A novel MRI-based classification of spinal cord shape and CSF presence at the curve apex to assess risk of intraoperative neuromonitoring data loss with thoracic spinal deformity correction

Study design Retrospective cohort. Summary We present a simple classification system that is able to identify patients with increased odds of losing intraoperative neuromonitoring data during thoracic deformity correction. Type 3 spinal cords, with the cord deformed against the concave pedicle in the axial plane, have ×28 greater odds of losing monitoring data during surgery. Objectives Assess preoperative morphology of the spinal cord across the thoracic concavity to predict intraoperative loss of neuromonitoring data. Methods 128 consecutive patients undergoing surgical correction of a thoracic deformity with pedicle screw/rod constructs were included. Spinal cords were classified into 3 types based on the appearance of the cord on the axial-T2 MRI at the apex of the curve. Type 1 is defined as a circular/symmetric cord with visible CSF between the cord and the apical concave pedicle/vertebral body. Type 2 is a circular/oval/symmetric cord with no visible CSF between the concave pedicle and the cord. Type 3 is a spinal cord that is flattened/deformed by the apical concave pedicle or vertebral body, with no intervening CSF (Fig. 1 ). Results 128 patients were reviewed: 81 (63%) Type 1; 32 (25%) Type 2; and 12 (11.7%) Type 3 spinal cords. Lower extremity trans-cranial motor-evoked Potentials (MEPs) and/or somatosensory evoked potentials (SSEPs) were lost intraoperatively in 21 (16%) cases, with full recovery of data in 20 of those cases. On regression analysis, a Type 1 cord was protective against intraoperative data loss (OR = 0.17, p = 0.0003). Type 2 cords had no association with data loss (OR = 0.66, p = 0.49). Type 3 cords had significantly higher odds of intraoperative data loss (OR = 28.3, p