RTHP-39. SPINAL CORD DELINEATION BASED ON CT MYELOGRAM VERSUS T2 MRI IN SPINE SBRT CASES

PURPOSE/OBJECTIVE(S) Spinal cord (SC) delineation is critical in spine SBRT. While underestimating cord can lead to myelitis, overestimating SC may compromise PTV coverage, resulting in higher rates of tumor recurrence. CT myelogram as part of CT simulation is a radiological means of better visualizing the SC and minimizing MRI fusion error. The aim of this study is to compare SC contours based on CT simulation with myelogram and with fused MRI. MATERIALS/METHODS Eight patients with 9 spinal metastases treated with spine SBRT were evaluated. All patients had spinal MRI and myelogram performed prior to CT simulation. The spinal MRI was fused and the SC volume was contoured from the superior to the inferior endplate of the target vertebral body volume based on the (1) CT myelogram images and (2) fused T2 MRI images. Comparing MRI to myelogram-defined SC, centroid deviations (deviations in the center point of the cord) were assessed at every slice through the VB target volume, and maximum deviations of the cord were assessed; all are reported with descriptive statistics. The SC cc volume based on CT myelogram and MRI were compared using paired t-test. RESULTS Between MRI- versus myelogram- defined SC volumes, average deviations of the centroid point cord contours was 0.06cm, ranging from 0.01-0.1cm; average deviation of the outmost aspect of the contours were on average between 0.07-0.11cm. Finally, the volume of the cord was significantly smaller when defined on MRI as compared to myelogram - 2.778cc (SD=0.98) vs. 3.44cc (SD=1.43), p=0.03 - respectively There was only one case that had a SC volume based on T2 MRI that was bigger than that based on myelogram (3.5cc vs. 2.9cc). CONCLUSIONS These data suggest average deviations on MRI of approximately 0.1cm, largely undercontouring the SC volume based on MRI as compared to CT simulation with myelogram.