Log margin-to-thickness ratio improves disease-specific survival prediction in oral cancer: A single cancer centre database: XXXX

OBJECTIVE: We examined whether dynamic margin criteria margin-to-thickness (MTR) ratio has superior predictive value compared with the resection margin or tumour thickness alone in the survival outcome in oral squamous cell carcinoma (OSCC). DESIGN: This is a retrospective cohort study. SETTING: Oral squamous cell carcinoma patients treated in Kaohsiung Veterans General Hospital Cancer Center between January 2006 and December 2013. PARTICIPANTS: A cohort of 302 patients with OSCC who had undergone surgical management. MAIN OUTCOMES MEASURES: Log MTR was calculated for each patient, and survival data were analysed using a multivariable Cox regression model. Discriminative analysis was performed using chi-square, Akaike information criterion (AIC) and Harrell's C tests. RESULTS: After assessing for discriminative ability, the linear trend of log MTR surpassed those of resection margin and tumour thickness in chi-square, AIC and Harrell's C tests for the advanced pathologic T (pT) category. A multivariate Cox proportional hazard regression model revealed that log MTR <33% was associated with less favourable 5-year disease-specific survival (DSS) (P = 0.006) in the entire oral cancer study cohort. Other significant factors included perineural invasion (P = 0.021), pT category, (P = 0.005), pathologic N category (P < 0.001) and differentiation category (P = 0.022). CONCLUSIONS: Log MTR < 33% may be a predictor of less favourable outcome in the DSS of OSCC. Log MTR outperformed both resection margin and tumour thickness alone in terms of discriminative analysis. Our study could help in presurgical planning for high-risk patients and in aiding the decision-making process for adjuvant treatment.