Generative Artificial Intelligence for Orthognathic Planning: Patient-Specific 3-Dimensional Jaw Reference Forms...

Area of Science:

• Medical Imaging
• Artificial Intelligence in Surgery
• Orthognathic Surgery Planning

Background:

• Conventional cephalometry has limitations in capturing individual skeletal forms.
• Patient-specific, anatomically complete reference forms are needed for personalized surgical planning.
• Existing cephalometric methods may not fully represent complex facial anatomy.

Purpose of the Study:

• To compare surgical plans generated by a generative artificial intelligence (AI) model with those created by expert surgeons.
• To validate the efficacy of AI-guided surgical planning against traditional methods.
• To assess the similarity between AI-generated and surgeon-derived orthognathic surgery plans.

Main Methods:

• A prospective, in-silico validation study using CT/CBCT scans of patients who underwent orthognathic surgery.
• Surgical plans were created using AI-guided templates and by three expert surgeons independently.
• Plan similarity was quantified using the Dice Similarity Coefficient (DSC) and statistical equivalence was tested.

Main Results:

• The mean Dice Similarity Coefficient (DSC) between AI-guided and surgeon plans was 0.77 (±0.13), comparable to intersurgeon agreement (0.78 ±0.11).
• Statistical equivalence was confirmed, with AI-generated plans falling within the range of expert variability (P < .001).
• The Kolmogorov-Smirnov test showed no significant difference in plan distributions (P = .6).

Conclusions:

• AI-guided surgical planning is statistically equivalent to surgeon-generated plans for orthognathic surgery.
• The AI method produces surgical plans within the natural variability observed among expert surgeons.
• Generative AI offers a viable tool for personalized and precise orthognathic surgery planning.