Younghac Kim, Eun‐Hye Kim, Man Ki Chung, Han‐Sin Jeong, Young-Ik Son, Jinyong Kim, Nayeon Choi
Background.: This study investigates the effects of various clinical data on rapid recurrence (recurrence within 3months) of oral tongue squamous cell carcinoma (OTSCC) and develops prediction models using various combinations of data. Methods.: We divided 944 OTSCC patients into two groups based on timing of recurrence: the rapid recurrence group (RRG, N=89) and the non-rapid or no-recurrence group (NRRG, N=855). Treatment-related data were collected and analyzed to identify risk factors for overall survival (OS) and rapid recurrence. A logistic regression-based prediction model was developed to identify the likelihood of rapid recurrence, and validation was performed using 10-fold cross-validation, leave-group-out cross-validation, and a calibration plot. Results.: Significantly lower OS rates were identified with increasing age, the T3-4 and N1-3 stages, a higher maximum preoperative positron emission tomography standardized uptake value (PET SUVmax), and rapid recurrence. Among the 89 RRG patients, T2-4 and N3 stages, a positive resection margin, and higher PET SUVmax were significant predictors of rapid recurrence, and postoperative chemoradiation was a preventive factor. A nomogram was developed to facilitate easy prediction of rapid recurrence in real-world clinical practice by risk-stratifying each clinical variable. Conclusion.: Rapid recurrence strongly predicts poor oncologic outcomes in OTSCC. Our validated prediction model, incorporating imaging and pathological factors, might help identify high-risk patients and support more tailored treatment and follow-up strategies.