(P 283) USE OF MIXED REALITY IN PELVIC BONE TUMOR RESECTION: FIVE CASE STUDY

Objective: Precise preoperative planning and accurate resection are crucial in musculoskeletal tumor surgeries. Inaccurate planning or resection can lead to tumor remnants, resulting in recurrence, or functional impairment. Current preoperative planning generally relies on 2D data from X-rays, CT scans, and MRIs, but 3D data and compatible surgical support devices offer more accurate planning and resection, especially for the anatomically complex pelvis.
Currently, orthopedic surgical support devices include navigation systems, 3D printed models, and patient-specific cutting guides. However, these devices are often expensive and time-consuming to prepare. Recent advancements in Virtual Reality (VR) and Mixed Reality (MR) technologies have enabled relatively inexpensive and quick 3D visualization. This study investigates the usefulness of MR in the resection of pelvic bone tumors.

Methods: From April 2023 to June 2024, we performed surgeries on five patients with pelvic bone tumors using MR with patient’s consent. Preoperative CT and MRI images were processed using medical imaging software (Ziostation) to extract STL data of the pelvic bone and tumor. Using our custom-developed software (Bone Simulator), 3D surgical planning was conducted, with a resection margin of at least 20 mm from the tumor. 3D holograms were created using the Holoeyes MD service. During surgery, the holograms were viewed with a head-mounted display (Hololens2) for reference, and bone cutting was performed with a bone saw or chisel. For two cases involving the acetabulum, navigation was also used for partial acetabular resection and total hip arthroplasty (THA). We measured discrepancies between the planned resection plane and the actual resection plane.

Results: Histological types included two cases of undifferentiated sarcoma, one case of osteosarcoma, and two cases of solitary bone metastasis from renal cancer and hepatic cancer. Surgical procedures included partial resection of the acetabulum with THA in two cases, partial resection of the ilium in two cases, and hemipelvectomy in one case. Operating time ranged from 97 minutes to 475 minutes, and blood loss ranged from 155 ml to 1685 ml. One case had a positive surgical margin. Local recurrence was observed in two cases: the positive margin case and the liver cancer bone metastasis case. The maximum discrepancy between the planned resection plane and the actual resection plane ranged from 3 mm to 18 mm.

Conclusion: MR technology provides an inexpensive and rapid method for visualizing the tumor extent and preoperative planning in three dimensions, allowing for intraoperative reference. However, our study found a maximum discrepancy of 18 mm between the planned and actual resection planes, while previous studies using optical navigation for bone tumor resection reported a discrepancy of 1.6 mm between preoperative planning and actual resection. Although this discrepancy is within the planned surgical margin of less than 20 mm, it indicates that the accuracy of MR is still insufficient. To use MR effectively as navigation for pelvic bone tumor resection, it is necessary to develop a method to synchronize the hologram with the affected bone.