Data-Driven Comparison Of Acute Imaging Biomarkers For Spinal Cord Injury: A Prospective TRACK-SCI Pilot Study (S42.006)

Objective: To pilot-test the relationship between acute imaging measures after SCI, and determine the measures most closely associated with impaired neurologic recovery. Background: The validation of early spinal cord injury biomarkers is essential for accurate prognosis and clinical decision-making. Our team recently performed a retrospective study of SCI patients (N=95) to determine the relationship between acute MR biomarkers and their individual validity in predicting neurologic impairment. Here we extended our study of the predictive validity of MR biomarkers to a prospective SCI cohort. Design/Methods: Thirty-one patients from the Transforming Research and Clinical Knowledge in Spinal Cord Injury (TRACK-SCI) prospective trial at Zuckerberg San Francisco General Hospital were evaluated. Length of hyperintense T2 signal, sagittal grade, Brain and Spinal Injury Center (BASIC) axial score, maximum canal compromise (MCC), and maximum spinal cord compression (MSCC) were assessed. Neurologic impairment was determined using the American Spinal Injury Association Impairment Scale (AIS) at discharge. The relationship between MR measures was assessed using non-linear principal components analysis (NL-PCA). Pilot/feasibility correlation analyses were performed to compare MR measures and NL-PCA scores to AIS grade. Results: NL-PCA identified two principal components, which accounted for 79.7% of the variance. The first principal component (PC1) was comprised of the intrinsic cord abnormality measures (sagittal grade, BASIC score, and injury length) while the second component (PC2) consisted of the extrinsic compression measures (MCC and MSCC). Correlation effect size was highest in the intrinsic abnormality measures (mean = 0.379) compared to extrinsic (mean = 0.152). Observed power ranged from 0.48–0.66, and power calculations revealed that approximate sample size of 87 patients will be needed to achieve power of 0.8. Conclusions: These preliminary results are consistent with our prior retrospective findings suggesting that intrinsic MR measures may have greater value than extrinsic MR measures in predicting early neurologic impairment after spinal cord injury. Study Supported by: Craig H. Neilsen Foundation Special Project Award (#UCSF 2016-2020) to MSB; NIH R01 NS067092, Wings for Life WFL-US-006/14, Veterans Affairs 1I01RX002245-01, and Craig H. Neilsen Foundation Award to ARF. Disclosure: Dr. Huie has nothing to disclose. Dr. Talbott has nothing to disclose. Dr. Singh has nothing to disclose. Dr. Duong Fernandez has nothing to disclose. Dr. Tsolinas has nothing to disclose. Dr. Pascual has nothing to disclose. Dr. Pan has nothing to disclose. Dr. Kyritsis has nothing to disclose. Dr. Torres has nothing to disclose. Dr. Beattie has nothing to disclose. Dr. Bresnahan has nothing to disclose. Dr. Ferguson has nothing to disclose. Dr. Dhall has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with Depuy Synthes, Globus Medical. Dr. Whetstone has nothing to disclose.