Program Information

Measurement of Renal Pyruvate-To-Lactate Exchange with Hyperpolarized ¹³C MRI

E Adamson

E Adamson¹*, J Gordon² , K Johnson¹ , S Fain¹ , (1) University of Wisconsin - Madison, Madison, WI, (2) University of California-San Fransisco, San Francisco, California

Presentations

SU-E-QI-11 Sunday 3:00PM - 6:00PM Room: Exhibit Hall

Purpose:
Previous work [1] modeling the metabolic flux between hyperpolarized [1-¹³C]pyruvate and [1-¹³C]lactate in magnetic resonance spectroscopic imaging (MRSI) experiments failed to account for vascular signal artifacts. Here, we investigate a method to minimize the vascular signal and its impact on the fidelity of metabolic modeling.

Methods:
MRSI was simulated for renal metabolism in MATLAB both with and without bipolar gradients. The resulting data were fit to a two-site exchange model [1], and the effects of vascular partial volume artifacts on kinetic modeling were assessed. Bipolar gradients were then incorporated into a gradient echo sequence to validate the simulations experimentally. The degree of diffusion weighting (b = 32 s/mm²) was determined empirically from ¹H imaging of murine renal vascular signal. The method was then tested in vivo using MRSI with bipolar gradients following injection of hyperpolarized [1-¹³C]pyruvate (~80 mM at 20% polarization).

Results:
In simulations, vascular signal contaminated the renal metabolic signal at resolutions as high as 2 x 2 mm² due to partial volume effects. The apparent exchange rate from pyruvate to lactate (k_p) was underestimated in the presence of these artifacts due to contaminating pyruvate signal. Incorporation of bipolar gradients suppressed vascular signal and improved the accuracy of k_p estimation. Experimentally, the in vivo results supported the ability of bipolar gradients to suppress vascular signal. The in vivo exchange rate increased, as predicted in simulations, from k_p = 0.012 s^-1 to k_p = 0.020 s^-1 after vascular signal suppression.

Conclusion:
We have demonstrated the limited accuracy of the two-site exchange model in the presence of vascular partial volume artifacts. The addition of bipolar gradients suppressed vascular signal and improved model accuracy in simulations. Bipolar gradients largely affected k_p estimation in vivo. Currently, slow-flowing spins in small vessels and capillaries are only partially suppressed, so further improvement is possible.

Funding Support, Disclosures, and Conflict of Interest: Funding support: Seed Grant from the Radiological Society of North America, GE Healthcare, University of Wisconsin Graduate School


Contact Email: