Premium
TU‐H‐206‐03: Characterizing B1 Inhomogeneities in DCE MRI
Author(s) -
Gach H,
Mason N
Publication year - 2016
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4957648
Subject(s) - imaging phantom , flip angle , homogeneity (statistics) , nuclear medicine , nuclear magnetic resonance , ground truth , physics , materials science , magnetic resonance imaging , mathematics , computer science , optics , medicine , artificial intelligence , radiology , statistics
Purpose: Dynamic Contrast Enhanced (DCE) MRI is a valuable technique for measuring perfusion and permeability characteristics of tumors. Exogenous contrast concentrations are calculated based on changes in T 1 measured using fast 3D gradient echo (FLASH) sequences. However, the slab selective pulses used in 3D MRI may result in B 1 inhomogeneities across the volume of interest that can lead to errors in T 1 and thus the estimated gadolinium concentration. We compared three FLASH DCE sequences (GRE, TWIST, and VIBE) to determine their signal homogeneity across slices and the accuracy in calculating T 1 using acquisitions with variable flip angles. Methods: The sequences were tested at 3 T on a Siemens mMR (VB20P) using a doped water phantom 3.75 g/L NiSO 4 ‐ 6H 2 O + 5 g/L NaCl (T 1 = 104 ms) and a 2% agar, 0.67% NaCl phantom (T 1 = 1.71 s). 2D EPI B 1 maps and inversion recovery T 1 maps were acquired for ground truth. 3D MRI was acquired at different flip angles to generate a T 1 map. Regions of interest were drawn to measure signal inside the phantoms as a function of slice position. The T 1 for each slice ROI was fit to the FLASH steady‐state model of magnetization. Results: Based on the data, GRE gave the most uniform signal homogeneity and T 1 values in the middle slices of the 3D volume. The 3D VIBE sequence had the largest region of signal inhomogeneity compared to the 3D GRE and TWIST sequences. VIBE's B 1 inhomogeneity is inconsistent at low flip angles. However, VIBE resulted in more slices with T 1 values similar to the ground truth. Conclusion: The central 1/3 of the slices yielded signals that result in T 1 fits consistent with the ground truth. However, the remaining slices required some form of B 1 inhomogeneity correction for quantitative DCE analysis. The research was supported in part by NIH NCI Grant R01CA159471.