MRI-based Surgical Planning for Lumbar Spinal Stenosis

摘要

The most common reason for spinal surgery in elderly patients is lumbarspinal stenosis(LSS). For LSS, treatment decisions based on clinical andradiological information as well as personal experience of the surgeon showslarge variance. Thus a standardized support system is of high value for a moreobjective and reproducible decision. In this work, we develop an automatedalgorithm to localize the stenosis causing the symptoms of the patient inmagnetic resonance imaging (MRI). With 22 MRI features of each of five spinallevels of 321 patients, we show it is possible to predict the location oflesion triggering the symptoms. To support this hypothesis, we conduct anautomated analysis of labeled and unlabeled MRI scans extracted from 788patients. We confirm quantitatively the importance of radiological informationand provide an algorithmic pipeline for working with raw MRI scans.