SU‐G‐IeP1‐03: Comparing Arterial Input Function Measurements in DCE‐MRI Using MOLLI and Phase

Abstract

Purpose:To compare two different methods for measuring the arterial input function (AIF) in dynamic contrast‐enhanced (DCE)‐MRI.Methods:Five DCE‐MRI experiments were performed on an aqueous DCE‐MRI phantom (Shelley Medical) in a Siemens 3T Trio MRI scanner with 32 channel head coil and clinical DCE‐MRI protocol. MRI signal phase (φ) data were saved for offline processing. T1 relaxation measurements, using a Modified Look‐ Locker Inversion Recovery (MOLLI) pulse sequence, were performed preand post‐DCE. The input flow rate was set to 200 mL/min to mimic blood flow in the human body. The AIF of Gd injection was determined with two different methods. 1. “Phase‐only”: Pre‐injection baseline phase (φ0) was subtracted from phase‐vs‐time [φ(t)]. This quantity, φ(t) – φ0, was then used to compute the AIF. 2. “Phase+MOLLI”: The AIF value during the post‐injection steady‐state washout, denoted AIF(w), was calculated from MOLLI T1s and known relaxivity. Phase during washout was denoted φ(w). The quantity φ(t) – φ(w) was used to compute AIF – AIF(w). The final Phase+MOLLI AIF was then calculated by adding the MOLLI AIF(w). A “gold‐standard” AIF(w) was also obtained by sampling liquid in the input tube port of the phantom post‐DCE, then later measuring T1 with standard inversion recovery (IR). AIF(w) values from Phase‐only, MOLLI, and IR were compared using a two‐tailed paired t‐test.Results:AIF(w) from Phase‐only and MOLLI were significantly different (p = 0.04). AIF(w) from MOLLI and IR were the same (p = 0.89). The Phase‐only curve therefore incorrectly estimated AIF(w).Conclusion:There is currently no standard method for determining the AIF in DCE‐MRI. This work has shown that a Phase‐only AIF returns incorrect values for the washout portion of the curve. The Phase+MOLLI method could provide a more accurate and reproducible AIF in clinical DCE‐MRI, which could lead to better diagnoses.