Abstract
OBJECTIVE: The application of immunoscore (IS) in non-metastatic colorectal cancer has been validated. We performed a meta-analysis to evaluate the prognostic value of IS on survival in patients undergoing liver resection for colorectal liver metastasis (CRLM). METHODS: Following the PRISMA 2020 guidelines, the protocol was registered in PROSPERO under number CRD42024498641. PubMed, Embase, Web of science, and the Cochrane Register of Controlled Trials databases were searched. To evaluate and compare the prognostic value of the IS, we independently assessed both the liver metastasis (metastatic IS) and the matched primary tumor (primary IS) for each patient with CRLM. Survival indicators include OS, RFS, DFS, with sufficient data to calculate hazard ratios (HRs) and 95% confidence intervals (CIs). The literatures’ quality was assessed by Quality in Prognostic Studies (QUIPS) tool. A pooled analysis was performed using a random-effects or fixed-effects model. Meta regression, subgroup analysis, influence analysis, Baujat plot and Galbraith plot were used to explore the heterogeneity. Publication bias was assessed using Egger’s test, funnel plots, contour-enhanced funnel plot or trim-and-fill analysis. Sensitivity analyses were conducted to examine the robustness of the findings. All outcomes were calculated by Stata 15 (Stata Corporation, TX, USA) and R software (version 4.3.2). RESULTS: A total of 1395 participants from 9 retrospective studies were enrolled in this meta-analysis. The pooled analysis showed that a low metastatic IS in the liver lesions was associated with prolonged OS(HR = 0.38, 95% CI:0.23–0.63, I(2) = 83.0%, P<0.001, for high vs. low metastatic IS) and RFS(HR = 0.63, 95% CI:0.47–0.85, P = 0.002). Even after the exclusion of the two studies contributing most to heterogeneity, the core finding remained unchanged. Four studies explored the association between OS and primary IS and found no significant association (HR = 1.31; 95% CI:0.97–1.77, I(2) = 0%, P = 0.511) with moderate-to-high heterogeneity (I² = 58.4%, P = 0.51). The high primary IS group did not demonstrate better outcomes compared to low-IS group (HR = 1.31; 95% CI:0.97–1.77, I(2) = 0%, P = 0.511). The meta-analysis also demonstrated that high IS was associated with a lower incidence of synchronous metastases (n = 484, OR = 0.51; 95% CI:0.22–0.78, P = 0. 002) and multiple liver metastases (n = 326, OR = 0.63; 95% CI:0.40-1.00, P = 0. 05). No significant publication bias was detected in most vital outcomes and sensitivity analyses confirmed the stability of the primary outcomes. CONCLUSIONS: The IS derived from the metastatic liver lesion, but not from the primary tumor, serves as a robust prognostic biomarker for patients with CRLM. This paradigm shift emphasizes the need to immunologically characterize metastatic lesions for precise prognostication. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12885-026-15867-w.