Accelerated white matter lesion analysis based on simultaneous and quantification using magnetic resonance fingerprinting and deep learning.

Ingo Hermann,Sebastian Weingärtner,Achim Gass,Sara Llufriu,Ralf Schmidt,Alena‐Kathrin Golla,Frank G Zöllner,Martijn Nagtegaal,Benedikt Rieger,Eloy Martínez‐Heras,Wei‐Kai Lee,Wu Yu‐Te,Elisabeth Solana,Jia‐Sheng Hong,Lothar R Schad

doi:10.1002/mrm.28688

Abstract

To develop an accelerated postprocessing pipeline for reproducible and efficient assessment of white matter lesions using quantitative magnetic resonance fingerprinting (MRF) and deep learning. MRF using echo-planar imaging (EPI) scans with varying repetition and echo times were acquired for whole brain quantification of and in 50 subjects with multiple sclerosis (MS) and 10 healthy volunteers along 2 centers. MRF and parametric maps were distortion corrected and denoised. A CNN was trained to reconstruct the and parametric maps, and the WM and GM probability maps. Deep learning-based postprocessing reduced reconstruction and image processing times from hours to a few seconds while maintaining high accuracy, reliability, and precision. Mean absolute error performed the best for (deviations 5.6%) and the logarithmic hyperbolic cosinus loss the best for (deviations 6.0%). MRF is a fast and robust tool for quantitative and mapping. Its long reconstruction and several postprocessing steps can be facilitated and accelerated using deep learning.

Highlights

White matter (WM) lesions are a common brain imaging finding in multiple sclerosis (MS) affecting the central nervous system
The Huber loss (HL), which is a combination of mean absolute error (MAE) and mean square error (MSE)
We trained the networks using patches and the full input resolution and MAE with 3 different types of outputs: (1) the network was trained with a single output once for T1 and another for T2∗; (2) the network was trained with both T1 and T2∗ in a single network; and (3) the network was trained with 4 output maps T1, T2∗, WM, and gray matter (GM) probability maps (4 outputs)

Summary

Introduction

White matter (WM) lesions are a common brain imaging finding in multiple sclerosis (MS) affecting the central nervous system. Quantification of relaxation times, such as T1, T2, and T2∗, is increasingly receiving interest for providing additional information beyond qualitative imaging.3-5 most quantitative methods suffer from long acquisition times as the acquisition of multiple qualitative images is required. This renders quantitative MRI susceptibility to intra-scan motion. Due to interscan motion and image distortion, multiple successive scans commonly need to be co-registered in order to allow for joint analysis

Methods

Results

Discussion

Conclusion