Detection of deep venous thrombosis by magnetic resonance imaging

Objective: To determine the accuracy of gradient recalled echo magnetic resonance imaging in assessing deep venous thrombosis. Design: This is a retrospective review of a prospective clinical experience in 216 consecutive patients studied using gradient recalled echo magnetic resonance imaging. Sixteen patients were unavailable for follow-up and 1 study was technically suboptimal, leaving 199 studies as the basis of this report. Results: In 79 cases with confirmatory venography (n = 54), ultrasound (n = 16, thigh veins only), or computed tomography (n = 9, pelvic veins only), magnetic resonance imaging was 97 percent sensitive, 95 percent specific, and 96 percent accurate. Including cases that were normal by magnetic resonance imaging, not anticoagulated, and with uneventful follow-up as true normal cases, the corresponding sensitivity, specificity, and accuracy of magnetic resonance imaging were as follows: 97 percent, 98 percent, and 97 percent. Conclusion: Magnetic resonance imaging, using gradient recalled echo acquisitions, is capable of accurately diagnosing acute deep venous thrombosis.