B mapping using pre-learned subspaces for quantitative brain imaging.

Purpose: To develop a machine learning-based method for estimation of both transmitter and receiver B fields desired for correction of the B inhomogeneity effects in quantitative brain imaging.

Theory And Methods: A subspace model-based machine learning method was proposed for estimation of B and B fields. Probabilistic subspace models were used to capture scan-dependent variations in the B fields; the subspace basis and coefficient distributions were learned from pre-scanned training data. Estimation of the B fields for new experimental data was achieved by solving a linear optimization problem with prior distribution constraints. We evaluated the performance of the proposed method for B inhomogeneity correction in quantitative brain imaging scenarios, including T and proton density (PD) mapping from variable-flip-angle spoiled gradient-echo (SPGR) data as well as neurometabolic mapping from MRSI data, using phantom, healthy subject and brain tumor patient data.

Results: In both phantom and healthy subject data, the proposed method produced high-quality B maps. B correction on SPGR data using the estimated B maps produced significantly improved T and PD maps. In brain tumor patients, the proposed method produced more accurate and robust B estimation and correction results than conventional methods. The B maps were also applied to MRSI data from tumor patients and produced improved neurometabolite maps, with better separation between pathological and normal tissues.

Conclusion: This work presents a novel method to estimate B variations using probabilistic subspace models and machine learning. The proposed method may make correction of B inhomogeneity effects more robust in practical applications.