Minimal bone resorption after open treatment of mandibular condylar head fractures

Open treatment of condylar base and neck fractures is widely recommended, whereas treatment of condylar head fractures is still controversial and just is removal of osteosynthesis material. In this study, bone resorption and remodelling after open treatment of condylar head fractures were three-dimensionally (3D) assessed and correlated with clinical parameters in a medium follow-up. Of 18 patients with 25 condylar head fractures who underwent open reduction and internal xation, clinical data and cone beam computed tomography (CBCT) datasets were analysed. Condylar processes were segmented in the postoperative and follow-up CBCT scans. Volumetric and linear changes were measured using a sophisticated 3D-algorithm. In the course after surgery, patients function and pain improved signicantly. Low rates of postoperative complications were observed. All 3D measurements showed no signicant bone resorption during the follow-up period. Open reduction of condylar head fractures leads to good patients outcomes and low rates of long-term complications. This study underlines the feasibility and importance of open treatment of condylar head fractures and may help to spread its acceptance as the preferred treatment option.


Introduction
Mandibular fractures are the most common craniofacial fractures with 25%-45% of cases involving the condylar process [1,2] . The treatment of these condylar fractures has been controversial for years. Whereas for fractures located at the condylar neck and base, open treatment has become "gold standard" [3][4][5] with superior function and outcomes compared to closed treatment; however, for condylar head fractures the discourse is ongoing. Despite existing nomenclature systems [2,6,7] , there is often no distinct differentiation between the types of condylar process fractures made in the literature [8][9][10] which has resulted in Nussbaum et al. concluding that a meta-analysis is not feasible due to the inconsistency of reported fractures and nomenclature in the underlying studies [11] . This is made more convoluted due to the different levels of fracture of the condylar process representing completely different, non-comparable types of diseases, morbidity, complications, and outcomes. The latest IBRA position paper did not clearly recommend open treatment for condylar head fractures (CHF) [3] ; however, in more recent studies, open treatment has been suggested and seems to be superior to closed treatment [3,[12][13][14][15] . In the authors' opinion, the so-called closed treatment is likely to be no treatment at all, as any spontaneous repositioning of the fractured condylar head against muscular traction seems highly unlikely to occur. Open treatment of CHF is considered the only way to achieve long-term restoration of function and anatomically stable results with reconstruction of vertical height and prevention of occlusal disturbances [14] .
Osteosynthesis with two positional screws has proven to be appropriate for treating CHF [16] . Although the removal of osteosynthesis material was formerly only recommended in cases of implant failure [17,18] , some of the latest studies using volumetric and three-dimensional evaluation of bone resorption strongly recommend standardised removal four to six months after surgery to prevent screw protrusion and reduce the rate of subsequent joint disorders [15,[19][20][21] .
Previous studies investigated bone resorption after open treatment of CHF using different techniques and made adverse conclusions in terms of screw/plate removal [15,18,19] . The resorption seems to effect the lateral and dorsal aspects of the condylar head, rather than vertical height and articulating surface [19,22,23] .
In this study, we evaluated bone remodelling and resorption in our patient cohort after open treatment of CHF using three-dimensional and volumetric methods in a medium-term follow-up. The surgical and clinical parameters were found to be correlated.

Patient selection
In this retrospective two-centre analysis, the Department of Oral and Maxillofacial Surgery, Hannover Medical School, Hannover, Germany and the Department for Oral and Maxillofacial Surgery, Leipzig University Hospital, Leipzig, Germany, were screened for patients who presented with CHF between 2016 and 2020. Patients had to meet the following inclusion criteria: 1. Displaced condylar head fracture type M or P according to Neff et al. [2] , regardless of the degree of fragmentation (none, minor, major) and vertical apposition (complete, partial, or lost).
2. Open reduction and internal xation (ORIF) were performed using positional screws with or without additional mini-plate osteosynthesis.
3. Complete patient documentation during follow-up of at least six months. 4. Postoperative and follow-up cone beam computed tomography (CBCT) scans of appropriate quality allowing for 3D-segmentation.

Clinical parameters and outcome measures
Pre-, intra-, and postoperative data from patient records was collected and analysed. In addition to aetiology and clinical parameters, TMJ-related parameters-such as maximal interincisal opening (MIO) laterotrusion, deviation and occlusion-and TMJ-related pain and joint noises were observed at two time points (T1 and T2).

3D segmentation and analysis
CBCT scans were performed in clinical routine intra-or immediately postoperatively (T1) for veri cation of repositioning and positioning of osteosynthesis, as well as during the follow-up examinations (T2).
For the three-dimensional analysis of condylar heads, the whole mandibular rami were segmented, using MITK Workbench (German Cancer Research Center, (DKFZ) Division of Medical Image Computing, Heidelberg, Germany). DICOM data sets were imported into MITK Workbench, which features an "Otsu" algorithm for automated threshold-based segmentation [24] . A standardised con guration was applied

Results
Screening of department databases yielded 30 patients meeting the inclusion criteria, of which 12 patients were lost to follow-up. Eighteen patients with eligible datasets at both time points were included in this study. These patients led to a total of 25 openly treated CHF.

Discussion
There are several previous studies on bone remodelling after surgical treatment of CHF [15,18,19] , each of which uses a different analytical approach. Recent studies have described either three-dimensional or volumetric measurements; therefore, we decided to combine both measurements of volumetric alterations and two-dimensional changes of the condylar head, which were modi ed according to Skroch et al. 2020 [19] . By using sophisticated automated segmentation algorithms in the present study, the volume bias during segmentation is reduced. The splitting of mandibular segmentations for volumetric assessment has been modi ed in this study, as we have seen a bias in the determination of the condylar head baseline due to fracture-related alterations in anatomy. The previously described projection of a sphere into the pole zone of the condylar head could be prone to bias [2] ; therefore, we set the cutting line lower, through the most inferior point of the sigmoid notch and the posterior boundary of the mandibular foramen ( g. 1) to prevent interference with osteosynthesis material and trauma-related deformities of the condylar head and neck. These points are reliably recognisable during the segmentation process. This does not affect the total volumetric resorption measurements, but the calculated relative resorption could be estimated to be lower as the reference volume increased. During segmentation and volume measurement, osteosynthesis material outside the condylar heads was excluded. As no surgery to the TJR occurred between T1 and T2 in this study, another reason for bias could be avoided.
In CBCT scans during follow-up in this study, remodelling of the fractured condylar heads was frequently observed (Table 1), even though there was no signi cant change in the condylar head volumes and linear dimensions. These ndings match those of previous studies on this topic, which could not prove any signi cant volumetric changes either [15,19] . Johner et al. observed a volumetric decrease of 15.29%, without any signi cance as well [15] . In our study the mandibular ramus height decreased by a median of 0.84 mm without being statistically signi cant; however, a similar amount of loss of vertical ramus height was observed in previous studies [18,19,22] . The in uence of comminution, dislocation, fracture location, and the presence of additional fractures is unlikely due to the small sample sizes. The studies mentioned before were also unable to demonstrate statistically signi cant correlations [19,20] . A general removal of osteosynthesis material was not performed in this study, except from those cases with osteosynthesis failure (n=1) or intraarticular screws (n=4). Still, the removal of osteosynthesis is point of discussion, unfortunately our data do not allow for any advice on this issue.
Most other studies compared patients' TMJ function using the well-established helkimo index. This index was initially described in 1978 by Helkimo [25] for the classi cation of patients with temporomandibular joint disorders. Since then, it has been widely used for the classi cation and documentation of any kind of TMJ disorder, including fractures of the TMJ; however, this index does not su ciently re ect the characteristics of patients with TMJ fracture or deformities in the course of TMJ trauma. To mitigate this, we used MIO, laterotrusion, and joint pain separately to compare patients' functions. Our clinical ndings show signi cant improvements in TMJ function and pain between T1 and T2, with no severe TMJ disorders or long-term complications at T2. Open treatment of CHF is supported by the ndings of our study and those of previous studies with low resorption and good clinical patient outcomes. Therefore, we strongly recommend the open treatment for condylar head fractures, as it is still not widely performed. However, further prospective clinical trials are required to con rm this trend and avoid shortcomings of a retrospective approach.

Declarations
Author contributions (names must be given as initials)

Funding
This research did not receive any speci c grant from funding agencies in the public, commercial, or notfor-pro t sectors.