A deep learning-based technique for the diagnosis of epidural spinal cord compression on thoracolumbar CT

• Published in: European spine journal Vol. 32; no. 11; pp. 3815 - 3824
• Main Authors: Hallinan, James Thomas Patrick Decourcy, Zhu, Lei, Tan, Hui Wen Natalie, Hui, Si Jian, Lim, Xinyi, Ong, Bryan Wei Loong, Ong, Han Yang, Eide, Sterling Ellis, Cheng, Amanda J. L., Ge, Shuliang, Kuah, Tricia, Lim, Shi Wei Desmond, Low, Xi Zhen, Teo, Ee Chin, Yap, Qai Ven, Chan, Yiong Huak, Kumar, Naresh, Vellayappan, Balamurugan A., Ooi, Beng Chin, Quek, Swee Tian, Makmur, Andrew, Tan, Jiong Hao
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.11.2023; Springer Nature B.V
• Subjects: Compression; Deep learning; Diagnosis; Epidural; Magnetic resonance imaging; Medicine; Medicine & Public Health; Neurosurgery; Original Article; Spinal cord; Spine; Surgical Orthopedics; Metastatic epidural spinal cord compression; Deep learning; CT; Epidural spinal cord compression; MRI; Artificial intelligence
• ISSN: 0940-6719; 1432-0932
• DOI: 10.1007/s00586-023-07706-4

Purpose To develop a deep learning (DL) model for epidural spinal cord compression (ESCC) on CT, which will aid earlier ESCC diagnosis for less experienced clinicians. Methods We retrospectively collected CT and MRI data from adult patients with suspected ESCC at a tertiary referral institute from 2007 till 2020. A total of 183 patients were used for training/validation of the DL model. A separate test set of 40 patients was used for DL model evaluation and comprised 60 staging CT and matched MRI scans performed with an interval of up to 2 months. DL model performance was compared to eight readers: one musculoskeletal radiologist, two body radiologists, one spine surgeon, and four trainee spine surgeons. Diagnostic performance was evaluated using inter-rater agreement, sensitivity, specificity and AUC. Results Overall, 3115 axial CT slices were assessed. The DL model showed high kappa of 0.872 for normal, low and high-grade ESCC (trichotomous), which was superior compared to a body radiologist (R4, κ = 0.667) and all four trainee spine surgeons (κ range = 0.625–0.838)(all p < 0.001). In addition, for dichotomous normal versus any grade of ESCC detection, the DL model showed high kappa (κ = 0.879), sensitivity (91.82), specificity (92.01) and AUC (0.919), with the latter AUC superior to all readers (AUC range = 0.732–0.859, all p < 0.001). Conclusion A deep learning model for the objective assessment of ESCC on CT had comparable or superior performance to radiologists and spine surgeons. Earlier diagnosis of ESCC on CT could reduce treatment delays, which are associated with poor outcomes, increased costs, and reduced survival.