A.C. Larson, D. Wang, A. Bangash, T.K. Rhee, K.T. Sato, T. Paunesku, G. Woloschak, R. Salem, R.A. Omary; Chicago/US

Purpose
Iterative monitoring of liver tumor perfusion during transcatheter arterial embolization (TAE) would facilitate selection of reproducible sub-stasis embolic endpoints. Dynamic first-pass contrast enhanced (CE) MRI permits reproducible perfusion measurements. However, with conventional intravenous (IV) injections, the number of iterative measurements is limited by cumulative dose. Intra-arterial (IA) injections reduce contrast dose while targeting delivery to the tumor. The purpose of this work was to assess the use of IA injections for iterative semi-quantitative perfusion measurements during TAE of liver tumors in the VX2 rabbit model. We tested the hypothesis that iterative IA injections with first-pass dynamic MRI can detect serial reductions in tumor perfusion during TAE.

Material and methods
We surgically implanted VX2 carcinoma into the left liver lobe of 4 New Zealand White rabbits. After 2-4 weeks, via femoral access and angiographic guidance we positioned a 2-F catheter to super-selectively deliver Embospheres (Biosphere Medical) to each tumor. We transferred animals to a 1.5T clinical MRI scanner (Siemens Magnetom Sonata) for first-pass CE perfusion scans using saturation recovery GRE MRI (TR/TE/TI = 2.5/1.2/100 ms, 5mm slices, 2.1s acquisition time for 7 slices repeatedly sampled for 1 min following IA injection of 2mL 0.5% Gd contrast agent). We performed dynamic CE scans at baseline and after each subsequent injection of Embospheres (~0.3-0.5 million spheres/injection, 10 min between injections). Serial embolization and imaging was continued until suspected stasis (lack of tumor enhancement). Stasis was confirmed by follow-up angiography. Separate regions-of-interest (ROI) were drawn to measure the time course of signal enhancement within both tumor tissues and hepatic arteries. We calculated the area-under-the-curve (AUC) for each signal enhancement time course. Tumor AUC, normalized by hepatic artery AUC, served as our semi-quantitative tumor perfusion measurement. We compared pre- and post-embolization tumor AUC using a matched pair t-test with alpha = 0.05.

Results
Eight liver tumors (0.5-3.0 cm diameter) were iteratively imaged during TAE in four VX2 rabbits. Tumor AUC decreased significantly from a pre-embolization baseline of 0.62±0.17 to 0.12±0.10 (mean±SD, normalized-units, p<0.001) after embolization to stasis. Serial reduction of tumor AUC was clearly depicted in each study (decreasing tumor AUC with increasing Embosphere dose).

Conclusion
Iterative IA contrast injections with dynamic first-pass MRI can detect serial reductions in tumor perfusion during TAE. Future studies are necessary to optimize iterative IA injection protocols and to compare MRI measured perfusion changes with gold-standard microsphere measurements. Hybrid MR-IR approaches for iteratively monitoring liver tumor perfusion during catheter-directed embolotherapy may permit optimal selection of embolic endpoints and could potentially serve to standardize future chemoembolization endpoints.