Claudin 18.2 Expression and Outcomes in Resected Gallbladder and Biliary Tract Adenocarcinomas: A Pilot Clinicopathologic Study.

AimsBiliary tract cancers (BTCs) have dismal outcomes and limited validated predictive biomarkers. We investigated the immunohistochemical expression of CLDN18, PD-L1, and HER2 in resected gallbladder and biliary tract adenocarcinomas and evaluated their associations with clinicopathologic features and survival outcomes.MethodsFifty-three surgically resected adenocarcinomas from the gallbladder (n = 26) and biliary tract (extrahepatic cholangiocarcinoma (eCCA) (n = 25), intrahepatic cholangiocarcinoma (iCCA) (n = 2)). Whole-tissue sections were stained for CLDN18.2, PD-L1 (SP263), and HER2. CLDN18 expression was defined as any unequivocal membranous staining; high-level positivity was defined as moderate-strong (2+/3+) membranous staining in ≥75% of tumor cells. PD-L1 was scored by TPS and CPS (CPS ≥1 positive).ResultsCLDN18 expression was detected in 28/53 (53%) and was more frequent in gallbladder carcinoma (GBC) than non-gallbladder tumors (p = .05); high-level CLDN18 positivity was present in 11/53 (21%). PD-L1 positivity was infrequent by TPS (5/53, 9%) but higher by CPS (19/53, 36%); all TPS-positive tumors occurred in the GBC subgroup. HER2 immunoreactivity was observed in 7/53 (13%), predominantly low-level. CLDN18 expression and high-level positivity were associated with improved DFS (p = .028 and p = .009) and RFS (p = .031 and p = .035). In multivariable models, margin positivity independently predicted worse OS and DFS (OS HR ∼3.5; DFS HR 3.17), while high-level CLDN18 independently predicted improved DFS (HR 0.27). PD-L1 and HER2 showed limited survival discrimination overall.ConclusionsHigh-level CLDN18 positivity identifies a prognostically favorable subset of resected BTC, independently associated with lower recurrence risk. These findings support CLDN18.2 as a clinically relevant biomarker for risk stratification and potential trial selection, warranting prospective multicenter validation with standardized scoring and sampling strategies.