Assessing the predictive capability of machine learning models in determining clinical outcomes for patients with cervical spondylotic myelopathy treated with laminectomy and posterior spinal fusion

Cervical spondylotic myelopathy (CSM) results from spinal cord compression. Laminectomy with posterior spinal fusion (LPSF) treats CSM patients. This study uses machine learning models (MLMs) to predict clinical outcomes in CSM patients undergoing LPSF. A retrospective analysis was conducted on 329 CSM patients. Neurological outcomes were evaluated using the modified Japanese Orthopaedic Association (mJOA) scale. Machine learning models were used to predict clinical outcome. The RF model identified preoperative mJOA scale, age, symptom duration, and MRI signal changes as significant variables associated with poor clinical outcomes. The study highlights the effectiveness of machine learning models in predicting the clinical outcomes of CSM patients undergoing LPSF.

• 1
  Age, preoperative mJOA scale, and symptom duration were independent predictors of poor clinical outcome.
• 2
  The RF model identified preoperative mJOA scale, age, symptom duration, and MRI signal changes as the most crucial variables.
• 3
  RF demonstrated the highest accuracy of 0.922, followed by SVM at 0.901, k-NN at 0.887, and LR at 0.876.