Deep learning-automatic 3D analysis of regional condylar remodeling and skeletal relapse following bimaxillary surgery: a two-year follow-up study

The aim of this study was to quantify postoperative regional condylar remodeling and its role in skeletal relapse after bimaxillary surgery using CBCT scans. 40 patients with mandibular hypoplasia who underwent bimaxillary surgery were analyzed. CBCT scans were acquired preoperatively, one-week postoperatively and two-years postoperatively. Deep-learning based 3D regional condylar volume analysis was performed and skeletal relapse was quantified through voxel-based matching. The maxilla and mandible were advanced by a mean of 2.87 mm and 8.37 mm, respectively, with a corresponding mean sagittal relapse of .57 mm and 1.62 mm. Postoperative volume loss was observed in 78% of condyles (mean loss = 242 mm 3 ; 16%). Linear regression analyses identified age, gender and magnitude of advancement as significant predictors of relapse. Condylar volume loss significantly mediated the relationship between surgical advancement and skeletal relapse, accounting for 29.1% of the total effect. No statistically significant correlation was found between specific condylar regional sub-volumes and directions of surgical movement. These findings underline the importance of condylar volume stability in predicting long-term skeletal stability among patients who undergo bimaxillary surgery with large advancements. The results also indicated that condylar remodeling is a generalized phenomenon related to surgical movements rather than a localized process restricted to specific anatomical regions of condyles.