3562 Background: CDKN2A encodes the inhibitor of the cyclin-dependent kinases p16 and is frequently methylated in colorectal cancer, thus reducing the exp levels of p16 protein. In pre-clinical models of colorectal cancer cells, the demethylation of CDKN2A gene induces topoisomerase I upregulation increasing the sensitivity to irinotecan. In metastatic colorectal cancer (mCRC) pts, the role of p16 expression has been poorly investigated. Methods: Tumour samples of pts randomized in the phase III TRIBE2 study comparing upfront FOLFOXIRI/bevacizumab (bev) versus (vs) FOLFOX/bev were assessed for p16 immunohistochemical exp. A validation analysis in patients treated with FOLFIRI/bev in the TRIBE study was conducted. Results: Out of 679 patients enrolled in the TRIBE2 study, 231 tumours were assessed for p16 exp. Overall, 152 (66%), and 79 (34%) were classified as positive (pos) (IHC 2+ and 3+) and negative (neg) (IHC 0 and 1+), respectively. In the p16 pos group, 69 (45%) and 83 (55%) pts received FOLFOXIRI/bev and FOLFOX/bev, respectively, while in the p16 neg cohort, FOLFOXIRI/bev was administered in 37 (47%) and FOLFOX/bev in 42 (53%) cases. Pts with p16 neg tumours had more frequently an ECOG-PS of 1-2 (16% vs 6%, p=0.0098) and a BRAF mut tumour (26% vs 7%, p=0.0002). No PFS difference was observed between p16 neg and pos patients, while a trend for a longer OS was shown in favour of p16 pos pts (25.5 vs 21.3 months; HR: 0.74, 95%CI: 0.53-1.02, p=0.068) that was not confirmed at the multivariate analysis (p=0.61). Among patients with p16 pos tumours, those treated with FOLFOXIRI/bev reported longer PFS (12.8 vs 9.4 months, HR: 0.55, 95%CI: 0.39-0.78, p=0.0008) and OS (30.0 vs 21.4 months, HR: 0.66, 95%CI: 0.46-0.95, p=0.026) than those receiving FOLFOX/bev. On the other hand, no significant difference was observed among p16 neg patients in terms of both PFS (9.0 vs 9.4 months, HR: 0.91, 95%CI: 0.57-1.46, p=0.69) and OS (21.3 vs 19.5 months, HR: 0.95, 95%CI: 0.58-1.57, p=0.85), thus suggesting a differential treatment effect according to p16 exp (pinteraction for PFS =0.12; pinteraction for OS =0.19). Among patients treated with FOLFIRI/bev in the TRIBE study (N=58) patients with p16 pos tumours reported longer PFS and OS than those with p16 neg. (HR for PFS: 0.45, 95%CI: 0.22-0.92, p=0.026; HR for OS: 0.41, 95%CI: 0.20-0.85, p=0.013). Conclusions: In mCRC, p16 pos exp seems associated with higher benefit from irinotecan. Prospective confirmation in independent randomized series is warranted.