PD-L1 Expression in Biliary Tract Cancer: Comparison Across Antibody Clones and Role as a Predictor of Response to Chemoimmunotherapy: A Meta-Analysis

Abstract

PURPOSEPD-L1 positivity in biliary tract cancers (BTCs) is reported from 4% to 76%. BTC clinical trials have not demonstrated PD-L1 expression as a predictor of response to chemotherapy combined with immune checkpoint inhibitor (chemo-ICI). This meta-analysis examines PD-L1 positivity rates in BTC and association between PD-L1 expression and outcomes in patients treated with chemo-ICI.MATERIALS AND METHODSObservational studies or clinical trials reporting tissue-based PD-L1 expression by Tumor Proportional Score/Combined Positive Score using immunohistochemistry were included. Clinical trials of BTC treated with chemo-ICI were included to assess PD-L1 expression on treatment response. PubMed, Embase, Web of Science, and Cochrane Library were searched for relevant studies before November 15, 2023. Methods of PD-L1 assessment, including antibody clone, cutoff for PD-L1 positivity, and anatomical subtype, were analyzed. Overall survival (OS) and objective response rates (ORRs) were the main outcomes. The generic inverse variance method and random-effect model were used to assess pooled effect sizes.RESULTSFifty-six studies met eligibility criteria. Among 7,768 patients, pooled PD-L1 positivity was 30%. Positivity rates varied significantly by antibody clone (5H1, 58% v SP142, 17%; P =.02). Clinical trials reported a higher positivity rate compared with observational studies (48% v 26%; P <.01). Across five phase I/II clinical trials (194 patients), PD-L1 ≥1% patients tended to have a better ORR than PD-L1 <1% patients (64% v 46%; P =.08). In two randomized controlled trials (874 patients), PD-L1 ≥1% had a statistically significant improvement in OS (hazard ratio [HR], 0.83; P <.01), while PD-L1 <1% did not (HR, 0.85; P =.2).CONCLUSIONGiven the high PD-L1 positivity rate seen in this study, as well as a possible signal of predictive response for chemo-ICI treatment, PD-L1 expression should be further explored as a predictive biomarker.