A New and Easy-to-Use Survival Score for Patients Irradiated for Metastatic Epidural Spinal Cord Compression.

PURPOSE: A survival score was created in 2008 to improve treatment personalization of patients irradiated for metastatic epidural spinal cord compression (MESCC). Since then, targeted therapies improved survival of patients with cancer, which may decrease this score's predictive value. A new score appears necessary. METHODS AND MATERIALS: Two hundred sixty-four patients receiving radiation therapy without surgery in prospective trials (2010-2021) were included. A dose-fractionation regimen plus 15 factors were analyzed: age, sex, tumor type, interval tumor diagnosis to MESCC, MESCC sites, affected vertebrae, additional bone lesions, other distant lesions (yes or no), number of organs involved by metastases, time developing motor deficits, ambulatory status, sensory function, sphincter dysfunction, pain, and distress. Six-month survival rates (%) of independent prognostic factors were divided by 10 and summed for each patient. The score was compared with the previous tool for predicting death ≤6 months and survival ≥6 months. RESULTS: In a multivariate analysis, tumor type (P = .001), other distant lesions (P < .001), and ambulatory status (P < .001) were significant. Based on 6-month survival rates, 4 groups (8-9, 10-13, 14-17, and 18 points) were created with 6-month survival rates of 12.8%, 34.7%, 62.8%, and 90.0%, respectively (version A). For version B, "other distant lesions" was replaced by "number of organs involved by metastases." Version B included 4 groups (8-10, 11-14, 15-16, and 17 points) with 6-month survival rates of 11.1%, 42.0%, 68.6%, and 91.7%, respectively. Positive predictive values to predict death ≤6 months were 87.2% (version A) and 88.9% (version B) versus 76.6% (3 groups) and 84.6% (5 groups) for the previous score. Positive predictive values to predict survival ≥6 months were 90.0% and 91.7% versus 59.0% and 64.3%. CONCLUSIONS: Both versions of the new score were more precise than the previous tool. Version B appears slightly superior to version A but requires more extensive diagnostic staging that may not be readily available when emergently treating.