Deep venous thrombosis: diagnostic value of non-contrast-enhanced MR venography using electrocardiography-triggered three-dimensional half-Fourier FSE

Abstract

We prospectively assessed the diagnostic accuracy of non-contrast-enhanced MR venography using both the flow-refocused fresh-blood imaging (FR-FBI) and the swap phase-encode arterial double-subtraction elimination (SPADE) techniques for detecting deep vein thrombosis (DVT), as compared using conventional X-ray venography as the reference standard. Forty-one legs of 32 consecutive patients (eight men, 24 women; mean age +/- standard deviation, 69.4 +/- 15.3 years) suspected of having deep vein thrombosis and thus examined using conventional X-ray venography underwent MR FR-FBI and SPADE. Twenty-five of the 32 patients had nonmagnetizing, metal implants they had received during hip or leg surgery. Two radiologists independently assessed the MR venograms as either diagnostic or nondiagnostic and with either the presence or absence of thrombi. The sensitivities of FR-FBI and SPADE for diagnosing thrombus were 100% (53 of 53) for both reviewers. Nondiagnostic segments were excluded from this analysis. The corresponding specificities were 100% (238 of 238 for reviewer A) and 99.6% (237 of 238 for reviewer B). The interobserver agreement regarding the MR images for the assessment of thrombosis was high (kappa = 0.92). Non-contrast-enhanced MR venography using SPADE and FR-FBI is highly accurate and reproducible for diagnosing DVT. This is especially advantageous for patients who have received nonmagnetizing, metal implants during orthopedic surgery.