Tarsal coalition: a blinded comparison of MRI and CT.

OBJECTIVE: To determine how well MRI can detect tarsal coalition compared with CT, the current imaging standard.

MATERIALS AND METHODS: Coronal and axial CT and MRI were performed within 3 weeks of each other on 40 feet in 20 consecutive patients referred with symptoms of possible tarsal coalition. Scans were read independently in blinded fashion by different imagers. Coalitions were either complete (osseous) or incomplete (non-osseous). Results were compared with available surgical data and clinical follow-up.

RESULTS: Both modalities prospectively identified 15 coalitions (9 patients) and each missed 1 calcaneonavicular coalition. Twenty-three of the remaining 24 feet were negative for coalition on both CT and MRI. An atypical incomplete talocalcaneal coalition seen on CT was not identified prospectively on MRI.

CONCLUSION: MRI is very good for detecting tarsal coalition and has a high rate of agreement with CT, the imaging "gold standard." When clinical suspicion for coalition is high, CT remains a more cost-effective diagnostic modality. If other causes for ankle pain are also entertained, MRI can be performed and provide nearly equivalent diagnostic accuracy for detecting tarsal coalition.