RESUMO
RATIONALE AND OBJECTIVES: We sought to test the hypothesis that transcatheter hepatic artery delivery of dilute gadolinium (Gd) in rabbits can be monitored in real-time using magnetic resonance imaging (MRI). MATERIALS AND METHODS: Catheters (2F) were inserted via a femoral access into the hepatic arteries of six New Zealand White rabbits under radiographic guidance. After transfer to a 1.5-T MRI scanner, 26 separate hepatic artery injections of 2 mL of 4% Gd and 14 sham injections were performed. Real-time imaging of all injections was acquired using two-dimensional projection inversion recovery-gradient echo. Films of these 40 injections, as well as 10 random repeats, were independently reviewed in a randomized, blinded fashion by two Certificate of Added Qualification-certified interventional radiologists. Observers reported (i) if Gd injection occurred and (ii) if so, the location of delivery. For each observer, we compared sensitivity/specificity for real-time visualization of contrast injection and accuracy of injection localization. Interobserver and intraobserver variability was assessed using the kappa statistic. X-ray digital subtraction angiography was the gold standard for all MRI studies. RESULTS: Both observers had a sensitivity of 100% and a specificity of 93%. Accuracy for intrahepatic contrast delivery was 77% for both observers. Accuracy for extrahepatic delivery was 92% and 96%, respectively. Both interobserver and intraobserver agreement was outstanding. CONCLUSIONS: In rabbits, MRI allows for accurate real-time monitoring of transcatheter hepatic artery delivery of contrast agent. Localization accuracy is higher outside the liver than within the liver. These results can be used as a baseline reference for comparing the accuracy of delivery of Gd-tagged therapies in the future.
