High‐resolution neural network‐driven mapping of multiple diffusion metrics leveraging asymmetries in the balanced steady‐state free precession frequency profile

Abstract

We propose to utilize the rich information content about microstructural tissue properties entangled in asymmetric balanced steady‐state free precession (bSSFP) profiles to estimate multiple diffusion metrics simultaneously by neural network (NN) parameter quantification. A 12‐point bSSFP phase‐cycling scheme with high‐resolution whole‐brain coverage is employed at 3 and 9.4 T for NN input. Low‐resolution target diffusion data are derived based on diffusion‐weighted spin‐echo echo‐planar‐imaging (SE‐EPI) scans, that is, mean, axial, and radial diffusivity (MD, AD, and RD), fractional anisotropy (FA), as well as the spherical coordinates (azimuth Φ and inclination ϴ) of the principal diffusion eigenvector. A feedforward NN is trained with incorporated probabilistic uncertainty estimation. The NN predictions yielded highly reliable results in white matter (WM) and gray matter structures for MD. The quantification of FA, AD, and RD was overall in good agreement with the reference but the dependence of these parameters on WM anisotropy was somewhat biased (e.g. in corpus callosum). The inclination ϴ was well predicted for anisotropic WM structures, while the azimuth Φ was overall poorly predicted. The findings were highly consistent across both field strengths. Application of the optimized NN to high‐resolution input data provided whole‐brain maps with rich structural details. In conclusion, the proposed NN‐driven approach showed potential to provide distortion‐free high‐resolution whole‐brain maps of multiple diffusion metrics at high to ultrahigh field strengths in clinically relevant scan times.

Flexible and Cost-Effective Deep Learning for Fast Multi-Parametric Relaxometry using Phase-Cycled bSSFP

“…Those reflect anisotropies in tissue microstructure with a correlation to diffusion metrics, e.g. in WM [38][39][40][41] , or the sensitivity to chemical shift, which can be exploited for fat fraction mapping 42 . Comparable to MIRACLE, this results in an underestimation of T 1 and T 2 in brain tissues with respect to gold standard spin-echo-based reference methods 15,16 .…”

SPARCQ: A new approach for fat fraction mapping using asymmetries in the phase‐cycled balanced SSFP signal profile

“…34 Such asymmetries were observed in gray matter, white matter, and muscle, [34][35][36] were stronger at high magnetic fields, 37 and can be avoided by reducing TR. 38 Profile asymmetries were used to obtain diffusion metrics, 39 and recent work suggests an interplay between the contributions of microstructural, chemical shift, and chemical exchange effects to the asymmetry of the bSSFP profile. 40 Although phase-cycled bSSFP was used for water-fat signal separation, [41][42][43][44][45] its potential to quantify fat by analyzing profile asymmetries has not been investigated.…”