The digitally planned mandibular reconstructions to correct defects with MFFF due to oncologic resections showed high accuracy and satisfactory results. The adequate functional and aesthetical results, with a high rate of facial symmetry achievement, are clinical indicators of the complete DSP's accuracy independently of the number of compromised mandibular segments.
The virtues of DSP and custom-made guides and plates for mandibular reconstructive surgery have been explored through different approaches; descriptive outcomes, comparative assessment with conventional surgical planning, presurgical and postsurgical accuracy comparisons, among others (6, 10, 14–16, 20, 22, 23, 25, 26). The general conclusion has been that CAD/CAM techniques improve efficiency during the preoperative planning and through the surgery given by the accuracy of fibular and mandibular cutting guides with a positive impact on the contouring symmetry more challenging to achieve with the traditional freehand technique (10, 14, 16, 20, 23, 25).
Oncological surgical management aims to resect the tumoral tissue while reconstructing the resultant defect. Mandibulectomies frequently create extensive defects leading to significant aesthetic deformities and functional impairment that can affect patients' speech and the ability to swallow (5, 7, 24, 27–29). Hence, it is expected that more prominent defects with more than two compromised mandibular segments would have worse or unpredictable outcomes. Additionally, the longer the fibula flap must be to restore the anatomy, the higher the donor site complications and worse outcomes.
Our report's novelty lies in the analysis of surgical variables and mid and long-term outcomes according to the number of resected and reconstructed mandibular segments.
Half of our patients presented mild defect sizes; in 55.5%, only one mandibular segment was involved, which could have favored the overall satisfactory outcomes. As expected, we observed a proportional increase of the resected mandibular volumes and harvested fibula segment lengths with the number of compromised mandibular segments. However, the lack of differences among those measurements could indicate that the DSP's precise calculations avoided unnecessary bone resections and fibula harvesting, minimizing the defect size and optimizing the tissular use.
DSP facilitates both the mandibular resections and the fibular harvest without the necessity for intraoperative adjustments nor freehand contouring, preventing prolonged vascular interruption to the flap (4, 23, 24). Previous studies have estimated the total surgical time and the MFFF ischemia time for both conventional and computer-aid planned mandibular reconstruction, showing the latter's advantage in terms of time-saving (1, 5, 25, 28). Bosc et al. (2016)(25) and Saini et al. (2019)(15) estimated a mean total operating time of 422 and 639 minutes, respectively, using computer-aid planning (25). Tasitano et al. (2016)(5) and Blanc et al. (2019)(20) showed a mean surgical time of 550.5 and 529 minutes for the conventionally planned reconstructions, compared to 435 and 441 minutes with DSP, respectively. In our case, the median surgical time for the whole cohort was 380 minutes, ranging from 340 minutes for one mandibular segment to 625 minutes for more complicated procedures of three compromised segments.
In terms of MFFF ischemia times, Succo et al. (2014)(1) and Rustemeyer et al. (2015)(28) observed a mean ischemia time of 105 and 98.6 minutes in the conventional group compared to 75 and 70.7 minutes for the DSP, respectively. In our cohort, the observed median ischemia time was 76 minutes that ranged from 60 to 113 minutes according to the mandibular segments, comparable to those previously reported for DSP.
The custom-made guides optimized the operative times and facilitated the planned osteotomies to be replicated during surgery (5, 6, 22). In our study, most of the surgeries were accurate, showing a good correlation and excellent fit between the preoperative virtual 3D models and the procedure's results. Bony contacts of the fibula segments were adequate, and the flaps fitted precisely into the defects. We obtained 95.5% intraoperative accuracy that, in the end, rendered excellent results in the mid and long-term. Four patients required minor onsite modifications, probably due to intraoperative changes on the resection extension because of tumor expansion or reduction due to neoadjuvant therapy.
The mandibular reconstructions' primary goals are achieving facial symmetry or an acceptable aesthetic appearance and an adequate function (5, 10, 27–29). In our sample, the overall functional results indicated that more than 60% of the followed patients had preserved speech and masticatory function. Additionally, the general appearance was excellent in 74% of the patients, and the facial symmetry was achieved in 92%. These results were not affected by the number of compromised mandibular segments, indicating that good functional and esthetic outcomes can be expected with computer-aided surgical planning, even in more complicated cases.
Our technical solution for the full DSP and the customized design of the guides and implants in our patients showed precision and reliability during the actual procedure. Furthermore, the stratified analysis by the number of compromised mandibular segments showed no significant differences in the overall outcomes, indicating that our DSP was accurate and produced suitable results independently of the procedure's complexity.
This study's limitations include its retrospective nature and the consequent selection bias, and the inability to carry out a longitudinal follow-up due to poor adherence of patients to the standardized postoperative clinical evaluations. Since most of our cases are planned and performed with DSP, we did not compare our findings with conventionally planned mandibular reconstruction. Even though this is a widely used technology for complex surgical procedures in our setting, there are no similar literature reports regarding sample size, length of follow-up, and demographics.