Reproducibility of foot structure measurements in neuropathic diabetic patients using magnetic resonance imaging

PURPOSE: To assess the intra- and interobserver agreement of commonly reported foot structure measurements in diabetic patients with neuropathy using magnetic resonance imaging (MRI). MATERIALS AND METHODS: In 23 neuropathic diabetic patients and five age-matched healthy controls, sagittal-plane MR images of the forefoot were obtained to assess joint configuration and plantar fat-pad thickness on two different occasions by the same observer and once by a different observer. The degree of intrinsic muscle atrophy was scored from coronal plane images on two different occasions by two observers. RESULTS: The intraclass correlation coefficients (ICCs) between occasions and between observers were >0.94. The mean differences (bias) and the limits of agreement (LoA = mean +/- 2 SDs) were small for the metatarsal-phalangeal (MTP) joint angle, toe angle, and plantar fat-pad thickness (bias </= 0.8 degrees or 0.2 mm, LoA </= 3.8 degrees or 0.8 mm), but larger for interphalangeal joint angles (bias </= 3.4 degrees , LoA </= 8.8 degrees). The weighted kappa for intrinsic muscle atrophy was 0.94. CONCLUSION: Static foot structure data can be assessed reliably using MRI. Because changes in foot structure contribute to the development of foot ulcers in neuropathic patients, MRI may be a useful technique to assess risk of ulceration in these patients