The purpose of this study was to determine which of the most commonly used flaps restore contour more accurately in mandibular body reconstructions using conformance analyses and virtual measurements.
Using normal computed tomography (CT) scans and a 3D software, mandibular body defects were virtually created. “Single shot” and osteotomized fibula flaps (SS-FF and O-FF), iliac crest flaps (ICF) and scapular tip flaps (STF) were digitally harvested and coregistered to reconstruct those defects. Conformance analyses were performed by calculating the root mean square (RMS) for overall and contour conformance.
Ten patients normal CT scans were included. The STF demonstrated improved overall conformance compared with the ICF, the SS-FF and the O-FF (RMS = 2.03 mm vs 4.53 mm vs 2.76 vs 2.37 mm, respectively; p<.001). Similar trends were seen for contour conformance in STF compared with the ICF and the SS-FF (RMS = 2.48 mm vs 4.50 mm vs 3.28 mm, respectively), whereas the O-FF performed better than STF (RMS = 1.85 mm vs 2.48 mm; p<.001).
The osseous component of the STF resembles the mandibular body more accurately than the one in the ICF and FF without the need for an osteotomy. Future clinical studies can help to elucidate the clinical impact of these virtual findings.
One of the objectives while reconstructing mandibular body defects is to achieve symmetrical contour of the mandible. The fibula flap (FF) and iliac crest flap (ICF) have been historically the most commonly used options to reconstruct mandibular defects. Our tertiary care center also uses the scapular tip flap (STF), especially to reconstruct mandibular body defects where osteotomies are not always needed. The STF provides a donor site with a favorable long-term morbidity profile, reliable anatomy, large caliber pedicle, amenability for dental implants, vessels which are less affected by atherosclerotic disease, and allows early ambulation which is critical in the elderly population. , In addition, from our experience, the scapular bone closely replicates the curvature of the mandibular body. There have been conformance studies which have analyzed the morphological similarities between the scapular bone with the palate and orbital floor. These virtual planning studies have contributed to support the use of the STF to reconstruct palatal and orbital defects.
In a similar manner, and based on our intraoperative clinical and postoperative imaging findings using the STF, we endeavored to confirm our clinical observation by comparing the morphological similarities between the mandibular body and the bone segment utilized in the FF, ICF, and STF. Using 3-dimensional (3D) conformance measurements, we aimed to determine which reconstructive option resembles more accurately the shape of the mandibular body, using conformance as the main outcome. There are no previous studies comparing the contour of the different flaps and these specific mandibular defects.
Institutional research ethics board approval was obtained for this project. The study population comprised anonymized patients computed tomography (CT) scans randomly selected from an institutional imaging database between January 1, 2005, and September 30, 2015. Patients were included if they had high-resolution CT imaging of the head and neck, chest, pelvis, and lower limbs. Patients were excluded from the study if there was evidence of any skeletal abnormality, prior surgical intervention and/or severe image artifacts. We used normal CT scans (without any abnormality) to have unchanged anatomy to compare with, as using images with mandibular tumors, osteoradionecrosis or other anomalies would have precluded this.
Data Collection and Study Variables
Using a technique developed in the Guided Therapeutics (GTx) Laboratory at the University Health Network, a conformance analysis was performed. Briefly, the whole method consists in importing images into a computer workstation, where bone and soft tissues are differentiated by intensity threshold. Regions of bone specific to the mandibular body, fibula, iliac crest, and scapula were then defined. The external surfaces of mandibular body and the different bony flaps were then manually coregistered. Volume renderings were compared, by creation of conformance maps, which express the distance between each point of bony flap with respect to the closest point of the mandibular bone. The resulting distances were presented in a color-coded map depicting the morphologic discrepancy between the flaps and the mandibular body. This determined that when a portion of the external surface of the flap was misaligned as compared with the mandible, the conformance that resulted was negatively affected.
The analysis was carried out using STL files, comparing surface differences between 2 models, which indicates conformance of 2 structures. STL files were created using CT image (DICOM) data, and then by using Mimics version 18.0 (Materialise ® , Leuven, Belgium) 3D visualization software. For each patient, the mandible, scapulae, iliac crest, and fibulas were virtually extracted.
Each mandible was divided in 3 segments (symphysis-parasymphysis and 2 mandibular body–which included a short segment of the angle), and mandibular body-angle defects were created. The scapulae, iliac crest, and fibulas were segmented and the STF, ICF, and FF were digitally harvested based on known surgical principles. The STF was segmented including 5 cm of the tip and 8 cm (cranio-caudal) x 2 cm (medio-lateral) of the lateral border. The iliac crest was virtually harvested including 10 cm along the crest and 4 cm of cranio-caudal thickness. The fibula segment was extracted as a 10 cm segment at the midportion of the fibula. The segmented bones were virtually superimposed to the ipsilateral body-ramus of the mandible ( Figs 1 A,B). For FFs, we decided to perform an additional analysis and include it as a “straight shot” (SS-F) (no osteotomy), and those with a single osteotomy (O-FF) in the point of maximal angulation of the mandibular body. The 2 fibula segments were placed to most similarly align with the latero-inferior aspect of the mandible. The exceeding portions of the flaps were trimmed to obtain the best achievable morphologic similarity to the mandible, always respecting the most likely entry point of the vascular pedicle ( Figs 1 C and 2 ). Two surgeons performed these virtual manipulations independently to avoid biases. Of note, ipsilateral scapulae were chosen as donor sites as in practice it is more common to use the ipsilateral side for patient positioning purposes (both the neck extension and rotation of the torso face the same side).
Overall conformance between each bony flap and the matched bony-angle segment of the mandible was measured as the mean root-mean-square (RMS) conformance distance (mm) ( Fig 1 D). RMS conformance distance is a measure of morphologic similarity as described in prior studies, and its value is inversely proportional to the morphologic similarity between structures. In addition, contour conformance was calculated, and was defined as the conformance between the lateral surface of each flap already positioned and the lateral aspect of the mandible (comparison of the lateral aspects only) ( Fig 1 E). Overall conformance expresses how all the surfaces of the bony flap align with all the mandibular surfaces. Contour conformance was also included to focus on clinically relevant surface in terms of facial contour restoration. Average of median, maximal, and minimal individual conformance data were extracted for each reconstruction.
Conformance results were grouped into 3 classes in accordance with the distance of virtual contact points between the different flaps and the mandibular body: Class 1 included points ≤ 1 mm from the mandible; Class 2 between 1-3 mm; Class 3 ≥ 3 mm. The percentage distribution in Classes 1 to 3 was measured for each conformance simulation ( Fig 1 F-H and Fig 2 ).