Abstract No. 115: Comparison between noncontrast gradient recalled echo (GRE) and contrast-enhanced MRV with blood pool agent to detect deep vein thrombosis (DVT)

Abstract

To assess image quality, vessel visualization, and presence of deep venous thrombosis (DVT) on noncontrast gradient recalled echo (GRE) as compared to contrast-enhanced (CE) MRV using a blood pool agent (gadofosveset trisodium). 25 consecutive patients with concern for DVT within the abdomen, pelvis, or lower extremities were evaluated with MRV. Conventional noncontrast GRE and contrast-enhanced T1-weighted sequences were performed as part of our standard magnetic resonance venography protocol. All vein segments were assessed for vein visualization (1: poor; 4: excellent), vein homogeneity (1: poor; 4: excellent), venous thrombosis (acute/chronic; negative), and confidence pertaining to the presence or absence of clot (1: unsure; 4: confident). The reference standard was determined by the two senior radiologists in consensus, using all available imaging. The sequences were reviewed retrospectively by six blinded readers; two senior radiologists, two junior radiologists, and two junior radiology residents. A total of 650 segments were analyzed. A total of 42 venous segments contained acute or chronic DVT. Mean venous visualization measured 3.1 for GRE and 3.7 for CE MRV (p<0.0001). Mean vein homogeneity measured 2.8 for GRE and 3.7 for CE MRI (p<0.0001). Mean diagnostic confidence levels measured 3.2 for GRE and 3.8 for CE MRI (p<0.0001). Overall sensitivity and specificity was 72% and 93% for GRE and 78% and 96% for CE MRI, respectively. However, for the two junior radiology residents, the sensitivity improved from 69% to 75% and specificity improved from 83% to 95%. Average time for study interpretation measured 10.1 minutes for GRE and 5.5 minutes for CE MRI (p=0.0018). Contrast-enhancement provides valuable information for the diagnosis of venous thrombosis in the abdomen, pelvis, and lower extremities. Image quality and time-to-interpretation using a contrast-enhanced sequence is superior to conventional GRE. Sensitivity and specificity are only mildly improved for radiologists with experience in vascular imaging. For radiology residents, however, there is a more pronounced improvement in sensitivity and specificity.