Validation of distant metastases risk-groups in oral cavity squamous cell carcinoma patients treated with postoperative intensity-modulated radiotherapy.

Abstract

This study aimed to derive distant metastases (DM) risk-groups in oral cavity squamous cell carcinoma (OSCC) patients treated with postoperative intensity-modulated radiation therapy (PO-IMRT). OSCC patients treated with PO-IMRT were divided into discovery (2005-2012) and validation (2013-2014) cohorts. DM predictors were identified from multivariable analysis (MVA) to derive low- and high-risk groups in the discovery-cohort. The result was subsequently evaluated in validation-cohort. Overall 447 patients were included (discovery-cohort: n = 300, and validation-cohort: n = 147). Between the two cohorts, there were no significant differences in DM (p = 0.16) or OS (p = 0.26). MVA identified pN2-3 and histological grade 2-3 (G2-3) as DM predictors. High-risk group included patients who had both poor predictors (pN2-3 and G2-3), while low-risk group included patients with no or only one poor predictor. In discovery-cohort, 3-year distant control (DC) was 78% and 97% in high- and low-risk groups respectively (p < 0.001, concordance index = 0.72). In validation-cohort, risk-group classification performed similarly (concordance index = 0.73). The 3-year OS for high- versus low-risk group was 85% versus 95% in discovery-cohort (p < 0.001), and 74% versus 93% in validation-cohort (p < 0.001). A model (G2-3/pN2-3) which identifies high DM risk was validated internally. This model might be used to design future prospective studies investigating treatment intensification and/or DM surveillance.