Applying artificial intelligence to assess the impact of orthognathic treatment on gender-affirming facial features.

Abstract

Background: Orthognathic therapy aims to correct severe maxillofacial deformities and enhance facial appearance. Certain orthognathic osteotomies may achieve gender-affirming changes, shifting a face towards a more feminine or masculine appearance. The objective was to develop a convolutional neural network (CNN) to quantify orthognathic therapy's effect on gender-defining facial features, in relation to malocclusion and osteotomy.

Methods: A computational algorithm combining a face detector and a CNN trained on >0.5 million gender-labelled facial images and further fine-tuned on dedicated 7591 face images to predict perceived gender was applied on 2164 annotated facial images of 146 orthognathic patients. Perceived gender was graded and post-treatment scores were compared with pre-treatment scores, differentiating between corrected malocclusion and performed surgery.

Results: Female orthognathic patients demonstrated a significant feminizing shift (P < .001) particularly evident in the Class III female subgroup. This feminization effect was also observed in female patients with vertical, transverse and retrognathic facial disbalances. Male patients experienced strengthened masculine facial features especially in the Class II subgroup (P = .029), while the correction of vertical and transverse discrepancies resulted in minimal change. The most pronounced gender-affirming changes were achieved when bimaxillary osteotomies were performed in conjunction with chin osteotomy (P ≤ .002).

Implications: CNN-based models offer dependable possibilities to quantify orthognathic treatment's impact on gender-related facial features, underlining their value in treatment planning and patient education. The results suggest that orthognathic corrections can significantly alter gender-affirming facial traits, which must be considered when defining individual treatment goals.