To date, no clinical study has compared the preoperative and intraoperative 3D positions of fractured mandibular condyles following transoral endoscopically-assisted reduction and osteosynthesis. The primary aim of this study was to analyze and compare the final intraoperative position of the fractured mandibular condyle with a virtually simulated, idealized condylar position. The secondary aim was to assess the association between patient-, trauma-, and procedure-specific variables and intraoperative surgical accuracy.

In this retrospective cohort study, patients who underwent transoral endoscopically-assisted osteosynthesis of mandibular subcondylar fractures — without the use of transbuccal trocars — over a six-year period were included. Demographic, clinical, radiological, and procedural data were analyzed. The final intraoperative position of the condylar fragment, based on intraoperative 3D C-arm imaging, was compared with the anatomic ideal reduction simulated with the preoperative CT scan. Univariate analyses were performed to identify associations between clinical and surgical variables and intraoperative repositioning accuracy. The primary outcome was the geometric accuracy of reduction, quantified using the Dice coefficient, mean multiplanar deviation, and rotational deviation.

A total of 86 patients with 95 subcondylar fractures. Eleven cases were excluded due to incomplete data and insufficient quality for segmentation. The most common mechanism of injury was bicycle trauma (n = 35; 36.8 %), followed by violence (n = 26; 27.4 %) and tripping falls (n = 13; 13.7 %). The mean Dice coefficient was 0.62 ± 0.18, the mean multiplanar deviation was 2.79 ± 1.53 mm, and the average rotational deviation was 10.18° ± 6.17° Univariate analysis revealed that greater preoperative sideward displacement of the condylar fragment was significantly associated with higher multiplanar deviation (p = 0.03) and a lower Dice coefficient (p = 0.05). Moreover, a higher length of the condylar fragment was significantly associated with a lower risk of intraoperative rotational deviation (p = 0.003). There was no association between the time interval from trauma to surgery and the intraoperative surgical accuracy.

This study confirms that endoscopically-assisted intraoral osteosynthesis enables accurate reduction in most cases. However, complex fracture morphologies—particularly those with significant sideward displacement or short proximal segments—pose a higher risk for positional deviations. These findings support the use of intraoperative 3D imaging and virtual modelling to improve surgical precision and establish new quality benchmarks in maxillofacial trauma surgery.