SOX2 As a Biomarker of Pathological Response in Oral Cavity Squamous Cell Carcinoma Treated With Neoadjuvant Chemotherapy: A Retrospective Cohort Study

BACKGROUND: Neoadjuvant chemotherapy response is highly variable in locally advanced oral cavity squamous cell carcinoma. SOX2, a stemness-associated transcription factor, promotes chemoresistance through cancer stem cell maintenance. No prior study has evaluated pretreatment SOX2 protein expression as a predictor of pathological response to neoadjuvant chemotherapy in this disease. We present the first retrospective evaluation of this question, assessed against the Hirakawa tumour regression grading system. METHODS: Fifty-four patients with locally advanced oral cavity squamous cell carcinoma receiving neoadjuvant chemotherapy followed by curative surgery were analysed retrospectively. SOX2 immunohistochemistry was performed on paired pre- and post-treatment specimens in 47 patients, defined as the primary analytic cohort. Pathological response was graded using the Hirakawa tumour regression grading system. Kaplan-Meier and Cox proportional hazards analyses were performed for event-free and overall survival. Subgroup analyses stratified by NACT regimen were additionally performed. RESULTS: SOX2 negativity was significantly associated with a favourable pathological response (odds ratio 6.22, 95% confidence interval 1.32-29.29, p = 0.021), with a negative predictive value of 82.4%. This association was directionally consistent across NACT subgroups, which did not differ significantly in pathological response or survival. Good responders demonstrated significantly superior event-free survival (p = 0.006). Nodal metastasis was the dominant independent predictor of event-free survival on multivariate Cox analysis (hazard ratio 6.23, p = 0.026). CONCLUSIONS: SOX2 negativity at baseline is associated with a favourable neoadjuvant chemotherapy response in locally advanced oral cavity squamous cell carcinoma. The Hirakawa tumour regression grading system demonstrated strong prognostic discrimination for event-free survival. Prospective, externally validated evaluation of SOX2 within a multi-marker prediction panel is warranted; these findings should be regarded as hypothesis-generating pending such validation.