Abstract
Morphological changes of the condyles are often observed following orthognathic surgery. In addition to clinical assessment, radiographic evaluation of the condyles is required to distinguish the physiological condylar remodelling from pathological condylar resorption. The low contrast resolution and distortion of greyscale values in cone beam computed tomography (CBCT) scans have impeded an accurate three-dimensional (3D) rendering of the condyles. The current study proposes a novel semi-automated method for 3D rendering of condyles using CBCT datasets, and provides a clinical validation of this method. Ten patients were scanned using a standard CBCT scanning protocol. After defining the volume of interest, a greyscale cut-off value was selected to allow an automatic reconstruction of the condylar outline. The condylar contour was further enhanced manually by two independent observers to correct for the under- and over-contoured voxels. Volumetric measurements and surface distance maps of the condyles were computed. The mean within-observer and between-observer differences in condylar volume were 8.62 mm 3 and 6.13 mm 3 , respectively. The mean discrepancy between intra- and inter-observer distance maps of the condylar surface was 0.22 mm and 0.13 mm, respectively. This novel method provides a reproducible tool for the 3D rendering of condyles, allowing longitudinal follow-up and quantitative analysis of condylar changes following orthognathic surgery.
Morphological changes of the condyles are observed regularly following orthognathic surgery. As a result of altered mechanical load, condyle seating, and the intrinsic adaptive capability of the temporomandibular joints (TMJ), dimensional changes of the condyles may occur in the postoperative period. According to their magnitude and rate of progression, the postoperative morphological changes of the condyles can be either physiological or pathological. In contrast to the self-limiting, physiological form of condylar remodelling, patients who experience pathological progressive condylar resorption (PCR) may be confronted with postoperative relapse, anterior open bite (AOB), decrease of posterior facial height, and temporomandibular disorders. The distinction between condylar remodelling and condylar resorption is sometimes difficult to make, and requires radiographic assessment. The TMJ is a difficult region to depict due to the dense cranial bones that surround and overshadow the relatively small condyles.
Conventional two-dimensional (2D) imaging techniques (e.g., lateral cephalograms, orthopantomograms, and tomograms) have been used widely to depict the condyles. However, 2D radiographic images are not only often difficult to interpret, but the lack of reproducibility also impedes an accurate monitoring of changes to the condylar contour during the follow-up period. The 2D representation of the three-dimensional (3D) morphologically complex anatomy of the TMJ has major drawbacks in the analysis of condylar remodelling.
The emergence of cone beam computed tomography (CBCT) in the past years has overcome many shortcomings of 2D imaging techniques. CBCT offers a real-size dataset based on a single low-radiation-dose scan, with the potential of scanning the patient in an upright sitting position. From this dataset, 3D reconstructions and multiplanar cross-sections of the maxillofacial bony structures can be computed. In light of this new imaging modality, an accurate 3D quantitative analysis of postoperative morphological changes of the condyles could be performed, including the invaluable volumetric analysis of the condyles, a hiatus in the field up until now. However, the low contrast resolution and distortion of Hounsfield units (HU value) in CBCT scans have hampered an accurate 3D reconstruction and surface rendering of the condyles. In order to counteract this problem, a new method of post-processing of the CBCT data was developed to improve the process of 3D condylar reconstruction.
The purpose of this study was (1) to present a novel semi-automated method for both the qualitative and quantitative 3D virtual evaluation of the condyles using CBCT data, and (2) a clinical validation of this method.
Subjects and methods
Subjects
Ten adult Caucasian patients with a mandibular retrognathia (two males and eight females) who had undergone bilateral sagittal split mandible advancement surgery (BSSO) were randomly selected from the department’s 3D database. The mean age of patients at the time of surgery was 38.1 years (range 19–58 years).
A CBCT scan was made of all 10 patients at 1 week preoperatively. The patients were scanned vertically in a natural, seated position with the occlusal plane parallel to the horizontal positioning line of the scanner, using a standard CBCT scanning protocol. CBCT scanning (i-CAT, 3D Imaging System, Imaging Sciences International Inc., Hatfield, PA, USA) was performed in ‘extended field’ mode (field of view: 16 cm diameter/22 cm height; scan time: 2 × 20 s; voxel size: 0.4 mm) at 120 kV and 3–8 mA pulse mode with a radiation dose of 136 μSv for a single scan. Data from the CBCT were exported in DICOM format and rendered to a 3D virtual skull model using Maxilim ® software (Medicim NV, Mechelen, Belgium).
Image processing
Each 3D virtual skull model was processed following a step-by-step protocol, as outlined below.
Step 1. Standardized positioning of the 3D virtual skull model in the 3D viewer
A Cartesian 3D cephalometric reference frame was set up around the 3D virtual skull model in order to position the 3D model of each patient in a uniform position. This 3D reference frame was constructed around the hard tissue nasion and sella landmarks, as described by Swennen et al. Consequently, the infraorbitale and porion bilateral landmarks were identified. The position of both porions was double-checked on the sagittal slices of CBCT data and manually adjusted if necessary. The Frankfurter horizontal (FH) plane was constructed through the left and right infraorbitale and the virtual landmark half-way between the left and right porion. Lastly, the vertical plane was constructed perpendicular to the Frankfurter plane ( Fig. 1 A).
Step 2. Determination of the volume of interest (VOI)
After the correct positioning of the 3D skull model, the C-point was identified as the most caudal point of the sigmoid notch bilaterally. A plane that goes through the C-point and runs parallel to the Frankfurter plane, referred to as the C-plane, was subsequently constructed in a semi-automated way by Maxilim software for both sides. The cranial part of the condylar process dissected by the C-plane was defined as the condyle and represented the VOI of this study. In order to compute the landmark identification error and the consequent error in the determination of VOI, C-points were identified twice on the same 3D model with an interval of 4 weeks by the same operator to prevent analysis bias. The Euclidean distance between the two identified C-points on each side was calculated. Thus, a total of four C-planes and four VOIs were created for each patient, two on each side ( Fig. 1 B).
Step 3. Semi-automated virtual recontouring of condyles
After the determination of the VOI, the task was to accentuate the contour of the VOI globally using the differences in greyscale between the voxels that are adjacent to the VOI outline. The original DICOM file of the CBCT dataset was loaded into the image processing program ImageJ (National Institutes of Health, USA). By scrolling through the coronal slices of the condyles, a suitable greyscale cut-off value was selected manually in such a way that most parts of the condylar contour were made visible while the interarticular space in relation to the glenoid fossa was maintained. Using this greyscale cut-off value, the contour of VOI was enhanced automatically. Consequently, the VOI was exported in a med-file ( Fig. 2 A).
