Testing immune priming failure in laryngeal squamous cell carcinoma: direct evidence, unresolved links, and a compartment-resolved framework

Abstract

Direct evidence for a continuous immune-priming defect in laryngeal squamous cell carcinoma (LSCC) remains incomplete. This Review therefore does not treat failed priming as an established LSCC mechanism. Instead, it separates direct LSCC observations from evidence derived from mixed-site head and neck squamous cell carcinoma cohorts and mechanistic studies in other tumors, using this hierarchy to identify supported pathway nodes and unresolved connections. Direct LSCC studies support recurrent loss of MHC class I antigen-processing components, favorable associations between tumor HLA-DR and clinical outcome, accumulation of PD-L2-positive plasmacytoid dendritic cells, remodeling of dendritic cell states, and an association between tumor-cell STING abundance and advanced disease. None of these findings demonstrates antigen-specific priming. Cross-tumor evidence provides mechanistic context for cDC1 cross-presentation, CCR7-dependent migration, tumor-draining lymph-node priming, and functional cGAS-STING signaling, but these links require LSCC-specific validation. We therefore distinguish MHC class I from MHC class II presentation, cross-presentation from tumor-cell HLA-DR expression, lymph-node priming from intratumoral restimulation, and STING abundance from activation, downstream output, and function. Translational priorities are grouped into evidence-based tiers according to the strength of LSCC evidence and intervention tractability. The MHC-I antigen-processing machinery, PD-L2-positive pDC states, and functional cGAS-STING competence merit the earliest LSCC-specific validation, whereas CLEC9A-, XCR1-, and FLT3L-directed cDC1 strategies remain mechanistically compelling but rely mainly on evidence from other tumor systems. Future multicenter studies should incorporate HLA genotype, population diversity, paired tumor and draining-node sampling, functional assays, and bioinformatic or artificial-intelligence integration with external validation. The proposed sensing-low, presentation-deficient, regulatory-DC-dominant, and signal-present but functionally disconnected phenotypes remain research constructs rather than clinical selection tools. The framework is intended to test where priming fails, if it fails at all, without presenting an unproven LSCC-specific pathway as established biology.