Construction of a new automatic grading system for jaw bone mineral density level based on deep learning using cone...

Area of Science:

• Oral surgery
• Radiology
• Artificial Intelligence

Background:

• Accurate assessment of bone mineral density is crucial for successful oral implant surgery.
• Cone beam computed tomography (CBCT) provides detailed radiographic data for implant site evaluation.
• Current methods for bone density classification can be subjective and time-consuming.

Purpose of the Study:

• To develop and validate an AI-powered deep learning model for automatic classification of bone mineral density from CBCT images.
• To enhance preoperative decision-making and intraoperative tool selection in oral implant surgery.

Main Methods:

• Seventy patients with mandibular defects underwent CBCT scanning.
• 605 CBCT datasets were standardized and classified into five Hounsfield Unit (HU) density levels (Type 1-5).
• An AI deep learning model was trained and validated against manual classifications by experienced dental implant surgeons.

Main Results:

• The AI model demonstrated high accuracy in classifying bone mineral density, with HU mean values closely matching the annotated grading intervals.
• Statistical comparison confirmed the AI model's ability to replicate the density classifications determined by expert surgeons.
• The AI classification system provides a detailed assessment of bone density.

Conclusions:

• The developed AI deep learning method offers a reliable and objective approach to classifying bone mineral density for oral implant surgery.
• This AI tool can assist surgeons in optimizing drilling parameters and tool selection, potentially leading to better surgical planning and execution.
• The AI classification system contributes to more precise and personalized dental implant procedures.