| Abstract: |
Study Design. This is a retrospective, cross-sectional study. Objective. The primary aim was to identify the diagnostic yield of spine magnetic resonance imaging (MRI) in detecting malignant pathology in cancer patients with back pain. We also sought to evaluate the role of MRI extent (i.e. regional vs. total) in identifying malignant pathology. Summary of Background Data. No prior study has systematically investigated the yield of spine MRI in a large cohort of cancer patients. Methods. Spine MRI reports from 2017 to 2021 for back pain (acute and nonspecified chronicity) in cancer patients were reviewed to identify clinically relevant findings: malignant (1) epidural, (2) leptomeningeal, (3) intramedullary, (4) osseous disease, and (5) fracture. Logistic regression was used to evaluate the association between MRI extent and the presence of cancer-related findings. For patients with multiple MRIs, short-interval scans (≤4 mo) were evaluated to assess the yield of repeat imaging. Results. At least one cancer-related finding was identified on 52% of 5989 spine MRIs ordered for back pain and 57% of 1130 spine MRIs ordered specifically for acute back pain. The most common pathology was malignant osseous disease (2545; 43%). Across all five categories, most findings (77%-89%) were new/progressive. Odds of identifying a finding were significantly higher with total versus regional spine MRIs (P<0.001). Although only 14 patients had a positive regional MRI followed shortly by a positive total spine MRI, most of these repeat total spine MRIs (78%) identified findings outside the scope of the initial regional scan. Twenty-one patients had both computed tomography and MRI within 30 days of each other; eight (38%) had compression fractures appreciated on MRI but not on computed tomography. Conclusions. Our findings suggest imaging the total spine in cancer patients with back pain given higher odds of identifying malignant pathology and instances of capturing otherwise not visualized disease. Further work is warranted to confirm these findings. © 2024 Wolters Kluwer Health. All rights reserved. |
