Claudin 18.2 Expression in Cholangiocarcinoma: Clinicopathological Heterogeneity, Prognostic Significance, and Impact of Positivity Thresholds

Background/Objectives: Claudin 18.2 (CLDN18.2) has emerged as a clinically actionable therapeutic target in gastrointestinal malignancies; however, its expression profile and prognostic significance in cholangiocarcinoma (CCA) remain incompletely characterized. This study investigated CLDN18.2 expression in CCA and evaluated its associations with clinicopathological characteristics and overall survival (OS). Methods: This retrospective study included 124 patients with histologically confirmed CCA who were diagnosed between 2010 and 2025. CLDN18.2 expression was assessed by immunohistochemistry using two predefined positivity thresholds (≥10% and ≥75% of tumor cells showing moderate-to-strong membranous staining). Associations with clinicopathological variables were analyzed, and prognostic significance was evaluated using Kaplan–Meier survival analysis and Firth’s penalized Cox regression. Results: Using a 10% cutoff, CLDN18.2 positivity was detected in 10.5% (13/124) of cases; using a 75% cutoff, it was detected in 4.8% (6/124) of cases. At the 75% threshold, CLDN18.2 expression was significantly associated with anatomical location (p = 0.013) and histological subtype (p = 0.002), with the highest positivity observed in perihilar CCA (30.0%) and large-duct intrahepatic CCA (21.7%). The median OS was 9.5 months. CLDN18.2 positivity was not significantly associated with mortality in a univariable Firth’s penalized Cox regression at either the 10% cutoff (hazard ratio (HR) = 1.641, 95% confidence interval (CI): 0.883–2.816; p = 0.112) or the 75% cutoff (HR = 1.884, 95% CI: 0.762–3.877; p = 0.155). In multivariable Firth’s penalized Cox regression, CLDN18.2 positivity at the 10% cutoff was independently associated with an approximately two-fold increased risk of mortality (HR = 2.072, 95% CI: 1.090–3.666; p = 0.028), whereas no independent prognostic significance was observed at the 75% cutoff (HR = 1.906, 95% CI: 0.768–3.951; p = 0.150). The absence of adjuvant therapy was independently associated with increased mortality in both models (10% cutoff: HR = 4.870, 95% CI: 1.671–13.784; p = 0.004; 75% cutoff: HR = 4.296, 95% CI: 1.521–11.899; p = 0.006). Conclusions: CLDN18.2 expression in CCA varies according to anatomical location, histological subtype, and positivity threshold. The relatively high expression observed in perihilar CCA, and large-duct intrahepatic CCA, and the independent prognostic significance identified at the 10% cutoff support the potential role of CLDN18.2 as a prognostic biomarker and therapeutic target in selected patients with CCA. Given the limited sample sizes in some anatomical and histological subgroups, these findings should be interpreted cautiously and validated in larger prospective multicenter studies.