The genetic polymorphism underlying immune checkpoint inhibitor–induced endocrine toxicity.

Abstract

12024 Background: Immune checkpoint inhibitors (ICIs) have transformed cancer therapy by enhancing T cell–mediated anti-tumor immunity; however, immune-related adverse events (irAEs)—particularly endocrine toxicities (e.g., thyroiditis, hypophysitis, adrenal insufficiency; incidence 5–15%)—pose significant clinical risks, including lifelong hormonal dependency. Although germline genetic factors, especially human leukocyte antigen (HLA) variants, are implicated in irAE susceptibility, prior studies lack granularity in toxicity subtyping and comprehensive genetic characterization across diverse clinical contexts. Methods: We integrated high-resolution HLA genotyping with precise phenotypic annotation of endocrine irAEs (subclassified by organ specificity, onset, severity per CTCAE v5.0) in multi-center retrospective and prospective cohorts. A hypothesis-free genome-wide approach identified novel HLA loci, followed by functional validation of candidate variants. Machine learning models were trained to predict toxicity risk by integrating HLA markers with clinical covariates (ICI type, cancer diagnosis, baseline characteristics). Results: Subtype-specific HLA associations were identified (e.g., distinct HLA-DRB1 alleles linked to thyroiditis versus hypophysitis). Collinearity analysis revealed significant positive correlations between non-endocrine irAE pairs (e.g., capillary hyperpermeability and gastrointestinal ulcers; P < 0.01), suggesting shared immunopathogenic pathways. The integrated risk model demonstrated robust stratification performance for endocrine toxicity susceptibility. Cohort demographics reflected real-world ICI utilization patterns (predominantly lung cancer), while clinical response data underscored challenges in achieving durable remission (notable progression rates). Conclusions: This study delineates the HLA-driven genetic architecture underlying endocrine irAEs and elucidates interdependencies among irAE subtypes. Integration of genetic risk profiling with clinical parameters enables personalized toxicity risk assessment, advocating for proactive, multidisciplinary monitoring protocols. These findings advance precision immunotherapy by informing biomarker-guided patient selection, optimizing safety, and refining clinical decision-making frameworks for ICI administration. Clinical trial information: MR-11-25-020518.