Network meta-analysis of 36 RCTs with 8925 patients comparing defined treatment classes for resectable locally advanced gastric cancer, with consistent superiority of chemotherapy plus immunotherapy across survival and pathological endpoints, but limited by indirect comparisons and heterogeneous trial populations.
Small animal study with surrogate molecular endpoints (gene expression, protein levels) and no clinical outcomes; results are mechanistically interesting but require human validation and larger controlled trials.
A multi-algorithm machine learning signature study with external validation across four datasets and functional validation of a candidate gene, showing clear prognostic separation but limited by retrospective cohort design and surrogate endpoints rather than prospective clinical outcomes.
Retrospective cohort study confirming that combined NLR and ALBI grade moderately improve prognostic discrimination in HCC treated with TACE, with clear multivariable analysis but modest effect size and no practice-changing threshold.