• Ming-Chen Paul Shih, Walter J Rogers, and Klaus D Hagspiel.
• Department of Radiology, University of Virginia Health System, Charlottesville, Virginia 22908, USA.
• J Vasc Interv Radiol. 2006 Feb 1; 17 (2 Pt 1): 327-33.

PurposeTo compare the precision of magnetic resonance (MR)-guided versus fluoroscopy-guided placement of retrievable inferior vena cava (IVC) filters with use of real-time MR imaging strategies optimized for each device in an in vitro model and in an animal model.Materials And MethodsThree different retrievable IVC filters were used in this study, including the Recovery, Günther Tulip, and OptEase devices. Experiments were performed on a 1.5 T-MR system with pre-release interactive MR software. For each device, high-resolution real-time MR imaging was optimized with use of steady-state free precession and fast low-angle shot sequences with radial and cartesian trajectories and varying flip angles (10 degrees -70 degrees ) and a frame rate of 2 per second. A custom-built IVC phantom was filled with dilute gadolinium contrast agent at a concentration of 0.05 mmol/L simulating a blood T1 of 8 msec and T2 of 6 msec. Signal intensities were measured in regions of interest at the filter, the IVC lumen, and the background. The contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) were calculated. The sequence suited best for each device was chosen for in vitro filter placement in a custom-made IVC phantom. Each device was deployed five times each under MR and fluoroscopic guidance with use of identical techniques. Accuracy was measured as absolute deviation of the filter tip in millimeters from a target landing zone. Differences were assessed statistically with use of the paired t test. Each device was also placed in vivo in a swine model under MR guidance.ResultsAll three IVC filters could be clearly identified and positioned under fluoroscopic and MR imaging control. A cartesian true fast imaging sequence with steady-state precession with a flip angle of 30 degrees or 50 degrees resulted in optimal SNR and CNR for all three filters. The Tulip filter created more susceptibility artifacts than the other two. Filter placement accuracy was similar with MR and fluoroscopy whether comparing devices individually (P=NS) or as a group (P=NS). The mean absolute differences between MR and fluoroscopy were 0.088 mm for the OptEase filter, 0.41 mm for the Bard Recovery filter, and 0.34 mm for the Günther Tulip filter.ConclusionsMR-guided placement of retrievable IVC filters is feasible and as accurate as fluoroscopy-guided placement in an in vitro model. With optimized sequences, real-time MR has the potential to develop as a reasonable alternative to fluoroscopy.

12 keywords...

hide…