Abstract
BACKGROUND: Liver metastases (LMs) are related to poor efficacy of immune checkpoint inhibitor (ICI)-containing therapies. In the AtezoTRIBE trial, Immunoscore-Immune-Checkpoint (immunoscore-IC) was a predictor of benefit from atezolizumab in mismatch repair-proficient (pMMR) metastatic colorectal cancer (mCRC). PATIENTS AND METHODS: In pMMR patients enrolled in the AtezoTRIBE study, we investigated the association of LMs with immune-related biomarkers and treatment outcomes, and the predictive role of immunoscore-IC in the LMs group. RESULTS: Out of 202 pMMR patients, 151 (75%) had LMs. No differences in immune-related features were observed according to the presence or not of LMs, except for a lower prevalence of tumour-infiltrating lymphocytes-high tumours in the LMs group (33% versus 52%, P = 0.03). Worse outcomes were observed among patients with LMs [progression-free survival (PFS), P = 0.002; overall survival (OS), P = 0.011], also in multivariable models. The effect of adding atezolizumab to FOLFOXIRI/bevacizumab was independent from LMs in terms of PFS (P(int) = 0.990) and OS (P(int) = 0.800). Among patients with pMMR mCRC and LMs, those with immunoscore-IC-high but not those with immunoscore-IC-low tumours achieved benefit from atezolizumab, though in the absence of a statistically significant interaction effect (P(int) for PFS and OS = 0.166 and 0.473, respectively). CONCLUSIONS: LMs are associated with poor prognosis in pMMR mCRC and do not predict resistance to the addition of atezolizumab to FOLFOXIRI/bevacizumab. Immunoscore-IC seems to retain its predictive impact also among patients with LMs.