Artificial Intelligence–Assisted Clinical Decision Support for Persistent Dentoalveolar Pain: A Pilot Machine Learning Study

Background: Persistent Dentoalveolar Pain (PDAP) is a chronic orofacial pain condition characterized by persistent pain in the dentoalveolar region without identifiable odontogenic pathology. Clinical management remains challenging because of heterogeneous presentations and variable treatment responses. Artificial intelligence (AI) may support clinicians by identifying treatment patterns and facilitating personalized treatment selection.

Objective: To develop and evaluate a machine learning–based clinical decision support model capable of generating individualized treatment recommendations for patients with PDAP.

Methods: A retrospective dataset of 99 patients diagnosed with PDAP at a university-based Orofacial Pain clinic was analyzed. Clinical variables included demographic information, pain characteristics, pain location, pain severity, and prior dental treatment history. Treatment approaches were categorized into systemic pharmacologic therapy, topical/local therapy, supplements, multimodal therapy, and education/watchful waiting. A Random Forest classifier was trained using structured clinical features to predict treatment categories. Model performance was compared with multinomial logistic regression and evaluated using stratified train-test splitting and five-fold cross-validation.

Results: The machine learning model successfully generated ranked treatment recommendations based on patient-specific clinical features. Compared with multinomial logistic regression, the Random Forest model demonstrated superior classification performance, suggesting an improved ability to capture complex nonlinear relationships among clinical variables. Feature importance analysis identified baseline pain severity, pain pattern, pain location, and prior dental treatment history as influential predictors of treatment selection.

Conclusions: This pilot study demonstrates the feasibility of AI-assisted clinical decision support for PDAP management. Machine learning may help clinicians navigate complex non-odontogenic pain presentations, reduce unnecessary dental interventions, and support personalized treatment planning. Future work will focus on multicenter validation and prospective evaluation of clinical utility, usability, and clinician acceptance.