Osteomyelitis of the pelvis/hips in paralyzed patients: accuracy and clinical utility of MRI.

PURPOSE: The goal of our study was to determine the accuracy and clinical utility of MRI in the diagnosis of osteomyelitis of the pelvis/hips in paralyzed patients.
METHOD: In 44 paralyzed patients, 59 consecutive MR examinations of the pelvis/hips were evaluated prospectively. Criteria for diagnosis of osteomyelitis were based on those established in previous studies of complex, nonhematogenous osteomyelitis (diabetic foot). Average follow-up was 3 years. The standard of reference for the diagnosis of osteomyelitis was histologic/microbiologic results of surgical biopsy specimens or clinical follow-up. Note was made if decubitus ulcers, sinus tract, fistula, fluid collection, abscess, septic arthritis, joint effusion, bursitis, or heterotopic ossification was present on MRI. Comparison of the extent of infection by MRI and surgical margins was performed. Effect on surgical treatment was defined by absence of recurrent infection at the surgical site within 6 weeks of limited resection.
RESULTS: The criteria for diagnosis of osteomyelitis were fulfilled in 49 of 57 MR studies for an overall accuracy of 97%. There was one false-negative MR study. MRI for the diagnosis of osteomyelitis yielded a sensitivity of 98% and a specificity of 89%. There were 41 decubitus ulcers, 28 sinus tracts, 2 fistulae, 14 fluid collections, 15 abscesses, 9 hips with septic arthritis, 10 "bland" hip effusions, 5 cases of trochanteric bursitis, and 30 patients with heterotopic ossification. Twenty-one patients underwent limited surgical resection guided by MR findings in which only the enhancing area was resected. There was only one recurrence of osteomyelitis at the surgical margins.
CONCLUSION: MRI is accurate in the diagnosis of osteomyelitis and associated soft tissue abnormalities in spinal cord-injured patients. MRI can delineate the extent of infection in guiding limited surgical resection and preserving viable tissue.