1 Department of Stomatology, The First People’s Hospital of Yongkang (Yongkang First People’s Hospital), 321300 Yongkang, Zhejiang, China
2 Department of Stomatology, Cangnan Hospital of Wenzhou Medical University (The People’s Hospital of Cangnan), 325800 Wenzhou, Zhejiang, China
Correspondence to: Yi Wu, Department of Stomatology, Cangnan Hospital of Wenzhou Medical University (The People's Hospital of Cangnan), 325800 Wenzhou, Zhejiang, China (e-mail: wuyicnyy@163.com).
Editor: Alfredo Iandolo
Abstract
AIM: This study aimed to investigate the predictive value of preoperative radiographic indices, demographic characteristics, and psychological factors for mandibular third molar extraction difficulty, to develop a nomogram, and to interpret feature contributions using SHapley Additive exPlanations (SHAP) analysis.
METHODS: In this retrospective cohort study of 250 patients, demographic characteristics, including age, sex, and body mass index (BMI), together with radiographic indices such as root morphology and psychological factors assessed using the Modified Dental Anxiety Scale (MDAS), were analyzed in relation to surgical difficulty, defined as operative time exceeding 45 minutes. Multivariate logistic regression was used to construct a nomogram, which was validated through receiver operating characteristic (ROC) curves and decision curve analysis (DCA). Feature importance was explored using SHAP analysis, and the association between operative time and perioperative outcomes was assessed.
RESULTS: Multivariate logistic regression identified age, root morphology, Winter’s angulation, and preoperative dental anxiety (MDAS score) as key predictors of high surgical difficulty (p < 0.05). The resulting nomogram demonstrated excellent discrimination, with an area under the curve (AUC) of 0.91. SHAP analysis illustrated that age and Winter’s angulation contribute more to the model’s predictions, followed by root morphology. Longer operative time was independently associated with a higher risk of perioperative complications (odds ratio = 1.03, p < 0.05) and showed a positive correlation with pain intensity on postoperative day 1 (Spearman’s ρ = 0.712, p < 0.001).
CONCLUSIONS: Bulbous or curved root morphology, advanced age, high dental anxiety, Winter’s angulation, and male sex were associated with the difficulty of mandibular third molar extraction. The developed nomogram serves as a precise, clinically interpretable tool for preoperative risk stratification. Integrating psychological evaluation with anatomical assessment facilitates a holistic approach to surgical planning.
Keywords
- mandibular third molar
- extraction difficulty
- preoperative radiographic assessment
- root morphology
- demographic characteristics
- operative time

