A novel computer algorithm allows for volumetric and cross-sectional area analysis of indirect decompression following transpsoas lumbar arthrodesis despite variations in MRI technique

Abstract

Many patients present for neurosurgical spine evaluation with MRI studies conducted at facilities outside of the treating medical center. These images often vary widely in technique, for example, variation in slice thickness, number of slices, and gantry angle. While these images may be sufficient in conjunction with a physical exam to make surgical evaluations, we have found they are often incapable of being used for objective post-operative volumetric comparisons. In order to overcome this, we created a computer program that compensates for these variations in MRI technique. For this study, we examined patients who had undergone outside MRI pre-operatively and were deemed appropriate for a lateral retroperitoneal transpsoas lumbar interbody arthrodesis procedure. Volumetric analysis was performed on sagittal and axial T2-weighted pre- and post-operative MRI. The percentage change of central canal volume and foraminal area was calculated for each level. The authors identified five levels with MRI sufficient for volumetric analysis and eight levels (16 foramina) sufficient for foraminal cross-sectional analysis. Through use of our computer algorithm, average central canal volume and foraminal cross-sectional area was calculated to increase by 32.8% and 67.6% respectively following the procedure. These results are consistent with previous study findings and support the idea that restoration of the anterior column via a lateral approach can result in significant indirect decompression of the neural elements. Additionally, the novel algorithm created and used for this study suggests that it can achieve quick measurement and comparison of MRI studies despite variations in pre- and post-operative technique.