974. Correlation of Magnetic Resonance Imaging (MRI) and Bone Histopathology in the Diagnosis of Foot Osteomyelitis in Diabetic Patients.

Abstract

Background. Inaccurate diagnosis of osteomyelitis (OM) in the setting of foot infection may lead to inappropriate therapy. Recommended imaging to discern OM is magnetic resonance imaging (MRI), but studies have shown varied sensitivity and specificity. The goal of our study is to determine the accuracy of MRI by comparing with bone biopsy histopathology. Methods. Retrospective chart review of adult patients with diabetes mellitus and foot infection with concern for OM. They had MRI with contrast and subsequent bone biopsy, resection or amputation with histologic examination. A second, blinded musculoskeletal radiologist reviewed all reports in which there was diagnostic uncertainty. Exclusion criteria: foot tumor or recent foot surgery or injury. Results. 63 eligible patients with 87 episodes of foot infection (50 males, 13 females, mean age 55 years) met study criteria. Time gap between biopsy and MRI was mean 6 days (range 0 to 23 days). 72 MRIs were positive (including 7 highly suspicious MRI reads), 11 negative and 4 indeterminate. 53 of the histologic results were positive for OM and 34 were negative. Two scenarios were created by placing 4 indeterminate MRI reads as positive in scenario 1 and negative in scenario 2. Scenario 1 Scenario 2 Sensitivity % 98 96 Specificity % 29 38 Positive predictive value % 68 71 Negative predictive value % 91 87 Positive likelihood ratio 1.39 1.56 Negative likelihood ratio 0.06 0.10 Accuracy % 71 73 Disease prevalence % 61 61 Fisher exact test two-tailed p-value 0.0003 0.0001 In scenario 1, there was only 1 false negative MRI but 24 false positives. In scenario 2, there were 2 false negative MRI and 21 false positives. Out of 74 bone specimens of those not on recent antimicrobial therapy, bone cultures were positive for pathogenic organisms on 22 specimens without histologic evidence of OM. Conclusion. In both the above scenarios, sensitivity was high. Specificity was found to be lower than the reported specificity of MRI in the diagnosis of diabetic foot OM, which has been reported as 40% to 100%. Our data demonstrate that MRI is useful to rule out OM if negative but performs poorly in ruling in OM. Bone specimen for histopathology is essential to confirm the diagnosis. Bone culture is helpful to guide antimicrobial therapy rather than diagnosing OM. Prospective studies are warranted.