo sequences 51015840-31015840 Label probe Ampliconlength (nts) Assay location PCR efficacy in

MKI67Forward CGAGACGCCTGGTTACTATCAA

CY5-BHQ2 108 c ex 2-3NM 002417 98Reverse GGATACGGATGTCACATTCAATACC

Probe ACGGTCCCCACTTTCCCCTGAGC

CALM2Forward GAGCGAGCTGAGTGGTTGTG

YY-BHQ1 72 c ex 1-2NM 001743 993Reverse AGTCAGTTGGTCAGCCATGCT

Probe TCGCGTCTCGGAAACCGGTAGC

EGFR(ERBB1)

Forward CGCAAGTGTAAGAAGTGCGAAFAM-BHQ1 93 NM 201283 891Reverse CGTAGCATTTATGGAGAGTGAGTCT

Probe CCTTGCCGCAAAGTGTGTAACGGAAT

ERBB2(HER2)

Forward TCTGGACGTGCCAGTGTGAACY5-BHQ2 61 M11730 gen 972Reverse CCTGCTCCCTGAGGACACAT

Probe AGGCCAAGTCCGCAGAAGCCCT

ERBB3(HER3)

Forward CGGTTATGTCATGCCAGATACACROX-BHQ2 81 NM 001982 96Reverse GAACTGAGACCCACTGAAGAAAGG

Probe CTCAAAGGTACTCCCTCCTCCCGGG

ERBB4(HER4)

Forward GAGGCTGCTCAGGACCTAAGGATTO-BHQ1 75 NM 005235 92Reverse GAGTAACACATGCTCCACTGTCATT

Probe CACAGACTGCTTTGCCTGCATGAATTTCCY5 = cyanine dye 5 BHQ = black hole quencher FAM = 6-carboxyfluorescein ROX = carboxy-X-rhodamine ATTO = fluorescent dye (ATTO-TEC GmbHSiegen Germany)

One-step RT-qPCR was applied for the relative quantifi-cation of EGFR (ERBB1) ERBB2 (HER2) ERBB3 (HER3)ERBB4 (HER4) andMKI67mRNAexpression by using gene-specific TaqMan-based assays as described [22] Expressionlevels of the target genes as well as of the reference geneCalmodulin 2 (CALM2) were assessed in duplicate by RT-qPCR using the SuperScript III PLATINUMOne-Step quan-titative RT-PCR System (Invitrogen Karlsruhe Germany)on a Stratagene Mx3005p (Agilent Technologies BoblingenGermany) with 30min at 50∘C and 2min at 95∘C followed by40 cycles of 15 sec at 95∘C and 30 sec at 60∘C

Forty cycles of nucleic acid amplification were appliedand the cycle threshold (Cq) values of the target genes wereidentified Cq values were normalized by subtracting theCq value of the housekeeping gene CALM2 from the Cqvalue of the target genes (ΔCq) RNA results were thenreported as 40-ΔCq values which correlate proportionallywith the mRNA expression level of the target genes Thequantity of RNA following isolation (yield) was determinedby measuring CALM2 expression as a surrogate marker foramplifiable mRNA Samples with average CALM2 Cq valueslt32 were considered to have sufficient RNA and were eligiblefor further analysis Therefore 5 of the 106 extracted samples(success rate 95) had an average CALM2 CT value ofge32 and were therefore excluded CALM2 were selected ascontrol gene based on molecular studies in breast cancerpatients which showed a high stability of this gene [23] Thelengths of the amplicons detected by the EGFR (ERBB1)ERBB2 ERBB3 ERBB4 andMKI67 assays were 93 bp 61 bp81 bp 75 bp and 108 bp respectively with PCR efficiencies

[119864 = 1(10-slope)] of 891 972 960 920 and 997respectively A commercially available human reference RNA(Stratagene qPCR Human Reference Total RNA AgilentTechnologies Waldbronn Germany) was used as positivecontrol No-template controls were assessed in parallel toexclude contamination

Details of the primerprobe sets used for amplification ofthe target and reference genes are listed in Table 1

22 Statistical Analysis Overall survival (OS) was measuredfrom the date of randomization until death from any causeSurviving patients were censored at the date of last contactDisease-free survival (DFS) was measured from the date ofrandomization until recurrence of tumor secondary neo-plasm or death from any cause as described [20] Time-to-event distributionswere estimated usingKaplan-Meier analy-ses Continuous variables were presented asmedians with thecorresponding range and categorical variables as frequencieswith the respective percentages Associations ofmarker geneswith basic patient and tumor characteristics were examinedusing Fisherrsquos exact test for categorical variables and Mann-Whitney test for continuous variables Correlations betweenthe target genes were calculated using Spearmanrsquos rank corre-lation coefficient (Rho) Univariate Cox regression analyseswere performed to assess the relationship between markersandOS orDFSThe cut-offswith the highest predictive valuesfor OS and DFS were estimated using the partitioning testAll 119901 values were two-sided with 119901 values lt 005 indicatingstatistical significance Statistical analyses were performedwith SAS Jmp 100 (SAS Institute Cary NC USA) and Graph


mRNA expression detectableEGFR (ERBB1) n = 49 HER2 (ERBB2) n = 49HER3 (ERBB3) n = 49HER4 (ERBB4) n = 49

MKI67 n = 48

mRNA expression detectableEGFR (ERBB1) n = 52HER2 (ERBB2) n = 52HER3 (ERBB3) n = 52HER4 (ERBB4) n = 52

MKI67 n = 52

Assesed for eligibility n = 101 ptsAdjuvant treatment n = 62pts

Neoadjuvant treatment n = 39pts

FFPE tumor tissue sampleswere available

Included in study n = 86pts

Adjuvant treatment n = 56ptsNeoadjuvant treatment n = 30pts

mRNA expression in resected specimen

Adjuvant treatment n = 54pts

mRNA expression in pretreatment biopsies(n = 52 DFS OS analysis)


Excluded n = 15 ptsWithdrawl informed consent n = 1pts

No adequate material neoadjuvant n = 9ptsNo adequate material adjuvant n = 5pts

Figure 1 Participants and samples flow chart diagram

Pad Prism software (Version 5 La Jolla CA USA) Withregard to OS and DFS patient samples with pretherapeuticbiopsies (119899 = 55) were analyzed in order to exclude a possiblesampling error

3 Results

31 Patientsrsquo Characteristics and mRNA Expression Tumortissue samples were obtained from 86 patients (median age63 years range 44ndash83 years) A total of 52 biopsies and54 resection specimens from these patients were analyzedas described in the ldquoMaterial and Methodsrdquo Only biopsiesderiving prior to treatment were analyzed with regard toOS and DFS Regarding these biopsies 3052 biopsies werederived from the additional neoadjuvant treated patientgroup and 2252 biopsies from the solely adjuvant treatedpatients Resection specimens were exclusively investigatedin patients receiving solely adjuvant chemoradiotherapy toanalyze a possible sample effect of biopsies and resectedspecimens (119899 = 54) (Figure 1) Basic clinical and pathologicalcharacteristics are shown in Table 2 and the samples flowchart is described in Figure 1 Analyzing normalized mRNAexpression values of evaluated HER (ERBB) family markergenes and MKI67 lowest expression levels were observedfor ERBB3 (Median 33) and ERBB4 (Median 327) andhighest expression levels for ERBB2 (Median 359) andMKI67 (Median 356 Figure 2) using the two-way Mann-Whitney test Measured Median Cq values of analyzed geneswere between 274 (ERBB3 (283 ERBB2 307 MKI67 321EGFR)) and 326 (ERBB4) Using Spearmanrsquos correlations

Table 2 Patientsrsquo characteristics

Neoadjuvanttreatment

Adjuvanttreatment

119873 (patients) 30 56Age (years) median[range]

67[47ndash83]

62[44ndash75]

Gender 119899 ()Male 20 (67) 36 (64)Female 10 (33) 20 (36)

5-Fluorouracil 119899 () 15 (50) 27 (48)Capecitabine 119899 () 15 (50) 29 (52)Recurrence during follow-up119899 ()Local recurrence 1 (3) 3 (5)Metastasis 9 (30) 15 (27)

between the analyzed genes weak negative correlations werefound between EGFR (ERBB1) and HER2 (ERBB2 Rho 018119901 = 022) and moderate to high positive correlationsbetween HER4 and ERBB2 (HER2) ERBB3 (HER3) andERBB2 (HER2) and ERBB4 (HER4) and ERBB3 (HER3 Rhoranges from 044 to 064 119901 lt 001 in all cases)

32 Association of Gene Expression with Patient and TumorCharacteristics All biopsies (119899 = 52) irrespective of treat-ment modality showed no association between normalizedmRNA expression and age (le65 versus gt65 years) or gender


Table 3 Comparison of normalized gene expression values (40-ΔCq) in biopsies versus resected specimen in the adjuvant treatment groupSignificant lower expression in resected specimen was observed for HER2ndashHER4 and MKI67 (ns not significant)

GeneBiopsy (119899 = 22) Resected specimen (119899 = 54)

119901 valueGene expressionmedian (range)

Gene expressionmedian (range)

EGFR (ERBB1 HER1) 348 (329ndash361) 346 (332ndash361) nsHER2 (ERBB2) 360 (346ndash372) 354 (343ndash365) 0039HER3 (ERBB3) 331 (319ndash345) 321 (306ndash334) lt0001HER4 (ERBB4) 329 (298ndash345) 318 (291ndash343) lt0001MKI67 354 (335ndash368) 351 (304ndash364) 002

However higher EGFR (ERBB1) expression levels were asso-ciated with the risk for developing recurrence (median 352versus 347 119901 = 0036) or death during follow-up (median351 versus 347 119901 = 0038) Furthermore a trend wasobserved for higher HER3 expression levels among patientsstill alive during follow-up (median 332 versus 324 119901 =0083) No significant association of normalized gene expres-sionwas observed in relation to different treatment groups (5-fluorouracil versus capecitabine) Comparison of normalizedgene expression in biopsies versus resected specimen inthe adjuvant treatment group showed that all markers withthe exception of EGFR (ERBB1) displayed higher expressionlevels in preoperative biopsies than in resection specimens(Table 3) Therefore solely pretreatment biopsies were ana-lyzed with regard to DFS and OS

33 Association of Gene Expression with Disease-Free andOverall Survival For the EGFR (ERBB1) mRNA values theevaluated cut-off (18 high expression) was prognostic forboth OS and DFS while for the ERBB2 ERBB3 and MKI67no significant prognostic effect was found in the examinedcut-offs in terms of OS and DFS Concerning EGFR (ERBB1)58 deaths (63) and 78 relapses (88 6 distant metastasis)occurred in patients with high-expressing tumors comparedto 841 deaths (20) and 1341 relapses (31 8 distantmetastasis) in the low-expressing tumors Patients with highmRNA expression of EGFR (ERBB1) had increased risk fordeath (HR = 475 95 CI 151 to 1448 119901 = 00090) andincreased risk for relapse (HR = 704 95CI 26 to 1884119901 =00003) compared to patients with low-expressing tumorsKaplan-Meier curves forOS andDFS according to themRNAstatus of EGFR (ERBB1) are displayed in Figures 3(a) and 3(b)

For ERBB2 (HER2)mRNA expression (38 high expres-sion) a trend for a prognostic association was observed forOS (119901 = 0064) and no correlation for DFS Furthermorepatients with high MKI67 mRNA expression (82 highexpression) showed a trend for a reduced DFS using theKaplan-Meier method (119901 = 0075 Figures 3(c) 3(d) and3(f)) Concerning ERBB2 (HER2) 219 deaths (11) occurredin patients with high-expressing tumors in comparison to1130 deaths (37) in the low-expressing tumors Patientswith high mRNA expression of ERBB2 (HER2) had reducedrisk for death (HR = 027 95 CI 0041 to 099 119901 =0048) in the proportional hazardmodel compared to patientswith high-expressing tumors Kaplan-Meier curves for OS

EGFR

(ERB

B1)

ERBB

2 (H

ER2)

ERBB

3 (H

ER3)

ERBB

4 (H

ER4)

MKI

67

30

35

40

n

lowastlowastlowastlowast



49494949 48Nor

mal

ized

mRN

A ex

pres

sion

(40

-ΔC

q)

Figure 2 Distribution of mRNA expression values in analyzedbiopsies Normalized mRNA expression values (40-ΔCq) of RT-qPCR evaluated HER (ERBB) family marker genes and MKI67are presented Lowest expression levels were observed for ERBB3and ERBB4 and highest expression for ERBB2 and MKI67 (boxand whiskers plots ranging from minimum to maximum) Geneexpression values were compared by two-way Mann-Whitney test(lowastlowastlowastlowast119901 lt 00001)

according to the mRNA status of ERBB2 (HER2) are shownin Figure 3(c)

With regard to MKI67 1339 relapses (33) occurredin patients with high-expressing tumors in comparison to69 relapses (67) in the low-expressing tumors Patientswith high mRNA expression of MKI67 had reduced risk forrelapse (HR = 032 95 CI 013 to 088 119901 = 0029) inthe proportional hazard model compared to patients withlow-expressing tumors when adjusting for treatment groupKaplan-Meier curves forDFS andOS according to themRNAstatus ofMKI67 are shown in Figures 3(e) and 3(f)

4 Discussion

No molecular markers have been described thus far toidentify patients with LARC carrying a high risk for distant


Months since randomisation0 12 24 36 48 60 72 84 96 108 120

0

20

40

60

80

100

EGFR (ERBB1) high60 months survival 23

EGFR (ERBB1) low 60 months survival 76

p = 00057

Perc

enta

ge al

ive

(a)

0

20

40

60

80

100


EGFR (ERBB1) high60 months disease free 13

EGFR (ERBB1) low 60 months disease free 71

p lt 00001

Perc

enta

ge d

iseas

e fre

e

(b)


HER2 (ERBB2) low60 months survival 59

HER2 (ERBB2) high60 months survival 89

p = 00640

20

40

60

80

100

Perc

enta

ge al

ive

(c)

HER2 (ERBB2) low60 months disease free 56

HER2 (ERBB2) high60 months disease free 70

Months since randomisation0 12 24 36 48 60 72 84 96

p = ns0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(d)

MKI low60 months survival 51

MKI67 high 60 months survival 75

Months since randomisation0 12 24 36 48 60 72 84 96 108 120 132 144

p = ns0

20

40

60

80

100

Perc

enta

ge al

ive

(e)

p = 0075

MKI low60 months disease free 39

MKI67 high 60 months disease free 70


0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(f)

Figure 3 (a) Kaplan-Meier curve for overall survival (OS) according to the mRNA status of EGFR (ERBB1) The predictive values for EGFR(ERBB1) high and low expression were estimated using the partitioning test Five-year overall survival for low EGFR (ERBB1) expression was76 for high EGFR expression 60 119901 = 00057 (b) Kaplan-Meier curve for disease-free survival (DFS) according to the mRNA statusof EGFR (ERBB1) The predictive values for EGFR (ERBB1) high and low expression were estimated using the partitioning test Five-yeardisease-free survival for low EGFR (ERBB1) expression was 71 for high EGFR (ERBB1) expression 13 119901 lt 00001 (c) Kaplan-Meier curvefor overall survival (OS) according to the mRNA status of ERBB2 (HER2) The predictive values for ERBB2 (HER2) high and low expressionwere estimated using the partitioning test Five-year overall survival for low HER2 (ERBB2) expression was 59 for high HER2 (ERBB2)expression 89 119901 = 0064 (d) Kaplan-Meier curve for disease-free survival (DFS) according to the mRNA status of ERBB2 (HER2) Thepredictive values for ERBB2 (HER2) high and low expression were estimated using the partitioning test Five-year overall survival for lowHER2 (ERBB2) expression was 56 for high HER2 (ERBB2) expression 70 119901 = ns (e) Kaplan-Meier curve for overall survival (OS)according to the mRNA status ofMKI67The predictive values forMKI67 high and low expression were estimated using the partitioning testFive-year overall survival for lowMKI67 expression was 75 for highMKI67 expression 55 119901 = ns (f) Kaplan-Meier curve for disease-free survival (DFS) according to the mRNA status ofMKI67 The predictive values forMKI67 high and low expression were estimated usingthe partitioning test Five-year overall survival for lowMKI67 expression was 39 for highMKi67 expression 70 119901 = 0075


metastases The aim of this study was to assess the potentialprognostic value of ERBB family of receptors (EGFR (ERBB1)ERBB2 (HER2) HER3 (ERBB3) and HER4 (ERBB4) expres-sion) as well as MKI67 expression in patients with LARCtreated at a single center with chemoradiotherapy based oncapecitabine or fluorouracil within a phase 3 clinical trialExpression of EGFR (ERBB1)was prognostic for both OS andDFS in this analysis

The mRNA expression of the four HER family mem-bers has previously been assessed from a series of 100locally advanced rectal cancers treated with radiotherapy orradiochemotherapy and showed an association for EGFR(ERBB1) andHER3 gene expression with development of dis-tant metastases in LARC [19]The current analysis confirmedthat higher EGFR (ERBB1) expression levels were foundmorefrequently among patients developing metastases Patientswith high mRNA expression of EGFR (ERBB1) also hadan increased risk for death compared to patients with low-expressing tumors In contrast ERBB3 expression was notprognostic in our analysis However both studies are notcompletely comparable as in the present study analysis forERBB1ndash4 and MKI67 was based on a standardized RNAextraction method and uniformly staged and treated (5-FU based chemoradiotherapy) patients were investigated inour series Moreover Ho-Pun-Cheung and colleagues usedtwo different RNA extraction protocols (TRIzol and QiagenColumns) These differences may explain the variabilityin the results Comparably both studies used pretreatmentspecimen and RNA based PCR expression EGFR (ERBB1)immunohistochemistry and RT-qPCR showed a good overallconcordance of 81 in breast cancer patients as describedrecently [21] This suggests related mRNA gene expressionand protein levels for EGFR (ERBB1) Pretreatment tumorbiopsies were also investigated in another study comprising40 patients with LARC [24] Responders to neoadjuvant 5-fluorouracil-based chemoradiotherapy (patients with signif-icant tumor regression) showed significantly lower EGFR(ERBB1) gene expression levels than nonresponders (patientswith insignificant tumor regression) The 3-year DFS ratesof the patients with lower gene expression levels of EGFR(ERBB1) were significantly higher (90) than those of thepatients with higher gene expression levels (70) (119901 =0003) These data are in line with the results of our currentinvestigation where patients with high mRNA expressionof EGFR (ERBB1) had increased risk for relapse and deathcompared to patients with low-expressing tumors treatedwith capecitabine or 5-FUThis provides evidence for the useof EGFR targeted therapies in patients with high expressionof EGFR treated with capecitabine or 5-FU

The predictive value of MKI67 was investigated in arectal cancer patient cohort treated with neoadjuvant 5-FUbased chemoradiotherapy [25] A small set of formalin-fixedparaffin-embedded pretreatment tumor biopsies and post-therapeutical resection specimens were studied by immuno-histochemistry The results were compared with histopatho-logical tumor regression according to a standardized semi-quantitative scoring system Responders (patients with hightumor regression) showed a significantly lower MKI67expression thannonresponders in the pretherapeutical tumor

biopsies (812 versus 167 119901 lt 005) as well as in theposttherapeutical resection specimens (758 versus 143119901 lt 001) Despite obvious differences regarding the aimsand methods of both investigations it is surprising thatthe results appear to be diametrically opposed Clearly thenumber of pretreatment tumor biopsies in the cited publi-cation was relatively small (119899 = 22) As described in theResults sampling methods (ie biopsies versus resection)may explain differences in gene expression levels Howeverthis does not explain the difference between both datasets In primary breast cancer for instance high MKI67mRNA expression measured by RT-qPCR is predictive forachieving pathological complete remission (pCR) to neoad-juvant chemotherapy [26] RT-qPCR was superior to MKi67determined by immunohistochemistry [27] Even thoughbreast cancer and rectal cancer have different biologies theapplied techniques immunohistochemistry versus mRNAexpression analysis by RT-qPCR may be in part responsiblefor the observed differences

With regard to HER2ERBB2 we found a trend for areduced risk of death for HER2 high-expressing tumorsThe role of HER2ERBB2 expression in colorectal canceris very controversial In a series of 1645 patients usingimmunohistochemistry and chromogenic in situ hybridiza-tion (CISH) with standard protocols an overall positivityrate of 16 in CRC patients was described [28] Contrarilyusing immunohistochemistry (IHC) scoring and detectionof silver in situ hybridization amplification (SISH) HER2status was determined in patients with rectal cancer (119899 =264) Tumors with an IHC score of 3 or SISH ratios of ge20were classified HER2 positive HER2 status was found to bepositive in 124 of patients with pretreatment biopsies Inthis analysis patients with HER2 positivity showed a trendfor better DFS and a significant benefit in cancer-specificsurvival Five-year survival rate was 960 for patients withHER2 positive tumors (versus 800 for HER2 negativetumors) [29] Recently an Italian group has proposed specificcriteria for HER2 positivity in colorectal cancer [30] Thesame group has evaluated anti-HER2 treatment (lapatinib +trastuzumab) within the so-called HERACLES trial in HER2positive colorectal cancer patients as last line treatment andfound significant antitumor activity [31] Taken togetherour results support the findings of Conradi and coworkerspostulating a positive prognostic effect of HER2 gene expres-sion However further investigations regarding the optimalmethod of defining HER2 positivity are needed particularlysince targeting HER2 has resulted in excellent efficacy dataeven in heavily pretreated patients in the HERACLES trial

With the exception of EGFR (ERBB1) normalized geneexpression was higher in biopsies than in resection spec-imens This difference could be a result of the effect ofischemia Surgical procedures significantly affect the expres-sion of genes in colorectal cancer tissue seen as a significantdifference in the molecular composition of tissue speci-mens collected after tumor resection compared to specimenscollected via colonoscopy before tumor resection [32] Inthe present analysis normalized values were also affectedThis clearly implies that ischemia alters gene expression ina gene-specific manner This was also shown by Liu and


colleagues using microarray and PCR techniques in kidneycancer [33] which clearly underlines the importance of strictstandardization and documentation of preanalytical factorsTherefore only pretreatment biopsies were evaluated in thisstudy to exclude an influence of sampling method on geneexpression

5 Conclusion

The key findings of our analysis of ERBB family in patientswith locally advanced rectal cancer undergoing 5-fluorouracilbased chemoradiotherapy are as follows (i) EGFR (ERBB1)mRNA levels had a prognostic significance for DFS as wellas for OS (ii) Differences between pretreatment biopsies andresected specimen underline the importance of standardizingpreanalytical procedures in biomarker studies (iii) HighMKI67 mRNA expression correlated by trend with DFS (iv)A relevant proportion of patients had HER2 positive tumorsand therefore a better prognosis On the other hand HER2positivity may offer the possibility for studying trastuzumab-based treatment in this disease

Competing Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Authorsrsquo Contributions

Melanie Kripp and Kirsten Merx contributed equally

References

[1] MG Zampino R Labianca G D Beretta et al ldquoRectal cancerrdquoCritical Reviews in OncologyHematology vol 70 no 2 pp 160ndash182 2009

[2] S Galandiuk H S Wieand C G Moertel et al ldquoPatterns ofrecurrence after curative resection of carcinoma of the colonand rectumrdquo Surgery Gynecology and Obstetrics vol 174 no 1pp 27ndash32 1992

[3] A Martling T Holm L E Rutqvist et al ldquoImpact of a surgicaltraining programme on rectal cancer outcomes in StockholmrdquoBritish Journal of Surgery vol 92 no 2 pp 225ndash229 2005

[4] D Sebag-Montefiore R J Stephens R Steele et al ldquoPreoper-ative radiotherapy versus selective postoperative chemoradio-therapy in patients with rectal cancer (MRC CR07 and NCIC-CTG C016) a multicentre randomised trialrdquo The Lancet vol373 no 9666 pp 811ndash820 2009

[5] W Van Gijn C A Marijnen I D Nagtegaal et al ldquoPreopera-tive radiotherapy combined with total mesorectal excision forresectable rectal cancer 12-year follow-up of the multicentrerandomised controlled TME trialrdquoThe Lancet Oncology vol 12no 6 pp 575ndash582 2011

[6] J-P Gerard T Conroy F Bonnetain et al ldquoPreoperative radio-therapy with or without concurrent fluorouracil and leucovorinin T3-4 rectal cancers results of FFCD 9203rdquo Journal of ClinicalOncology vol 24 no 28 pp 4620ndash4625 2006

[7] J-F Bosset L Collette G Calais et al ldquoChemotherapy withpreoperative radiotherapy in rectal cancerrdquo The New EnglandJournal of Medicine vol 355 no 11 pp 1114ndash1123 2006

[8] J E Tepper M OrsquoConnell D Niedzwiecki et al ldquoAdjuvanttherapy in rectal cancer analysis of stage sex and localcontrolmdashfinal report of Intergroup 0114rdquo Journal of ClinicalOncology vol 20 no 7 pp 1744ndash1750 2002

[9] M JOrsquoConnell J AMartensonH SWieand et al ldquoImprovingadjuvant therapy for rectal cancer by combining protracted-infusion fluorouracil with radiation therapy after curativesurgeryrdquo The New England Journal of Medicine vol 331 no 8pp 502ndash507 1994

[10] C Rodel U Graeven R Fietkau et al ldquoOxaliplatin added tofluorouracil-based preoperative chemoradiotherapy and post-operative chemotherapy of locally advanced rectal cancer (theGerman CAOAROAIO-04 study) final results of the mul-ticentre open-label randomised phase 3 trialrdquo The LancetOncology vol 16 no 8 pp 979ndash989 2015

[11] Y S Hong B-H Nam K-P Kim et al ldquoOxaliplatin flu-orouracil and leucovorin versus fluorouracil and leucovorinas adjuvant chemotherapy for locally advanced rectal cancerafter preoperative chemoradiotherapy (ADORE) an open-label multicentre phase 2 randomised controlled trialrdquo TheLancet Oncology vol 15 no 11 pp 1245ndash1253 2014

[12] L L Gunderson J M Jessup D J Sargent F L Greene andA Stewart ldquoRevised tumor and node categorization for rectalcancer based on surveillance epidemiology and end results andrectal pooled analysis outcomesrdquo Journal of Clinical Oncologyvol 28 no 2 pp 256ndash263 2010

[13] J T Price H M Wilson and N E Haites ldquoEpidermal growthfactor (EGF) increases the in vitro invasion motility andadhesion interactions of the primary renal carcinoma cell lineA704rdquo European Journal of Cancer Part A vol 32 no 11 pp1977ndash1982 1996

[14] K M Woods Ignatoski T Maehama S M Markwart J EDixon D L Livant and S P Ethier ldquoERBB-2 overexpressionconfers PI 3rsquo kinase-dependent invasion capacity an humanmammary epithelial cellsrdquo British Journal of Cancer vol 82 no3 pp 666ndash674 2000

[15] Z Lu G Jiang P Blume-Jensen and T Hunter ldquoEpidermalgrowth factor-induced tumor cell invasion and metastasisinitiated by dephosphorylation and downregulation of focaladhesion kinaserdquoMolecular and Cellular Biology vol 21 no 12pp 4016ndash4031 2001

[16] Y Ueno H Sakurai S Tsunoda et al ldquoHeregulin-inducedactivation of ErbB3 by EGFR tyrosine kinase activity promotestumor growth and metastasis in melanoma cellsrdquo InternationalJournal of Cancer vol 123 no 2 pp 340ndash347 2008

[17] R Radinsky S Risin D Fan et al ldquoLevel and functionof epidermal growth factor receptor predict the metastaticpotential of human colon carcinoma cellsrdquo Clinical CancerResearch vol 1 no 1 pp 19ndash31 1995

[18] T Sasaki T Nakamura R B Rebhun et al ldquoModification ofthe primary tumor microenvironment by transforming growthfactor 120572-epidermal growth factor receptor signaling promotesmetastasis in an orthotopic colon cancer modelrdquo AmericanJournal of Pathology vol 173 no 1 pp 205ndash216 2008

[19] A Ho-Pun-Cheung E Assenat C Bascoul-Mollevi et alldquoEGFR and HER3 mRNA expression levels predict distantmetastases in locally advanced rectal cancerrdquo InternationalJournal of Cancer vol 128 no 12 pp 2938ndash2946 2011

[20] R-D Hofheinz F Wenz S Post et al ldquoChemoradiotherapywith capecitabine versus fluorouracil for locally advanced rectalcancer a randomised multicentre non-inferiority phase 3trialrdquoThe Lancet Oncology vol 13 no 6 pp 579ndash588 2012


[21] K Bohmann G Hennig U Rogel et al ldquoRNA extraction fromarchival formalin-fixed paraffin-embedded tissue a compari-son of manual semiautomated and fully automated purifica-tion methodsrdquo Clinical Chemistry vol 55 no 9 pp 1719ndash17272009

[22] A Koutras K T Kalogeras R M Wirtz et al ldquoEvaluation ofthe prognostic significance of HER family mRNA expression inhigh-risk early breast cancer a Hellenic Cooperative OncologyGroup (HeCOG) Validation Studyrdquo Journal of TranslationalMedicine vol 13 no 1 article 171 2015

[23] T Tramm B S Soslashrensen J Overgaard and J Alsner ldquoOptimalreference genes for normalization of qRT-PCR data fromarchival formalin-fixed paraffin-embedded breast tumors con-trolling for tumor cell content and decay of mRNArdquo DiagnosticMolecular Pathology vol 22 no 3 pp 181ndash187 2013

[24] Y Toiyama Y Inoue S Saigusa et al ldquoGene expression pro-files of epidermal growth factor receptor vascular endothelialgrowth factor and hypoxia-inducible factor-1 with special refer-ence to local responsiveness to neoadjuvant chemoradiotherapyand disease recurrence after rectal cancer surgeryrdquo ClinicalOncology vol 22 no 4 pp 272ndash280 2010

[25] C Jakob T Liersch W Meyer H Becker G B Barettonand D E Aust ldquoPredictive value of Ki67 and p53 in locallyadvanced rectal cancer correlation with thymidylate synthaseand histopathological tumor regression after neoadjuvant 5-FU-based chemoradiotherapyrdquoWorld Journal of Gastroenterol-ogy vol 14 no 7 pp 1060ndash1066 2008

[26] P A Fasching K Heusinger L Haeberle et al ldquoKi67chemotherapy response and prognosis in breast cancer patientsreceiving neoadjuvant treatmentrdquo BMC Cancer vol 11 article486 2011

[27] F Marme A Schneeweiss J Aigner et al ldquoAbstract P3-06-08 Ki-67 mRNA as a predictor for response to neoadjuvantchemotherapy in primary breast cancerrdquo Cancer Research vol72 no 24 supplement 2014

[28] B Ingold Heppner H-M Behrens K Balschun et alldquoHER2neu testing in primary colorectal carcinomardquo BritishJournal of Cancer vol 111 no 10 pp 1977ndash1984 2014

[29] L-C Conradi H Styczen T Sprenger et al ldquoFrequency ofHER-2 positivity in rectal cancer and prognosisrdquoThe AmericanJournal of Surgical Pathology vol 37 no 4 pp 522ndash531 2013

[30] E Valtorta C Martino A Sartore-Bianchi et al ldquoAssessmentof a HER2 scoring system for colorectal cancer Results from aValidation Studyrdquo Modern Pathology vol 28 no 11 pp 1481ndash1491 2015

[31] S Siena A Sartore-Bianchi S Lonardi L Trusolino C Mar-tino and K Bencardino ldquoTrastuzumab and lapatinib in HER2-amplified metastatic colorectal cancer patients (mCRC) theHERACLES trialrdquo in Proceedings of the ASCO Annual Meetingsuppl abstr 3508 Chicago Ill USA 2015

[32] K A David F T Unger P Uhlig et al ldquoSurgical procedures andpostsurgical tissue processing significantly affect expression ofgenes and EGFR-pathway proteins in colorectal cancer tissuerdquoOncotarget vol 5 no 22 pp 11017ndash11028 2014

[33] N W Liu T Sanford R Srinivasan et al ldquoImpact of ischemiaand procurement conditions on gene expression in renal cellcarcinomardquo Clinical Cancer Research vol 19 no 1 pp 42ndash492013

Submit your manuscripts athttpwwwhindawicom

Stem CellsInternational

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Immunology ResearchHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttpwwwhindawicom


proved advantage compared to bolus 5-FU [8] Only admin-istering 5-FU as continuous infusion during radiation led toimproved survival and increased time to relapse [9] Likewisethe addition of oxaliplatin to perioperative fluorouracil treat-ment has given diverging results in five clinical trials [10 11]

Altogether the identification of patients being at high riskfor distant metastases still represents a major challenge totailor the management of rectal cancer therapy For patientswith resected rectal cancermdashin addition to the adequacy ofsurgical excision (evaluated by the circumferential resectionmargins)mdashthe TNM classification is still the most reliableindicator of risk for systemic recurrence [12] However forpatients scheduled to undergo neoadjuvant chemoradiother-apy clinical stratification (eg withMRI scan) has limitationsespecially in view of the inability to exactly predict the nodalstatus Risk stratification based on molecular markers couldprovide better estimate of individual risk and tailored treat-ment In this regard especiallymolecularmarkers bearing thepotential to serve as therapeutic targets formedical treatmentare indispensable

Several drugs are licensed or in clinical evaluation forthe treatment of tumors expressing human epidermal growthfactor tyrosine kinase receptors (EGFR) Besides EGFRwhich is encoded by the ERBB1 gene the human epidermalgrowth factor receptor (HER) family of receptor tyrosinekinases includes three additional members HER2 (ERBB2)HER3 (ERBB3) and HER4 (ERBB4) Activation of the HERsignaling network has been shown to promote tumor inva-sion and metastasis both in vitro [13ndash16] and in vivo [17 18]

Mutations and overexpression of HER family genes arefrequently present in rectal cancer For instance a higher rateof HER2 amplification in high grade tumors was reportedand EGFR and HER3 mRNA expression was described to beassociated with the occurrence of metastases in patients withLARC [19]

Here we sought to evaluate (a) the mRNA expression ofERBB1ndash4 in pretreatment tumor tissue (b) the correlationof expression of each receptor along with tumor cell pro-liferation using MKI67 mRNA expression in patients withLARC treated with 5-FU-based chemoradiotherapy and (c)the impact of these markers on prognosis

2 Material and Methods

Patient Cohort Tumor tissue for this study stemmed frompatients participating in a phase III clinical trial conductedat the University Hospital of Mannheim between 2002 and2007 These patients had histologically confirmed LARC(adenocarcinoma cT3-4 any N or cT2 N+) This nonin-feriority trial compared 5-FU with the oral 5-FU prodrugcapecitabine for the perioperative treatment of LARCDetailsof the study protocol and the results have been publishedpreviously [20] In brief the trial began in 2002 as an adju-vant trial comparing capecitabine-based chemoradiotherapywith fluorouracil-based chemoradiotherapy Patients in thecapecitabine group were scheduled to receive two cycles ofcapecitabine followed by chemoradiotherapy (504Gy pluscapecitabine 1650mgm2) then three cycles of capecitabine

Patients in the fluorouracil group received two cycles ofbolus fluorouracil followed by chemoradiotherapy (504Gyplus infusional fluorouracil 225mgm2 daily) then two cyclesof bolus fluorouracil The protocol was amended in 2005to allow a neoadjuvant cohort in which patients in thecapecitabine group received chemoradiotherapy followed byTME surgery and five cycles of capecitabine and patients inthe fluorouracil group received chemoradiotherapy (504Gyplus infusional fluorouracil 1000mgm2 days 1ndash5 and 29ndash33)followed by surgery and four cycles of bolus fluorouracilPatients were randomly assigned to treatment groups ina 1 1 ratio The primary endpoint was overall survivalNoninferiority of capecitabine in terms of 5-year overallsurvival was tested 392 patients were evaluable with amedianfollow-up of 52 months Five-year overall survival in thecapecitabine group was noninferior to that in the fluorouracilgroup (76 [95CI 67ndash82] versus 67 [58ndash74] 119901 = 00004)Similar numbers of patients had local recurrences in eachgroup 12 [6] in the capecitabine group versus 14 [7] in thefluorouracil group 119901 = 067

All patients providing tumor tissue for the current anal-ysis were treated at a single center (University HospitalMannheim University of Heidelberg)

Both the clinical study protocol and the molecular inves-tigations reported here were approved by the local ethicscommittee

21 RNA Isolation from Formalin-Fixed Paraffin-Embedded(FFPE) Tissue and Quantitative Reverse Transcription-Poly-merase Chain Reaction (RT-qPCR) Assessment Fixation oftumor specimens followed standard protocols 10 bufferedformalin for at least 8 hours Hematoxylin-eosin (HE) stainedsections were evaluated for pathological stage accordingto the 2002 TNM classification of the American JointCommittee on Cancer classification For molecular analy-ses HE slides were reevaluated by an experienced pathol-ogist (SE) for confirmation of the presence of invasivetumor Each case was macrodissected for an invasive tumorareal comprising at least 30 of tumor cells One 10 120583msection was used for the isolation of RNA accordingto a fully automated high-throughput extraction work-flow which runs on an Xtract-XL liquid-handling robot(STRATIFYER Molecular Pathology GmbH Cologne Ger-many) The extraction solutions and chemicals are com-mercially available in Germany as part of the XTRAKTFFPE kit which is based on magnetic bead technology(STRATIFYER) [21] In brief FFPE sections were solubilizedand paraffin was melted by incubating with a lysis buffer ina Thermo-mixer Tissue was lyzed with Proteinase K Thelysates were then admixed with germanium-coated mag-netic particles in buffer-controlled conditions which enhancepreferential attachment of nucleic acid molecules to thesurface of the particles Purification was carried out bymeans of 3 consecutive washing cycles involving magneti-zation centrifugation washing and removal of the supernatant Nucleic acids were eluted with 100 120583L elution bufferand treated with DNase I The DNA-free RNA eluates werestored at minus80∘C until use










EGFR(ERBB1)



ERBB2(HER2)



ERBB3(HER3)



ERBB4(HER4)










MKI67 n = 48


MKI67 n = 52














3 Results




Adjuvanttreatment


67[47ndash83]

62[44ndash75]














EGFR

(ERB

B1)

ERBB

2 (H

ER2)

ERBB

3 (H

ER3)

ERBB

4 (H

ER4)

MKI

67

30

35

40

n




49494949 48Nor

mal

ized

mRN

A ex

pres

sion

(40

-ΔC

q)




4 Discussion




0

20

40

60

80

100



p = 00057

Perc

enta

ge al

ive

(a)

0

20

40

60

80

100




p lt 00001

Perc

enta

ge d

iseas

e fre

e

(b)




p = 00640

20

40

60

80

100

Perc

enta

ge al

ive

(c)




p = ns0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(d)




p = ns0

20

40

60

80

100

Perc

enta

ge al

ive

(e)

p = 0075




0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(f)











5 Conclusion


Competing Interests




References








































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

























EGFR(ERBB1)



ERBB2(HER2)



ERBB3(HER3)



ERBB4(HER4)










MKI67 n = 48


MKI67 n = 52














3 Results




Adjuvanttreatment


67[47ndash83]

62[44ndash75]














EGFR

(ERB

B1)

ERBB

2 (H

ER2)

ERBB

3 (H

ER3)

ERBB

4 (H

ER4)

MKI

67

30

35

40

n




49494949 48Nor

mal

ized

mRN

A ex

pres

sion

(40

-ΔC

q)




4 Discussion




0

20

40

60

80

100



p = 00057

Perc

enta

ge al

ive

(a)

0

20

40

60

80

100




p lt 00001

Perc

enta

ge d

iseas

e fre

e

(b)




p = 00640

20

40

60

80

100

Perc

enta

ge al

ive

(c)




p = ns0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(d)




p = ns0

20

40

60

80

100

Perc

enta

ge al

ive

(e)

p = 0075




0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(f)











5 Conclusion


Competing Interests




References








































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















MKI67 n = 48


MKI67 n = 52














3 Results




Adjuvanttreatment


67[47ndash83]

62[44ndash75]














EGFR

(ERB

B1)

ERBB

2 (H

ER2)

ERBB

3 (H

ER3)

ERBB

4 (H

ER4)

MKI

67

30

35

40

n




49494949 48Nor

mal

ized

mRN

A ex

pres

sion

(40

-ΔC

q)




4 Discussion




0

20

40

60

80

100



p = 00057

Perc

enta

ge al

ive

(a)

0

20

40

60

80

100




p lt 00001

Perc

enta

ge d

iseas

e fre

e

(b)




p = 00640

20

40

60

80

100

Perc

enta

ge al

ive

(c)




p = ns0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(d)




p = ns0

20

40

60

80

100

Perc

enta

ge al

ive

(e)

p = 0075




0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(f)











5 Conclusion


Competing Interests




References








































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

























EGFR

(ERB

B1)

ERBB

2 (H

ER2)

ERBB

3 (H

ER3)

ERBB

4 (H

ER4)

MKI

67

30

35

40

n




49494949 48Nor

mal

ized

mRN

A ex

pres

sion

(40

-ΔC

q)




4 Discussion




0

20

40

60

80

100



p = 00057

Perc

enta

ge al

ive

(a)

0

20

40

60

80

100




p lt 00001

Perc

enta

ge d

iseas

e fre

e

(b)




p = 00640

20

40

60

80

100

Perc

enta

ge al

ive

(c)




p = ns0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(d)




p = ns0

20

40

60

80

100

Perc

enta

ge al

ive

(e)

p = 0075




0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(f)











5 Conclusion


Competing Interests




References








































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















0

20

40

60

80

100



p = 00057

Perc

enta

ge al

ive

(a)

0

20

40

60

80

100




p lt 00001

Perc

enta

ge d

iseas

e fre

e

(b)




p = 00640

20

40

60

80

100

Perc

enta

ge al

ive

(c)




p = ns0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(d)




p = ns0

20

40

60

80

100

Perc

enta

ge al

ive

(e)

p = 0075




0

20

40

60

80

100

Perc

enta

ge d

iseas

e fre

e

(f)











5 Conclusion


Competing Interests




References








































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

























5 Conclusion


Competing Interests




References








































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of



































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Research Article Prognostic Impact of mRNA Expression ...Research Article Prognostic Impact of mRNA...

Documents

Transcript of Research Article Prognostic Impact of mRNA Expression ...Research Article Prognostic Impact of mRNA...