Abstract
OBJECTIVE: This study aimed to construct and validate region-specific nomograms to predict ameloblastoma recurrence and to investigate potential geographic differences in recurrence-related risk factors between South Korea and China. MATERIALS AND METHODS: A total of 816 patients with ameloblastoma treated between 2006 and 2023 were retrospectively analyzed from Yonsei University Dental Hospital (n = 372) and Hospital of Stomatology, Wuhan University (n = 444). Demographic, radiographic, and pathological variables were collected. Logistic regression analysis identified recurrence-associated predictors. Separate nomograms were developed and internally validated for each cohort. Model performance was evaluated using the area under the curve (AUC), calibration curves, and decision curve analysis. RESULTS: The South Korean nomogram incorporated six variables, including pathological type, malignant transformation, surgical method, imaging type, tumor size, and cortical bone destruction (AUC = 0.757). The Chinese model incorporated eight predictors, including sex, root resorption, and number of involved teeth (AUC = 0.787). Calibration and decision curves indicated strong agreement between predicted and observed outcomes and favorable clinical applicability. CONCLUSIONS: Both nomograms demonstrated good predictive accuracy and highlighted regional differences in risk factors. These findings support the integration of multicenter data to enhance recurrence prediction in ameloblastoma. CLINICAL RELEVANCE: The developed tools can guide individualized treatment planning and long-term follow-up, aiding clinicians in early identification of high-risk patients and optimizing surgical strategies.