Predictive Model for Bone Availability in Patients With Unilateral Cleft Lip and Palate: A Machine Learning Approach

Area of Science:

• Oral and Maxillofacial Radiology
• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Unilateral cleft lip and palate (UCLP) presents unique challenges in maxillary bone development and surgical reconstruction.
• Accurate assessment of cortical bone thickness is crucial for successful orthodontic treatment and dental implant placement in UCLP patients.
• Existing methods for bone thickness evaluation may lack precision and patient-specific detail.

Purpose of the Study:

• To develop and validate a machine learning model for predicting cortical bone thickness at any maxillary location in patients with UCLP.
• To identify key predictors of cortical bone thickness in the UCLP population.
• To enhance pretreatment risk assessment and surgical planning for UCLP patients.

Main Methods:

• A retrospective cross-sectional cohort study involving 50 UCLP patients and 50 controls.
• Cone-beam computed tomography (CBCT) data were analyzed to measure cortical bone thickness across multiple regions and depths.
• A Random Forest regression model was employed to predict bone thickness, with feature importance analysis conducted.

Main Results:

• The Random Forest model explained 75% of the variance in cortical bone thickness (R² = 0.75).
• Anatomical region, depth, age, and cleft status were significant predictors of bone thickness.
• The model demonstrated high accuracy, particularly in posterior regions and at greater depths, and identified high-risk zones with 89% sensitivity (AUC = 0.94).

Conclusions:

• The developed predictive model offers accurate, patient-specific estimations of maxillary cortical bone thickness in UCLP patients.
• This tool can significantly aid in pretreatment risk assessment, optimize surgical planning, and improve patient counseling.
• Further external validation is recommended to confirm the model's generalizability across different centers.