In this single-patient pilot study, we report details of the used methodology as well as preliminary results from the first patient, and discuss the systems' strengths, as well as challenges to be addressed in the future.
Ethical approval was obtained by the Charité - Universitätsmedizin Berlin ethics committee (EA1/062/21). This study was performed in accordance with the Declaration of Helsinki. Patients were considered for participation, if they were planned for reconstruction of one-segmental mandibular defects with fibula free flaps. After receiving detailed study information by the principal investigator prior to treatment, participants provided oral and written informed consent to participate in this study. Here, a 69-year-old male patient without relevant comorbidities or prior therapy undergoing primary reconstruction of a one segmental, left-sided mandibular defect with a fibula free flap following resection of oral squamous cell carcinoma participated. The patient had a prosthetically restored, full dentition to both second molars in the upper jaw and a shortened prosthetically restored dentition in the lower jaw following resection (to the first premolar in the third, and to the first molar in the fourth quadrant.
To mark the free flap and native mandibula segment in vivo positions for later analysis, a minimum of five biocompatible, unalloyed 1 mm F560 tantalum beads (X-medics A/S, Brøndby, Denmark) were implanted in the anterior mandibular segment (AM), fibula free flap (FFF) and posterior mandibular segment (PM)) in three dimensions intraoperatively as previously described for pelvic models. The beads are biocompatible and do not need to be removed. Beads were randomly positioned in previously drilled 1.2 mm bone cavities across the bone segment surfaces and secured using bone wax. The free flap was fixated using 3D-printed, patient-specific (3D-PSI) titanium plates as well as corresponding cutting and drilling guides (KLS Martin SE & Co. KG, Tuttlingen, Germany). A 2.0 mm reconstruction plate and two 1.0 mm mini plates were used for the posterior and anterior flap fixation, respectively, as previously described. Further surgical procedures were performed in typical manner. The reconstruction results and the bead positions were postoperatively evaluated using orthopantomography (OPTG) and cone-beam computed tomography (CBCT).
Fluoroscopy was conducted three weeks postoperatively with the patient in an upright sitting position. This time point was determined based on clinical criteria: patients are typically mobile and recovered enough after three weeks to undergo the described procedure, while bone healing has commonly not advanced to mineralization stage, thereby allowing for quantification of micromovements. To correct image distortion, the fluoroscopic system was calibrated using a Perspex calibration box (BAATEngineering B.V., Hengelo, The Netherlands) prior to image acquisition as described in previous studies.
Following instructions and practice, the patients performed maximum jaw movements, starting from a resting position: (a) maximum mouth opening and (b) maximum intercuspation. Fluoroscopic imaging was performed using a C-arm fluoroscope (Philips BV Pulsera, Philips Medical Systems DMC GmbH, Hamburg, Germany) with an acquisition frequency of 30 Hz. Each measurement was recorded three times (Fig. 1). During fluoroscopy, the patient received a total effective dose of 0.4 mSv (dose-area-product: 2830 mGy*cm).
Tantalum beads were identified in the postoperative CBCT and OPTG and assigned to the respective segments. Using the software Amira (Amira, Visage Imaging, Berlin, Germany), metallic objects were displayed in the CBCT based on thresholds and the tantalum beads were extracted. The position of each individual tantalum bead was defined and merged within the coordinate system of a separate 3D frame assigned to each of the respective segments (AM, FFF or PM). For each measurement, specific image files corresponding to the resting position and a maximum movement were extracted from the fluoroscopy. These image files and 3D frames were transferred to the commercially available software Model-based RSA (Medis specials B.V., Netherlands) for all subsequent analyses. The 3D frames assigned to the respective segments (AM, FFF or PM) could subsequently be superimposed on the scene file in x (anterior-posterior)-, y (superior-inferior)- and z (medio-lateral)-axis. Considering that the z-axis represents the sensitive out of image-plane direction (error in relative position and orientation up to 0.8 mm and 0.6°), translational changes in this direction were approximated to 0 mm. Rotational changes around the x- and y-axes were also neglected due to the aforementioned out of image-plane error. Thereby, these bead-derived 3D-objects acted as segment surrogates to facilitate quantification. Finally, the relative spatial position change between the 3D frames was measured for translational and rotational movements in the respective axes, with negative values indicating counterclockwise rotations or diverging translations (Fig. 2).
Data management and descriptive statistical calculations were performed in Microsoft Excel (Microsoft Corp., Redmond, WA, USA). Visualizations were created using the Python Programming Language (Google Colab).