Abstract 5268: Outcome prediction of metastatic colorectal cancer patients undergoing liver resection by analyzing serum metabolomics

Abstract

In metastatic colorectal cancer (mCRC) patients (pts), treatment strategies integrating liver resection with more effective therapies offer better 5-year survival rates than palliative chemotherapy alone. However, resectability is established only on clinical-morphovolumetric criteria, liver resection is a complex and costly procedure and relapse occurs in almost 2/3 of pts after potentially curative resection. Therefore, prompt identification of pts at higher risk of recurrence is critical to avoid not-beneficial, expensive procedures. Aberrant metabolism is an emerging hallmark of cancer and recent observations suggest that specific metabolic changes can be used to stratify pts for prognosis and drug-response. We evaluated by 600MHz NMR spectroscopy the metabolomics profiling on sera from 30 mCRC pts, enrolled in the Obelics trial (NCT01718873), which investigated different schedules of bevacizumab in combination with oxaliplatin plus fluoropyrimidines regimens, and subdivided on the basis of outcome, in good (R) vs bad responders (NR) according to PFS: 12 months or longer (R, n = 12) and shorter than 12 months (NR, n = 18). We compared the samples of the two mCRC pts groups, collected at response evaluation when resectability was established in case of appropriate tumor reduction and PCA, sPLS-DA and loading plots evidenced metabolites with statistically different levels between the two sub-groups. ROC curves were performed to identify the cutoff levels of these significant metabolites to be correlated with patient survival. In this way we demonstrated that low levels of 3-hydroxybutyrate and of hydroxyproline as well as high levels of histidine correlated with both poor progression free survival (PFS) and overall survival (OS). Notably, either 3-hydroxybutyrate or histidine are better predictor of both PFS and OS compared to pathologic response on resected metatastases. Lipidomics analysis confirmed clear differences between R and NR pts indicating statistically significant increase of lipids in NR pts, with both higher triglycerides and phospholipids correlating with poor PFS and OS. This latter effects, may reflect, at least in part a non specific inflammatory response; indeed a significant increase of pro-inflammatory cytokines was also demonstrated in NR pts sera by cytokinomics using multiplex ELISA approach. Finally, basal serum metabolomics analysis in both NR and R pts demonstrated that on-treatment evaluation is more informative than pre-treatment evaluation to stratify patients for outcome. Overall, these data suggest that NMR-based metabolomics is a potent and affordable method that could play a role in the prediction of mCRC outcome. Citation Format: Alfredo Budillon, Susan Costantini, Angela Sorice, Francesca Capone, Silvia Marchese, Elena Di Gennaro, Carlo Vitagliano, Fabiana Tatangelo, Alfonso De Stefano, Franco Bianco, Paolo Delrio, Francesco Izzo, Antonio Avallone. Outcome prediction of metastatic colorectal cancer patients undergoing liver resection by analyzing serum metabolomics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 5268.