Developing Semiautomated Methods to Measure Pre- and Postoperative Syrinx Volumes.

Abstract

Neurosurgeons evaluate MRI scans to document whether surgical treatment has reduced syrinx size. Manual measurement of syrinx volume is time-consuming and potentially introduces operator error and bias. Developing convenient semiautomated volumetric analysis methods may encourage their clinical implementation and improve syringomyelia monitoring. We analyzed 30 SPGR axial MRI scans from 15 pre- and postoperative Chiari I and syringomyelia patients using two semiautomated (SCAT and 3DQI) methods and a manual Cavalieri (CAV) method. Patients' spinal cord and syrinx volumes pre- and postoperatively were compared by paired t-test. A decrease in syrinx volume (mm3) after surgery was detected across all methods. Mean syrinx volume (± SD) measured by CAV (n = 30) was, preoperatively, 4515 mm3 ± 3720, postoperatively 1109 ± 1469; (p = 0.0004). SCAT was, pre, 4584 ± 3826, post, 1064 ± 1465; (p = 0.0007) and 3DQI was, pre, 4027 ± 3805, post, 819 ± 1242; (p = 0.001). 3DQI and CAV detected similar mean spinal cord volumes before (p = 0.53) and after surgery (p = 0.23), but SCAT volumes differed significantly (p = 0.005, p = 0.0001). The SCAT and 3DQI semiautomated methods recorded surgically related syrinx volume changes efficiently and with enough accuracy for clinical decision-making and research studies.