BACKGROUND CONTEXT: Accurate and automatic MRI measurements are relevant for assessing spinal cord compression severity in degenerative cervical myelopathy (DCM) and guiding treatment. The widely-used maximum spinal cord compression (MSCC) index has limitations. Firstly, it normalizes the anteroposterior cord diameter by that above and below the compression but does not account for cord size variation along the superior-inferior axis, making MSCC sensitive to compression level. Secondly, cord shape varies across individuals, making MSCC sensitive to this variability. Thirdly, MSCC is typically calculated by an expert-rater from a single sagittal slice, which is time-consuming and prone to variability. PURPOSE: This study proposes a fully automatic pipeline to compute MSCC. DESIGN: We developed a normalization strategy for traditional MSCC (anteroposterior diameter) using a healthy adults database (n = 203) to address cord anatomy variability across individuals and evaluated additional morphometrics (transverse diameter, area, eccentricity, and solidity). PATIENT SAMPLE: DCM patient cohort of n = 120. OUTCOME MEASURES: Receiver operating characteristic (ROC) and area under the curve (AUC) were used as evaluation metrics. METHODS: We validated the method in a mild DCM patient cohort against manually derived morphometrics and predicted the therapeutic decision (operative/conservative) using a stepwise binary logistic regression incorporating demographics and clinical scores. RESULTS: The automatic and normalized MSCC measures correlated significantly with clinical scores and predicted the therapeutic decision more accurately than manual MSCC. Significant predictors included upper extremity sensory dysfunction, T2w hyperintensity, and the proposed MRI-based measures. The model achieved an area under the curve of 0.80 in receiver operating characteristic analysis. CONCLUSION: This study introduced an automatic method for computing normalized measures of cord compressions from MRIs, potentially improving therapeutic decisions in DCM patients. The method is open-source and available in Spinal Cord Toolbox v6.0 and above.

NeuroPoly Lab Institute of Biomedical Engineering Polytechnique Montreal 2500 Chem de Polytechnique Montréal H3T 1J4 Québec Canada

NeuroPoly Lab Institute of Biomedical Engineering Polytechnique Montreal 2500 Chem de Polytechnique Montréal H3T 1J4 Québec Canada; Mila Quebec AI Institute 6666 Rue Saint Urbain Montréal H2S 3H1 Québec Canada; Department of Neurosurgery Faculty of Medicine and Dentistry Palacký University Olomouc 775 15 Hněvotínská 976 3 Nová Ulice 779 00 Olomouc Czechia; Department of Neurology Faculty of Medicine and Dentistry Palacký University Olomouc 775 15 Hněvotínská 976 3 Nová Ulice 779 00 Olomouc Czechia

NeuroPoly Lab Institute of Biomedical Engineering Polytechnique Montreal 2500 Chem de Polytechnique Montréal H3T 1J4 Québec Canada; Mila Quebec AI Institute 6666 Rue Saint Urbain Montréal H2S 3H1 Québec Canada; Functional Neuroimaging Unit CRIUGM University of Montreal 4545 Queen Mary Road Montreal H3W 1W4 Quebec Canada; Centre de recherche du CHU Sainte Justine Université de Montréal 4545 Chem Queen Mary Montréal H3W 1W4 Quebec Canada

Spinal Cord Injury Center Balgrist University Hospital University of Zurich Forchstrasse 340 8008 Zurich Switzerland

Spinal Cord Injury Center Balgrist University Hospital University of Zurich Forchstrasse 340 8008 Zurich Switzerland; Department of Neurophysics Max Planck Institute for Human Cognitive and Brain Sciences Stephanstraße 1a 04103 Leipzig Germany

Spinal Cord Injury Center Balgrist University Hospital University of Zurich Forchstrasse 340 8008 Zurich Switzerland; Department of Neurophysics Max Planck Institute for Human Cognitive and Brain Sciences Stephanstraße 1a 04103 Leipzig Germany; Wellcome Trust Centre for Neuroimaging Queen Square Institute of Neurology University College London 12 Queen Square London WC1N 3AR United Kingdom

Citace poskytuje Crossref.org

Monitoring morphometric drift in lifelong learning segmentation of the spinal cord