The challenges in differentiating tuberculous from pyogenic spondylitis using magnetic resonance imaging

Purpose: To analyze the diagnostic value of MRI for distinguishing tuberculous from pyogenic spondylitis confirmed by histology results and to determine the cut-off point score of MRI. Subjects and methods: Observational analytic design with a cross-sectional approach. Data were collected from radiology and pathology anatomy medical records, therefore no informed consent was obtained. We utilized diagnostic tests using 2×2 tables and receiver operating characteristic curve to obtain the value of the sensitivity, specificity, accuracy, and area under curve (AUC) of MRI. Twenty-eight samples were selected, consisting of 20 samples of tuberculous spondylitis and 8 samples of pyogenic spondylitis. Results: The radiological diagnosis using MRI had a sensitivity of 85%, specificity of 87.5%, positive predictive value of 94.4%, negative predictive value of 70%, and accuracy of 85.7%. AUC value was 91.6%, p = 0.001. MRI with a score ≥4 had sensitivity of 90%, specificity of 87.5%, positive predictive value of 94.7%, negative predictive value of 77.8%, and accuracy of 89.2%. Conclusion: MRI had a high diagnostic value with AUC value of 91.6% for distinguishing tuberculous from pyogenic spondylitis, confirmed by histology results. MRI, with a score of ≥4, had a higher diagnostic value compared with the reading of MRI without using scores.