525PD - HSA-MIR31-3P expression as a predictor of anti-EGFR response in wild-type KRAS patients with metastatic colorectal cancer

Date 29 September 2012
Event ESMO Congress 2012
Session Gatrointestinal tumors, colorectal
Topics Colon Cancer
Rectal Cancer
Translational Research
Presenter Pierre Laurent-Puig
Authors P. Laurent-Puig1, G. Manceau1, J. Bachet2, B. Chibaudel3, F. Liebaert4, O. Bouché5, F. Penault-Llorca6, M.D. Diebold7, T. André3, S. Imbeaud8
  • 1Umr-s775, Université Paris Descartes, Paris/FR
  • 2Gastroenterology Hepatology, Hôpital Pitié-Salpêtrière, Paris/FR
  • 3Department Of Medical Oncology, Hôpital Saint Antoine, FR-75012 - Paris/FR
  • 4R&d, Integragen, Evry/FR
  • 5Hopital Robert Debré, Reims/FR
  • 6Recherche, CJP-ERTICA, Clermont-Ferrand/FR
  • 7Hepatology-gastroenterology, Hopital Robert Debré, Reims/FR
  • 8Umr-s674, INSERM, Paris/FR

Abstract

In colorectal cancer, KRAS mutations are associated with resistance to anti-EGFR antibodies. A major challenge is to identify, in wild-type KRAS patients, markers that predict response to this therapy. We focused on miRNAs, which can play role in the resistance to anti-EGFR based chemotherapy for metastatic colorectal cancer (mCRC). We first analysed 1145 miRNAs in 84 colorectal tumors and 5 normal colon mucosae. We then conducted a study with 3 subgroups of patients. Group 1 is a retrospective series of patients treated by cetuximab and irinotecan, group 2 is a prospective collection of patient treated by cetuximab or panitumumab based chemotherapy and group 3 is a series of patients prospectively treated by panitumumab and irinotecan as third-line. Using a Cox proportional hazards model, fitted using principal components analysis of group 1, we identified a predictive signature of 11 miRNA linked to disease free survival (p < 0.01). Validation by RT PCR showed that all the survival information is associated with miRNA hsa-mir-31-3p. We tested expression of this miRNA and the presence or absence of mutation BRAF on group 1 of 33 patients, in a Cox model, and found a hazard ratio (HR) of 1.9 CI95% [1.1-2.9]. In the two prospective series (38 patients) the prognostic impact of hsa-mir31-3p the HR is estimated to 1.9 CI95% [1.1-3.1]. We applied the multivariate model obtained from group 1 to group 2 and 3 in order to predict the disease free survival of the patients. The accuracy of the prediction measured by the AUC is 0.77. The performance of the test remains stable from 10 weeks to 48 weeks. A Cox proportional hazards model based on the hsa-mir31-3p logged expression, fitted using principal components from group 1 and BRAF mutation status as a clinical covariate allowed us to classify group 2 and 3 according to a free progression survival risk score (P = 0.005) with a specificity of 62% [95% CI: 38%-82%] and a sensitivity of 82% [95% CI: 56%-96%] for the prediction model. We established a nomogram, taking into account mir-31-3p expression level, age, gender and BRAF mutation status, which predict the progression risk (P < 0.0001). It is the first tool for selecting individual patients with a wild-type KRAS tumor for anti-EGFR therapy.

Disclosure

P. Laurent-Puig: Received clinical research grant from Integragen.

F. Liebaert: Integragen S.A Employee.

All other authors have declared no conflicts of interest.