Pre-existing TCR clones drive major pathological responses in HNSCC patients treated with dual immune checkpoint inhibitors

Purpose: Immune checkpoint inhibitors (ICI) elicit variable responses in head and neck squamous cell carcinoma (HNSCC), yet mechanisms driving major pathological responses (MPR) remain poorly defined. We sought to evaluate longitudinal CD8 T-cell repertoire evolution to identify determinants of MPR. Experimental design: We analyzed high-resolution single-cell TCR sequencing data from paired pre- and post-treatment CD8 tumor-infiltrating lymphocytes (TILs) obtained from HNSCC patients treated with neoadjuvant anti-PD-1 combined with either anti-CTLA-4 or anti-LAG-3. Results: Contrary to "clonal replacement" hypothesis, post-treatment CD8 T-cell pools were dominated by pre-existing TCR clones regardless of clinical outcome. MPR was uniquely characterized by higher abundance and greater expansion magnitude of "super-expanded" clones. We developed the TCR Adaptivity Index (TAI) to quantify coordinate flux (expansion and contraction) of all TCR clones detected across pre- and post-treatment timepoints; this index emerged as the most significant parameter associated with MPR. Importantly, clonal expansion in non-MPR was "uncoupled" from the productive, therapy-induced transcriptional reprogramming-characterized by markers of effector vigor and cellular fitness-that was observed in MPR. Furthermore, expansion dynamics positively correlated with predicted tumor reactivity as calculated by the Tumor Reactive Signature (TRS) score. Finally, a TRS-integrated TAI remained significantly correlated with MPR. Conclusions: Dual ICI drives MPR predominantly through the adaptivity and functional reprogramming of pre-existing immunity. Successful therapy relies on a coordinate repertoire response that promotes transition of putative tumor-reactive super-expanders into productive functional states. TRS-integrated TAI provides a high-throughput framework incorporating clonal dynamics and functional reprogramming to predict therapeutic efficacy.