Resumen
Metastatic epidural spinal cord compression (MESCC) is a disastrous complication of advanced malignancy, and early diagnosis is important to prevent irreversible neurological injury. MRI is the mainstay of diagnosis for MESCC, but it is expensive, and routine screening of asymptomatic patients is not feasible. Staging CT studies are performed routinely as part of the cancer diagnosis and represent an opportunity for earlier diagnosis and treatment planning. In this study, we trained deep learning models for automatic MESCC classification on staging CT studies using spine MRI and manual radiologist labels as the reference standard. On a test set, the DL models showed almost-perfect interobserver agreement for the classification of CT spine images into normal, low, and high-grade MESCC, with kappas ranging from 0.873?0.911 (p < 0.001). The DL models (lowest ? = 0.873, 95% CI 0.858?0.887) also showed superior interobserver agreement compared to two radiologists, including a specialist (? = 0.820, 95% CI 0.803?0.837) and general radiologist (? = 0.726, 95% CI 0.706?0.747), both p < 0.001.