Longitudinal liquid biopsy identifies an early predictive biomarker of immune checkpoint blockade response in head and neck squamous cell carcinoma [bulk RNA-seq]

Immune checkpoint blockade (ICB) has improved outcomes in head and neck squamous cell carcinoma (HNSCC), yet there is an unmet need for predictive, blood-based biomarkers to guide treatment. We longitudinally profiled peripheral immune responses in a mouse HNSCC model treated with anti–PD-1, using single-cell and bulk RNA sequencing with paired TCR sequencing. Early expansion of effector memory T cells (Tem) and B cells was associated with ICB response. Here, we show that the LiBIO score, a composite of Tem and B cell signatures, robustly predicts ICB response in HNSCC blood and tumor cohorts, outperforming existing biomarkers. Furthermore, the LiBIO score generalizes to melanoma, non-small cell lung cancer, and breast cancer, demonstrating potential as a clinically applicable blood-based biomarker for ICB response across multiple cancer types. We employed our previously characterized 4MOSC carcinogen-induced orthotopic HNSCC preclinical model to evaluate immune dynamics during ICB therapy. Two cohorts of mice were used, with one cohort comprising 45 mice for bulk RNA sequencing (RNA-seq) and the other comprising 16 mice for single-cell RNA sequencing (scRNA-seq) and single-cell TCR sequencing. Lesions developed consistently at expected time points. Blood samples were collected at four key intervals: pre-treatment (Day 4) and three on-treatment time points (Days 9, 17, and 24), aligned with the administration of anti-PD-1 therapy. Tumor growth and immune responses were monitored longitudinally to assess treatment efficacy. Bulk RNA-seq provided an overview of immune dynamics in the larger cohort, while scRNA-seq and scTCR sequencing in the smaller cohort delivered high-resolution insights into immune repertoire changes, supporting the study’s focus on peripheral immune biomarkers for predicting ICB outcomes.