A novel composite residual viable tumor score redefines prognosis in oral squamous cell carcinoma following neoadjuvant immunochemotherapy

Abstract

Standard ypTNM staging may not capture residual viable tumor (RVT) across primary and nodal sites after neoadjuvant immunochemotherapy (NICT) for locally advanced oral squamous cell carcinoma (OSCC). Our aim was to derive and internally evaluate a composite RVT score. This retrospective cohort study was conducted at a tertiary referral center between June 2019 and December 2024. Eligible patients had clinical stage III–IV OSCC, and received at least two NICT cycles followed by curative-intent surgery. The predictor was the higher RVT percentage in the primary tumor bed or any involved lymph node. The primary and secondary outcomes were event-free survival (EFS) and overall survival (OS), respectively. Restricted cubic splines and cut-point analysis defined strata; multivariable Cox models, bootstrap-corrected C-indices, net reclassification improvement (NRI), and decision-curve analysis assessed associations and performance. The cohort comprised 353 patients with a mean age of 57.8 ± 9.5 years (range 30–79 years); 261 (73.9%) were men and 92 (26.1%) were women. Three strata were identified: ≤ 10% RVT ( n = 247), 11–40% RVT ( n = 68), and > 40% RVT ( n = 38). Compared with ≤ 10% RVT, the 11–40% and > 40% strata were independently associated with worse EFS (adjusted hazard ratio [HR] 4.87 [95% CI 3.12–7.60] and 9.25 [95% CI 5.41–15.80], respectively) and OS (adjusted HR 5.12 [95% CI 3.07–8.54] and 9.87 [95% CI 5.38–18.11], respectively); all p < 0.001. Discrimination improved versus ypTNM (optimism-corrected C-index 0.76 vs 0.70; NRI 0.42 [95% CI 0.29–0.55]). The derived composite RVT score identified distinct prognostic groups, and appeared to add prognostic information beyond ypTNM staging. External validation is required before clinical use.