Development and validation of a prediction model for risk stratification and outcome prediction in oral oncology patients

Oral squamous cell carcinoma (OSCC) is a prevalent malignancy with significant morbidity and mortality, particularly in low- and middle-income countries. Despite advancements in treatment, prognostic tools integrating clinical, histopathological, and molecular data remain underdeveloped, limiting personalized risk stratification and survival prediction. This study aimed to develop and validate a prediction model for overall survival (OS) and progression-free survival (PFS) in OSCC, incorporating clinical, histopathological, and molecular factors. A retrospective cohort of 132 patients with histopathologically confirmed OSCC was analyzed. Data on demographic, clinical (tumor stage, lymph node involvement), histopathological (tumor grade, perineural invasion), and molecular (HPV status) variables were collected. Logistic regression and machine learning algorithms were used to build the prediction model. Internal validation was conducted using bootstrapping, and model performance was assessed using the area under the receiver operating characteristic (ROC) curve, calibration plots, and decision curve analysis (DCA). The model demonstrated robust predictive performance, with an area under the ROC curve (AUC) of 0.85 for OS and 0.83 for PFS. Tumor stage, lymph node involvement, and HPV status were identified as key predictors of survival. Kaplan-Meier analysis showed steep declines in OS probabilities during the first 24 months, emphasizing the need for early interventions. Calibration plots indicated strong agreement between predicted and observed outcomes, supporting the model's reliability. This study developed a validated prediction model for OS and PFS in OSCC, demonstrating high discriminatory ability and calibration. Integrating clinical, histopathological, and molecular data enhances personalized risk stratification and treatment planning in oral oncology. • Developed a robust OSCC prediction model combining clinical, histopathological, and molecular variables. • Integrated HPV status, tumor stage, and lymph node involvement as key predictors of OS and PFS in OSCC patients. • Employed machine learning algorithms, enhancing model accuracy and clinical applicability for survival prediction. • Achieved high discrimination (AUC 0.85) and calibration, validating the model’s reliability in predicting outcomes. • Study findings emphasize tailored prognostic tools for personalized OSCC management and improved patient outcomes.