Pediatr Blood Cancer 2009;53:1035–1039

VEGF Expression as a Prognostic Marker in Osteosarcoma

Jyoti Bajpai, MD,1 Meharchand Sharma, MD,2 Vishnubhatla Sreenivas, PhD,3 Rakesh Kumar, MD,4

Shivanand Gamnagatti, MD,5 Shah Alam Khan, MS,6 Shishir Rastogi, MS,6

Arun Malhotra, MD,4 and Sameer Bakhshi, MD1*

INTRODUCTION

New blood vessel formation (angiogenesis) is a fundamental

event in the process of tumor growth and metastatic dissemination.

The vascular endothelial growth factor (VEGF) pathway is well

established as one of the key regulators of this process. Activation

of the VEGF-receptor pathway triggers a network of signaling

processes that promote endothelial cell growth, migration, and

survival from pre-existing vasculature. In addition, VEGF mediates

vessel permeability, and has been associated with malignant

effusions. More recently, an important role for VEGF has emerged

in mobilization of endothelial progenitor cells from the bone

marrow to distant sites of neovascularization. Due to its central role

in tumor angiogenesis, the VEGF/VEGF-receptor pathway has

become a major focus of research and antiangiogenic drug

development in oncology [1].

Increased VEGF production has been shown to be important in

the growth of various solid tumors in humans including osteosar-

coma, gastric, esophageal, colorectal, renal, lung, and breast

carcinomas [2–4]. In osteosarcoma, patients with VEGF positive

tumors have poorer disease-free and overall survival compared with

those with VEGF negative tumors [3,5]. VEGF expression in pre-

treated osteosarcoma specimens is predictive of eventual develop-

ment of pulmonary metastasis; further circulating VEGF levels by

ELISA were found to be significantly higher in patients with

osteosarcoma who had pulmonary metastasis [5]. Prognostic value

of post-neoadjuvant chemotherapy (NACT) VEGF expression is

largely unexplored. Therefore, in the present study we investigated

the prognostic potential of VEGF expression at baseline as well as in

post-NACT surviving tumor cells in relation with histologic

necrosis, an established robust prognostic factor in osteosarcoma

[6,7].

MATERIALS AND METHODS

This is a prospective, diagnostic study conducted at our institute

from January 2006 to December 2008. Treatment naive osteosarcoma

patients with adequate organ function for receiving NACT and

adequate biopsy sample for analysis were eligible for the study. After

a detailed history and examination, patients were subjected to initial

investigations of MRI and staging workup, which included CT chest

and bone scan. Patients were staged as per AJCC staging system [8].

NACT included three cycles of cisplatin (40 mg/m2) and doxorubicin

(25 mg/m2), both for 3 days every 3 weeks. Following NACT, the

patient was reevaluated, underwent radical resection of tumor (limb

salvage or amputation) and assessment of histologic necrosis in the

resected specimen. The initial biopsy block was reviewed for the

grade, histologic subtype, and VEGF expression by immunohisto-

chemistry (IHC). The resected tumor specimen post-NACT was

examined for histopathologic necrosis and VEGF expression. All the

slides were coded and evaluated by a pathologist, who was blinded

with regard to the clinical status and results of VEGF staining or

histopathologic necrosis of the patient. This study was approved by

the ethics committee and the institutional review board.

IHC Analysis

After an initial review of all the available hematoxylin and eosin

(H&E) stained slides of the biopsy and surgical specimens, one

Background. The vascular endothelial growth factor (VEGF)pathway is the key regulator of angiogenesis. In osteosarcomabaseline VEGF is of proven prognostic value but prognosticpotential of post-NACT VEGF expression is largely unexplored.Procedure. Treatment naive patients with osteosarcoma weresubjected to initial staging workup followed by three cycles ofneoadjuvant chemotherapy (NACT) and surgery; resected tumorswere assessed for histological necrosis by Huvos grading. Initialbiopsy and resected tumor specimens post-NACT were examined forVEGF expression by immunohistochemistry. Positive VEGF expres-sion was considered when intensive positive staining was observedin >10% of the tumor cells. VEGF expression at baseline wascompared with grade of tumor; pre-NACT and post-NACT VEGFexpression were compared with histological necrosis. Receiveroperating characteristic curves were generated to assess best

threshold and predictability. Results. A total of 31 patients wererecruited with median age of 17 years (range 5–66 years); male/female ratio was 25:6; 23 patients (74%) were non-metastatic. Atbaseline, there was 90% concordance between positive VEGFexpression and higher histological grade (28/31); baseline VEGFexpression did not correlate well with stage and histologicalnecrosis. Twenty-one (67%) were poor and 10 (33%) were goodhistologic responders; post-NACT VEGF expression as well as VEGFchange following NACT significantly correlated with histologicalnecrosis. Conclusion. Positive VEGF expression in survivingtumor cells post-NACT in resected tumors appears to be an importantnegative prognostic factor in osteosarcoma which may helpfuture therapies to be identified according to the angiogenic potentialof the disease. Pediatr Blood Cancer 2009;53:1035–1039.� 2009 Wiley-Liss, Inc.

Key words: angiogenesis; necrosis; neoadjuvant chemotherapy; osteosarcoma; vascular endothelial growth factor (VEGF)

� 2009 Wiley-Liss, Inc.DOI 10.1002/pbc.22178Published online 20 July 2009 in Wiley InterScience(www.interscience.wiley.com)

——————1Department of Medical Oncology, Dr. B. R. A. Institute Rotary

Cancer Hospital, All India Institute of Medical Sciences, New Delhi,

India; 2Department of Pathology, All India Institute of Medical

Sciences, New Delhi, India; 3Department of Biostatistics, All India

Institute of Medical Sciences, New Delhi, India; 4Department of

Nuclear Medicine, All India Institute of Medical Sciences, New Delhi,

India; 5Department of Radiodiagnosis, All India Institute of Medical

Sciences, New Delhi, India; 6Department of Orthopedics, All India

Institute of Medical Sciences, New Delhi, India

*Correspondence to: Sameer Bakhshi, Associate Professor of Pediatric

Oncology, Department of Medical Oncology, Dr. B. R. A. Institute

Rotary Cancer Hospital, All India Institute of Medical Sciences, New

Delhi 110029, India. E-mail: sambakh@hotmail.com

Received 4 February 2009; Accepted 2 June 2009

paraffin-embedded tissue block was selected from each case in

which viable tumor cells were present. Five microns thick sections

were recut and routine H&E stained sections of each case

were reviewed and diagnosis reconfirmed. IHC was done by

streptavidin–biotin peroxidase complex method using monoclonal

antibodies to anti-human VEGF rabbit monoclonal immuno-

globulin G antibody (dilution 1:100) (BIO SB, Santa Barbara,

CA). For the negative controls we used pancytokeratin (M/s; Dako,

Glostrup, Denmark) and human glioblastoma multiforme was taken

as positive controls.

The cell types with positive staining for VEGF were defined

morphologically by using H&E staining. Our analysis was

semiquantitative wherein we counted 100 surviving tumor cells

and positive VEGF expression was considered when intensive

positive staining of VEGF was observed in>10% of the tumor cells.

Further subdivision included grade I as 11–25%, grade II as 26–

50%, and grade III as 51–100% cells showing positive staining of

VEGF [9,10].

Histopathologic Response Assessment

Tumor necrosis was graded as per Huvos pathologic tumor

response grading wherein grade I is <50% necrosis; grade II is

50–89% necrosis; grade III is 90–99%necrosis; and grade IV is

100% necrosis. Grades III and IV (�90% necrosis) were considered

as good responders while grades I and II (<90% necrosis) were

considered as poor responders [11].

Statistical Analysis

Stata software version 9.1 was used for the analysis of data. At

baseline VEGF expression of tumor cells in the biopsy specimens

were compared with histologic grade of tumors; histopathologic

necrosis was compared with baseline and post-NACT VEGF

expression. Receiver operating characteristic (ROC) curves were

generated to assess the best threshold and predictability.

RESULTS

A total of 31 osteosarcoma patients were recruited for the study

with a median age of 17 years (range 5–66 years); the male/female

ratio was 25:6 and 23/31 (74%) patients were non-metastatic.

Osteoblastic osteosarcoma was the most common histopathological

subtype in 14 patients (45%). Lower end of femur was the most

common site seen in 13 (42%) patients. Patients distribution

according to AJCC staging showed stage IIa in 9 (29%), IIb in 14

(45%), stage IVa in 2 (7%), and stage IVb in 6 (19%) patients. Limb

salvage with reconstruction surgery could be performed in 23/31

patients; remaining eight patients were amputated. Twenty-one

(67%) were poor histologic responders (grade I: 16 and grade II: 5

patients) necrosis while 10 (33%) were good histologic responders

(grade III: 8 and grade IV: 2 patients). The mean necrosis in the

resected specimens was 47� 37% (range: 10–100%).

VEGF Expression of Tumor Cells

In the baseline biopsy specimen VEGF expression was negative

in 3/31 (10%) patients while post-NACT it was negative in 6/31

(19%) patients. At baseline, all the 31 cases were of histologic high

grade and 28 of these were VEGF positive; thus, there was a

concordance of 90% between the positive expression of VEGF and

histologic grade. Amongst the VEGF positive patients, 22/28 (79%)

patients at baseline and 10/25 (40%) patients in post-NACT group

had grade III positive expression. The mean percentage of tumor

cells expressing VEGF at baseline in biopsy specimens was

78.38� 34.65 while post-NACT the mean percentage of surviving

tumor cells in the resected specimen was 48.7� 36%; the mean

change following NACT (baseline� post-NACT) was 29.7�47.5% (Table I) (Figs. 1A–D and 2A–D)

Post-NACT VEGF expression as well as the change in VEGF

expression following chemotherapy (baseline VEGF� post-NACT

VEGF) showed significant association with histologic necrosis. The

area under the ROC curve for the post-NACT VEGF expression was

91.7% with 95% confidence interval (CI) of 81–100%, which

implies that in 91.7% cases the post-NACT VEGF expression could

discriminate the responders and non-responders correctly (Fig. 3A).

Further, the best threshold value of post-NACT VEGF expression

was observed to be 30%, at which level the sensitivity was 90.0%

(CI: 71.4–100%) and specificity 90.5% (CI: 77.9–100%) for good

histological response. Similar results were also observed for the

change in VEGF expression with a threshold value of 50%, at which

level the sensitivity was 70% (CI: 41.6–98.4%) and specificity 67%

(CI: 46.5–86.7%) (Fig. 3B). Baseline VEGF expression did not

correlate with disease stage (mean percentage of tumor cells

expressing VEGF in stages I and II was 73.3� 37.9 while in stages

III and IV it was 93.1� 17.5, P¼ 0.17). Further, baseline VEGF

expression was not correlated well with histologic necrosis (area

under curve was 48%, CI 28.2–67.9%) (Fig. 3C).

Survival Characteristics

All patients were followed up for 21.9 months, with a median

follow-up of 9.4 months. Among patients with post-NACT VEGF

expression �30%, the event free survival at the end of 20 months

follow-up was 70.8% as against 39.8% among the group with>30%

post-NACT VEGF expression (P¼ 0.23) (Fig. 4A). Similar results

were also noted in the overall survival between the two groups

(90.9% versus 68.3%, respectively, P¼ 0.28) (Fig. 4B).

Pediatr Blood Cancer DOI 10.1002/pbc

TABLE I. VEGF Expression at Baseline and Post-NACT in OS

VEGF

grade

Tumor cells

positive for

VEGF (%)

Number of

patients

baseline

Number of

patients post-

NACT

VEGF negative (�10%)

0 0–10 3 6

VEGF positive (>10%)

I 11–25 1 5

II 26–50 5 10

III 51–100 22 10

Variable

(% of tumor cells positive)

Baseline Post-NACT

Mean VEGF� SD 78.38� 34.65 48.7� 36

Mean VEGF change� SD 29.7� 47.5

1036 Bajpai et al.

DISCUSSION

The importance of VEGF in cancer progression has been

validated in other organ cancers [12,13]; in sarcoma also VEGF

expression was found to have association with stage and grade of

tumors, as well as survival and pulmonary metastasis [3,14]. In the

present study, we hypothesized that the cells surviving after NACT

to be more important because the surviving population of the cells

may be the one that results in recurrence or subsequent metastases.

Our approach differs from those in previous studies, which have

examined the prognostic influence of VEGF expression in

diagnostic biopsy specimens before any NACT is given [3]. The

result in our patient population suggests that baseline VEGF

expression had significant association with histologic grades of

tumor with a concordance of 90% between the VEGF expression

and histologic grade. However, there was no significant association

of baseline VEGF expression with stage of the disease, which may

have been due to the disproportionate stage distribution in our

cohort. Similar lack of association with stage was found in a study by

Kaya et al. [3] in which disproportionate stage distribution was

proposed as an explanation.

Previous investigators have identified various factors associated

with a poor prognosis in patients with osteosarcoma; the most

consistent factor identified is a poor response (<90% necrosis) after

NACT [6,7]. In previous studies although baseline VEGF correlated

with survival, it did not significantly correlate with necrosis [3,14].

In the present study we also could not find an association of baseline

VEGF with necrosis. The reason for this may lie in the fact that

necrosis and VEGF expression may be independent variables for

survival.

There is one retrospective study that explored the prognostic

value of post-NACT VEGF status of surviving cells in which more

than 25% VEGF positivity of post-NACT surviving cells was

associated with poor survival; however, there was no significant

correlation with histologic necrosis [9]. In the present study, post-

NACT VEGF expression in the surviving cells was significantly

associated with favorable histologic necrosis. Further, decrease

in VEGF expression following NACT from baseline was also

significantly associated with favorable histologic necrosis

which suggests that in the chemotherapy era, chemosensitivity of

the disease may be an important factor for ultimate outcome. Thus,

pre-NACT VEGF may be more relevant if patients do not receive

treatment but with effective chemotherapy it is the post-NACT

VEGF expression which may be more meaningful. Our study was

not aimed at survival end points because of short follow-up and

inclusion of both metastatic and non-metastatic patients of

osteosarcoma. However, Kaplan–Meier survival curves showed a

positive trend in favor of those patients who had �30% VEGF

expression in surviving tumor cells post-chemotherapy, in com-

parison to those who had >30% VEGF expression. This difference,

however, was not statistically significant. The basic demographic

features of osteosarcoma in our center revealed that approximately

one-fourth of our patients were metastatic at presentation which is in

contrast with the metastatic rate of 11.4–20% reported by other

investigators [15,16]. This may be the result of referral

patterns as our center is a major tertiary care center in Northern

India and an increased proportion of advanced cases maybe be

referred. It is, however, difficult to comment on whether the disease

biology in this part of the country results in a more aggressive

disease at the outset.

Pediatr Blood Cancer DOI 10.1002/pbc

Fig. 1. Good responder: Pre-chemotherapy biopsy (H&E) slide

(A) and pre-chemotherapy biopsy slide showing 100% VEGF

expression (B); post-chemotherapy resection specimen of same patient

(H&E) slide showing extensive areas of necrosis (>90%) (C) and post-

chemotherapy resection specimen of same patient showing no VEGF

expression of surviving cells (D). [Color figure can be viewed in the

online issue, which is available at www.interscience.wiley.com.]

Fig. 2. Poor responder: Pre-chemotherapy biopsy (H&E) slide (A)

and pre-chemotherapy biopsy slide showing 100% VEGF expression

(B); post-chemotherapy resection specimen of same patient (H&E)

slide showing focal areas of necrosis (10%) with extensive viable areas

(C) and post-chemotherapy resection specimen of same patient showing

75% VEGF expression of surviving cells (D). [Color figure can be

viewed in the online issue, which is available at www.interscience.

wiley.com.]

VEGF in Osteosarcoma 1037

High VEGF expression of surviving tumor cells in post-

chemotherapy resected tumors appears to be an important

negative prognostic factor in osteosarcoma. Suppression of tumor

angiogenesis, for example, by inhibition of the action of VEGF, has

shown promise in animal models as a potential new therapeutic

strategy for treatment of osteosarcoma [17]. If the results obtained in

the present study can be reproduced in a larger cohort of the patients

then it will further establish role of post-NACT VEGF as a

prognostic factor in osteosarcoma and may serve as a platform to

build future therapies according to the angiogenic potential of the

disease and provide means to test antiangiogenic molecules in

osteosarcoma.

REFERENCES

1. Hicklin DJ, Ellis LM. Role of the vascular endothelial growth factor

pathway in tumor growth and angiogenesis. J Clin Oncol 2005;23:

1011–1027.

Pediatr Blood Cancer DOI 10.1002/pbc

Fig. 3. Association of VEGF expression in surviving tumor cells post-NACT with histologic necrosis (A); association of change in VEGF

expression following NACT from baseline with histologic necrosis (B); and association of VEGF Expression in tumor cells pre-NACT with

histologic necrosis (C).

Fig. 4. Kaplan–Meier Survival curves with respect to post-chemotherapy VEGF expression of surviving tumor cells (median follow-

up¼ 9.4 months) showing event free survival (A) and overall survival (B). Y-axis¼ Survival proportion; VEGF 0¼ post-chemotherapy

�30%VEGF expression of surviving tumor cells; VEGF 1¼ post-chemotherapy >30%VEGF expression of surviving tumor cells.

1038 Bajpai et al.

2. Handa A, Tokunaga T, Tsuchida T, et al. Neuropilin-2

expression affects the increased vascularization and is a

prognostic factor in osteosarcoma. Int J Oncol 2000;17:291–295.

3. Kaya M, Wada T, Akatsuka T, et al. Vascular endothelial growth

factor expression in untreated osteosarcoma is predictive of

pulmonary metastasis and poor prognosis. Clin Cancer Res 2000;

6:572–577.

4. Voest EE, D’Amore PA. Tumor angiogenesis and microcirculation.

New York, NY: Marcel Dekker; 2001.

5. Kaya M, Wada T, Kawaguchi S, et al. Increased pre-therapeutic

serum vascular endothelial growth factor in patients with early

clinical relapse of osteosarcoma. Br J Cancer 2002;86:864–

869.

6. Saeter G, Hoie J, Stenwig AE, et al. Systemic relapse of patients

with osteogenic sarcoma: Prognostic factors for long term survival.

Cancer 1995;75:1084–1093.

7. Enneking WF, Spanier SS, Goodman MA. A system for the surgical

staging of musculoskeletal sarcoma. Clin Orthop 1980;153:106–

120.

8. Greene FL, Page DL, Fleming ID, et al. AJCC Cancer Staging

Manual, 6th edition. New York: Springer; 2002.

9. Charity RM, Foukas AF, Deshmukh NS, et al. Vascular endothelial

growth factor expression in osteosarcoma. Clin Orthop Relat Res

2006;448:193–198.

10. Choi JY, Jang KT, Shim YM, et al. Prognostic significance of

vascular endothelial growth factor expression and microvessel

density in sophageal squamous cell carcinoma: Comparison with

positron emission tomography. Ann Surg Oncol 2006;13:1054–

1062.

11. Rosen G, Marcove RC, Huvos AG, et al. Primary osteogenic

sarcoma: Eight years experience with adjuvant chemotherapy.

J Cancer Res Clin Oncol 1983;106:55–67.

12. Knopp MV, Weiss E, Sinn HP, et al. Pathophysiologic basis of

contrast enhancement in breast tumors. J Magn Reson Imaging

1999;10:260–266.

13. George ML, Dzik-Jurasz AS, Padhani AR, et al. Non-invasive

methods of assessing angiogenesis and their value in predicting

response to treatment in colorectal cancer. Br J Surg 2001;88:1628–

1636.

14. DuBois S, Demetri G. Markers of angiogenesis and clinical

features in patients with sarcoma. Cancer 2007;109:813–819.

15. Kager L, Zoubek A, Potschger U, et al. Primary metastatic

osteosarcoma: Presentation and outcome of patients treated on

neoadjuvant Cooperative Osteosarcoma Study Group protocols.

J Clin Oncol 2003;21:2011–2018.

16. Bacci G, Bertoni F, Longhi A, et al. Neoadjuvant chemotherapy for

high-grade central osteosarcoma of the extremity: Histologic

response to preoperative chemotherapy correlates with histologic

subtype of the tumor. Cancer 2003;97:3068–3075.

17. Tsunemi T, Nagoya S, Kaya M, et al. Postoperative progression of

pulmonary metastasis in osteosarcoma. Clin Orthop Relat Res

2003;407:159–166.

Pediatr Blood Cancer DOI 10.1002/pbc

VEGF in Osteosarcoma 1039