An Artificial Intelligence-Based Cosmesis Evaluation for Temporomandibular Joint Reconstruction

Area of Science:

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

Background:

• Temporomandibular joint (TMJ) reconstruction after condylar resection presents significant challenges in mandibular reconstruction.
• Objective assessment of aesthetic outcomes for TMJ reconstruction methods is limited.
• Simple TMJ reconstruction using fibular grafts is a potential but under-evaluated option.

Purpose of the Study:

• To develop an artificial intelligence (AI)-based system for evaluating aesthetic outcomes in simple TMJ reconstruction.
• To assess the functional outcomes of simple TMJ reconstruction using fibular grafts.
• To compare the aesthetic and functional results of simple TMJ reconstruction with native condyle preservation.

Main Methods:

• Retrospective review of patients undergoing segmental mandibular resection and reconstruction with fibular flaps (2011-2020).
• Development of a facial recognition AI to calculate mandibular asymmetry values as a primary aesthetic outcome measure.
• Comparison of outcomes between simple TMJ reconstruction (n=10) and native condyle preservation (n=18).

Main Results:

• No significant difference in postoperative mandibular asymmetry values between simple TMJ reconstruction and native condyle preservation groups.
• No significant differences observed in functional outcomes, including deglutition and speech.
• The AI-based asymmetry evaluation system proved effective for assessing aesthetic outcomes.

Conclusions:

• AI-based asymmetry evaluation is a valuable tool for assessing aesthetic outcomes in mandibular reconstruction.
• Simple TMJ reconstruction with a fibular graft end demonstrates promising aesthetic and functional results comparable to native condyle preservation.
• This technique offers a viable alternative for TMJ reconstruction following condylar resection.