Summary
Fibrinolytic therapy can result in rapid lysis of deep vein thrombi (DVT), but its use is limited by the failure of many patients to respond and by the increased risk of bleeding complications in comparison with anticoagulant therapy alone. Treatment could be improved by the ability to select patients most likely to respond. Since magnetic resonance (MR) imaging may be sensitive to thrombus age and structure, properties related to thrombolytic sensitivity, we have evaluated the ability of MR imaging to predict the response of DVT to thrombolytic therapy. Nine patients with venographically documented proximal DVT were treated with streptokinase, and MR imaging using a gradient recall echo sequence was performed before and after therapy. The proximal leg veins were divided into nine segments in each patient, and thrombus was present in 55 segments. The MR appearance of the thrombus in each segment was evaluated prior to therapy, and the amount of clot lysis was determined by comparing pre- and post-treatment MR images. MR imaging accurately identified thrombus in all cases in comparison with venography and also identified proximal extension into pelvic veins that was not identified venographically. Prior to treatment, the thrombus in eight of nine patients varied in appearance in different vein segments consistent with a course of progressive extension over time. Thrombi with low MR signal occurred more often in patients with symptoms of four days or less, and were also more common in partially occluded segments (10/12, 83%) than with total occlusion (7/43, 16%) (p < .001). Clot lysis correlated with the appearance of the thrombi occurring in 9/17 (53%) segments with a uniform low signal intensity compared to 2/38 segments (5%) with areas of high or intermediate intensity (p < .001). Clot lysis was also correlated with partial vein occlusion and with symptom duration of 4 days or less. However, even among partially occluded segments and in patients with shorter duration of symptoms, clot lysis occurred significantly more often in segments with low MR signal intensity. The association of low MR signal intensity with both shorter symptom duration and incomplete vein obstruction suggests that this appearance identifies newly formed thrombi which may be responsive to thrombolytic therapy. We conclude that MR imaging may be useful in predicting response and selecting patients for thrombolytic therapy.
