PURPOSE: Availability data are scarce and primarily retrospective in patients with brain metastasis (BM) from gastrointestinal (GI) cancers. The objective of this cohort was to determine prognostic factors for survival outcomes in patients with BM from GI cancers. METHODS: METACER is a national multicentric prospective cohort study which included patients with BM diagnosis during a histologically proven digestive cancer follow-up between 2010 and 2014. The primary endpoint was overall survival (OS). The secondary endpoints were Progression-Free survival (PFS), prognostic factors, and BM-free survival as time from disease diagnosis to BM diagnosis. RESULTS: METACER included 130 patients, with colorectal cancer (CRC) (N = 105) and eso-gastric (N = 25) cancer (EGC). The median OS was 6.6 months: 7.1 months (95%CI: 4.7-9.7) in CRC patients and 5.2 months, (95%CI: 1.9-7.6) in EG patients (p = 0.827). In multivariate analysis, cerebral BM location (versus cerebellar), BM surgery, performance status (0-1 versus 2), and a unique BM were significantly associated with prolonged OS. BM-free survival were 30.8 months (95%CI:25.2-36.9) in CRC patients and 7.8 months (95%CI:3.8-13.6) in EGC patients (p <
0.001). In synchronous metastatic disease, BM-free survival were 18.6 months (95%CI:13.1-25.2) in CRC patients and 3.7 months (95%CI:0.03-7.8) in EGC patients (p <
0.001). CONCLUSION: BM in GI cancers are of poor prognosis. BM surgery should be considered in case of unique brain lesion. In metastatic settings, EGC patients have shorter BM-free survival than CRC patients.