Prospective evaluation of candidate SNPs of VEGF/VEGFR pathway in metastatic colorectal cancer (mCRC) patients (pts) treated with first-line FOLFIRI plus bevacizumab (BV). C. Cremolini1, F. Loupakis1,2, D. Yang2, L. Salvatore1, W. Zhang2, T. Wakatsuki2, M. Schirripa1, S. Lonardi3, C. Antoniotti1, G. Aprile4, G. Masi1, F. Graziano5, A. Ruzzo6, S. Lucchesi7, M. Ronzoni8, M.K.H. Maus9, G. Bocci10, G. Tonini11, A. Falcone1 and HJ Lenz2 1.U.O. Oncologia Medica 2, Azienda Ospedaliero-Universitaria Pisana, Istituto Toscano Tumori, Pisa, Italy; 2. University of Southern California Norris Comprehensive Cancer Center, Los Angeles, CA; 3. Istituto Oncologico Veneto, IRCCS, Padova, Italy; 4.Azienda Ospedaliero-Universitaria di Udine, Udine, Italy; 5. UOC Oncologia, A. O., Pesaro, Italy; 6. Section of Biochemistry and Molecular Biology, Department of Biomolecular Sciences, University of Urbino, Urbino, Italy; 7. U.O. Oncologia Medica, Ospedale , Pontedera, Italy; 8. Dipartimento di Oncologia, Istituto Scientifico San Raffaele, Milano, Italy; 9. Response Genetics Inc. , Los Angeles, CA; 10. Division of Chemotherapy and Pharmacology Department of Internal Medicine, University of Pisa, Pisa, Italy; 11. Università Campus Bio-Medico, Rome, Italy; University of Pisa, Pisa, Italy Background Patients and Methods Loupakis et al, BMC 2011 No association of VEGF rs833061 C/T variants with PFS was found 100 To detect an HR for PFS of 1.7 for VEGF rs833061 T/T variant compared to C-, with α and β errors of 0.05 and 0.20 respectively, 199 events were required. Finding Reference VEGF-A rs699947 Predictive Schneider et al. JCO ‘08 VEGF-A rs699946 Predictive Lambrechts et al. ESMO ‘11 60 40 20 0 20 40 60 Time At the univariate analysis, no association of other candidate SNPs with PFS was found, except for VEGFR2 12505758 C/T variants CC (N= 11) median PFS: 10.7 mos CT (N= 107) median PFS: 9.5 mos TT (N= 306) median PFS: 10.9 mos Germ-line DNA was extracted by peripheral blood and candidate SNPs were analyzed by PCR and sequencing C- (N= 118) median PFS: 9.5 mos TT (N= 306) median PFS: 10.9 mos CC CT TT 80 60 40 60 40 20 0 0 20 40 60 CC/CT TT 80 20 0 HR: 1.40 (1.07-1.84) Log-rank test p=0.015 100 Log-rank test p=0.047 100 Percent survival rs number 80 0 As secondary endpoint, we planned a confirmatory analysis of the following candidate SNPs. • A potential prognostic or predictive role of other VEGF, VEGFR1 and 2 and EPAS1 SNPs was suggested by other retrospective series and by the post-hoc analysis of phase III randomized trials as reported below Gene TT TT (N= 147) median PFS: 10.2 mos CC- (N= 276) median PFS: 10 mos HR: 1.17 (0.91-1.50) Log-rank test p=0.218 Percent survival • The association of the T/T variant of VEGF rs833061 C/T SNP with worse PFS and OS was retrospectively reported by our group in a cohort of 111 pts treated with first-line FOLFIRI plus BV but not in a historical cohort of pts treated with first-line FOLFIRI Based on our retrospective findings, we planned a prospective validation trial in mCRC patients treated with first-line FOLFIRI+BV Percent survival • No predictors of benefit from BV have been so far identified Results 0 20 40 60 Time Time VEGFR-1 rs9582036 Predictive Lambrechts et al. ECCO ‘09 VEGFR-1 rs7993418 Predictive Claes et al. AACR ‘12 VEGFR-2 rs11133360 Predictive Lambrechts et al. ESMO ‘11 VEGFR-2 rs12505758 Prognostic Lambrechts et al. ESMO ‘11 Sex M/F VEGFR-2 rs2305948 Predictive Gerger et al. CCR ‘11 Age (median) EPAS-1 rs4145836 Prognostic Lambrechts et al. ESMO ‘11 ECOG PS 0/1/2 357/63/4 84/15/1 Primary site Right/Left/Rectum 110/187/127 26/44/30 Liver-only disease Yes/No 136/288 32/68 The prospective validation is an essential step on biomarkers’ way toward clinical practice. Number of mts sites 1/>1 194/230 46/54 Future studies on predictors of benefit from BV should provide a comprehensive overlook on the complexity of tumoral angiogenesis not only from the genetic perspective. [email protected] MAIN PATIENTS’ CHARACTERISTICS N=424 % 252/172 59/41 62 At the multivariate analysis, including Kohne score, mucinous histology, ECOG PS, LDH levels and primary tumor site as covariates, the association of VEGFR2 12505758 C- variants with shorter PFS was still significant (HR: 1.402 [1.079-1.822], p=0.012).Significance was lost when applying multiple testing correction. Conclusions This prospective trial does not confirm the predictive impact of VEGF rs833061 C/T SNP suggested by our retrospective experience. Also previous findings on other candidate SNPs are not confirmed.