Accuracy of virtually 3D planned resection templates in mandibular reconstruction.

Objective: Since reconstruction of composite defects in the head and neck region is a challenging and demanding problem for head and neck surgeons, surgical aids have been sought for decades. The purpose of this study was to evaluate the accuracy of prefabricated surgical resection templates used in mandibular segmental resections in comparison to the virtual surgical plan.

Materials And Methods: A prospective study was performed in 11 consecutive patients, with a primary T4 oral squamous cell carcinoma or osteoradionecrosis of the mandible. Preoperatively, a CBCT scan was acquired to delineate the size and extension of tumor invasion; a virtual patient-specific resection template was designed based on this information. Intraoperatively, the resection template was positioned on the mandible and secured using four fixation screws. Postoperatively, a CBCT scan was acquired. This scan was superimposed on the preoperative scan. Positioning of the resection template and inclination of the resection planes were evaluated on the virtual head model. In order to test the interobserver reliability of these new measurement methods, two different observers executed all measurements.

Results: The mean shift of the proximal resection templates was 3.76 mm (standard deviation [SD] 3.10 mm). For the distal resection templates, the mean shift was 3.06 mm (SD 1.57 mm) with no significant interobserver difference (ICC = 0.99). An absolute mean deviation of 8.5° (SD 5.3°) was found for the proximal resection angle and 10.4° (SD 5.0°) for the distal resection angle. Again, no significant interobserver differences were found (ICC = 0.98).

Conclusion: The resection templates used in this study proved reasonably accurate. Although the concept of virtual surgical planning aids significantly in mandibular reconstruction with microvascular free flaps, further improvement of resection accuracy is necessary for further improvement of reconstruction accuracy.