Improving the definition of glioblastoma resection borders with intraoperative cone-beam computed tomography

To assess the effectiveness of intraoperative cone-beam computed tomography (CBCT) in defining tumor resection margins during glioblastoma surgery.

This study included glioblastoma patients who underwent surgery between 2017 and 2023. All patients underwent pre- and postoperative magnetic resonance imaging (MRI) and intraoperative CBCT before and after craniotomy. Tumors and cavities were segmented by two neurosurgeons, and fusion images were generated. Interrater and intermethod agreements were evaluated using intraclass correlation coefficients (ICC), and volume shapes were compared using Dice and Hausdorff similarity metrics.

A total of 21 glioblastoma tumors and 20 cavities were analyzed. The median segmented volume was 32.8 cm³ (CBCT) versus 34.4 cm³ (MRI), with excellent agreement between modalities (ICC ≥0.90). The dice index was 0.79, and the Hausdorff 95% index was 14% relative to tumor diameter. Clinical outcomes were analyzed for 18 cases, where the total resection of contrast-enhancing tumor was considered feasible preoperatively. Intraoperative CBCT suggested no residual tumor in 13 (72%) cases, all of which were confirmed as total removal on postoperative MRI. CBCT detected suspected residual contrast-enhancing lesions in 3 (17%) cases (mean volume: 1.05 cm³, range: 0.42-1.59 cm³); additional resection achieved total removal in all cases. Conversely, postoperative MRI identified small residual tumors in 2 (11%) cases (0.73-1.57 cm³) that were not detected by intraoperative CBCT.

CBCT provides intraoperative visualization of glioblastomas, postcraniotomy bleeding, and cavities, enhancing resection margin definition. It is a promising tool for improving glioblastoma resection in a hybrid operating room.