# Newly Identified Transcriptomic Biomarkers and Gene Signature of Pathological Complete Response to Induction Chemoimmunotherapy in Locally Advanced Head and Neck Squamous Cell Carcinoma

## TL;DR

This study identifies gene signatures and biomarkers that predict successful treatment response in head and neck cancer patients receiving chemoimmunotherapy.

## Contribution

The study introduces a novel 7-gene signature (CheckRad-7) that effectively predicts pathological complete response and survival in head and neck squamous cell carcinoma patients.

## Key findings

- Baseline CD8+ T-cell density, stromal tumor lymphocyte infiltration, and PD-L1 proportion score are associated with pathological complete response.
- A new 7-gene signature (CheckRad-7) achieved an AUC of 0.902 in predicting treatment response and prognosis in HNSCC patients.
- T and B-cell-related pathways and IFN-gamma signaling are significant predictors of pathological complete response.

## Abstract

Neoadjuvant therapies incorporating immune checkpoint inhibitors (ICIs) have shown promise in locally advanced head and neck squamous cell carcinoma (HNSCC). However, biomarkers for pathological complete response (pCR) remain undefined. In the CheckRad‐CD8 trial (NCT03426657), we performed RNA sequencing on pre‐ and post‐treatment biopsies from 77 locally advanced HNSCC patients treated with induction chemoimmunotherapy. Of these, 42 patients achieved pCR, while 35 had residual disease (RD). Differentially expressed genes (DEGs) and pathways were identified using DESeq2 and gene set enrichment analysis. Tumor immune microenvironment analysis, utilizing eight RNAseq deconvolution methods, assessed 266 gene signatures and 38 curated immunotherapy signatures. Baseline intratumoral CD8+ T‐cell density, stromal tumor lymphocyte infiltration, and combined PD‐L1 proportion score were associated with pCR. Pretreatment analysis identified 830 DEGs between pCR and RD, with T and B‐cell‐related pathways enriched in pCR samples. Logistic regression models indicated the significance of T and B cells and IFN‐gamma signaling in predicting pCR. Furthermore, a new CheckRad‐7‐gene signature, with an AUC of 0.902 in the training cohort, effectively predicted pCR and survival, serving as a robust biomarker for prognosis and treatment response in HNSCC.

Scheme and pathological characteristics of the CheckRad‐CD8 trial cohort used for the transcriptome analyses. (A) Flow diagram depicting the patient cohort and study scheme as the basis for whole transcriptome profiling of samples from locally advanced HNSCC patients treated with induction chemoimmunotherapy (iCIT). Patients with pCR or increased CD8+ T cells entered radioimmunotherapy (RIT), patients with stable or decreased CD8+ T cells received chemoradiation or salvage surgery; (B) boxplot of intratumoral CD8+ cell density for samples from the three pathologic response groups. Intratumoral CD8+ cell density of samples from patients with pCR, PPR, and NPR was compared using the Wilcoxon rank‐sum test; (C) boxplot of sTIL for samples from the three pathologic response groups. sTIL of samples from patients with pCR, PPR, and NPR were compared using the Wilcoxon rank‐sum test; (D) boxplot of TPS for samples from the three pathologic response groups. TPS of samples from patients with pCR, PPR and NPR were compared using the Wilcoxon rank‐sum test; (E) boxplot of CPS for samples from the three pathologic response groups. CPS of samples from patients with pCR, PPR, and NPR were compared using the Wilcoxon rank‐sum test.