Premium
TU‐H‐206‐06: B1 Inhomogeneity Correction for Measuring T1 From DCE MRI
Author(s) -
Gach H,
Mason N,
Wang H,
Tsien C,
Benzinger T,
Robinson C
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 - 10.1118/1.4957651
Subject(s) - imaging phantom , flip angle , physics , measure (data warehouse) , pulse sequence , ground truth , pulse (music) , nuclear magnetic resonance , nuclear medicine , magnetic resonance imaging , optics , computational physics , computer science , artificial intelligence , medicine , database , detector , radiology
Purpose: 3D MRI sequences typically used in DCE MRI suffer from B1 inhomogeneities associated with the slab‐selective RF pulse. B 1 inhomogeneities affect the calculation of tissue T 1 often used to solve the two‐compartment model. Previously, the inhomogeneities were corrected based on Bloch equation simulations of the RF pulse. However, the RF pulse specification may not be readily available or the simulation may not represent the actual scan conditions. The objective of this work was to directly measure B 1 inhomogeneities and apply an inhomogeneity correction to minimize quantification errors in T 1 fitting. Methods: We used a 3D GRE sequence with varied flip angles (2–30°) to measure T 1 ‐weighted signals in phantoms at 3 T on a Siemens mMR. 2D EPI B 1 maps were acquired to measure the B 1 inhomogeneity associated with the RF coils. Inversion recovery T1 maps were acquired for ground truth. Two B 1 inhomogeneity correction methods were compared: 1) A 3D segmented EPI sequence measured the RF slab selective excitation profile; and 2) The T 1 's of phantom standards were used to normalize the 3D GRE images. Results: Only the central 1/3 of the 3D GRE slices yielded accurate T 1 values without correction. The slice profiles measured with 3D EPI were consistent between measurements made with different flip angles. The two correction approaches resulted in similar T 1 values for at least 2/3 of the slices. Approach 2 yielded more slices with values comparable to the ground truth. However, Approach 1 resulted in less error than Approach 2 for the upper 2/3 of the slices. Conclusion: Using a phantom standard with a short T 1 allowed us to shorten the acquisition time of the 3D EPI sequence to less than 1 minute or increase spatial resolution. The B 1 inhomogeneity correction technique requires further testing in vivo for DCE MRI studies.