Medical Student Carle Illinois College of Medicine, Minnesota, United States
Introduction: Properly assessing the size of glioblastoma multiforme (GBM) pre-and postoperatively influences the calculated extent of resection and decisions in further management. While certain image analysis software can provide volumetric segmentations of a given tumor, routine MRI reporting only allows 2-dimensional diameter measurements in the axial, coronal, or sagittal planes. We sought to determine if volumes calculated from these diameters accurately represent 3D volumetric segmentations.
Methods: We retrospectively identified 16 patients that underwent resection of grade IV gliomas at our institution. We utilized Velocity dosimetry software (Varian) to volumetrically segment pre-operative T1 gadolinium-enhanced MRI scans. We utilized the native picture archiving and communication system at the hospital to identify the maximum diameter of the tumor in three planes (axial, coronal, sagittal). These diameters were then converted into spherical volumes (v = 4/3*𝛑*r^3) and plotted against the corresponding segmented tumor volumes.
Results: 16 patients were identified with a mean age of 57.1 (SD = 14.9) years. The mean size of pre-operative tumor was 30.1 (SD = 22.0) cm3. The maximum diameter was 5.59 cm. Segmented tumor volumes were plotted against the corresponding spherical volumes obtained from axial, coronal, or sagittal diameters, respectively. Linear regression for the axial (R^2 = 0.92), coronal (R^2 = 0.88), and sagittal (R^2 = 0.73) plots were separately calculated.
Conclusion : Axial, coronal, and sagittal diameters can serve to reasonably correlate with segmented GBM tumor volume from T1 gadolinium-enhanced MRI. Future work will explore how deviations from a spherical tumor shape perturbs this correlation.
