Motion Correction for Native Myocardial T1 Mapping using Self-supervised Deep Learning Registration with Contrast Separation.

Researchers

Journal

Modalities

Models

Abstract

In myocardial T1 mapping, undesirable motion poses significant challenges because uncorrected motion can affect T1 estimation accuracy and cause incorrect diagnosis. In this study, we proposed and evaluated a motion correction method for myocardial T1 mapping using Self-supervised Deep learning based Registration with contrAst seParation (SDRAP). A sparse coding based method was firstly proposed to separate the contrast component from T1-weighted (T1w) images. Then, a self-supervised deep neural network with cross-correlation (SDRAP-CC) or mutual information (SDRAP-MI) as the registration similarity measurement was developed to register contrast separated images, after which, signal fitting was performed on the motion corrected T1w images to generate motion corrected T1 maps. The registration network was trained and tested in 80 healthy volunteers with images acquired using the modified look-locker inversion recovery (MOLLI) sequence. The proposed SDRAP was compared with the Free Form Deformation (FFD) registration method regarding: 1) Dice Coefficient (DSC) and Mean Boundary Error (MBE) of myocardium contours; 2) T1 value and Standard deviation (SD) of T1 fitting; 3) Subjective evaluation score for overall image quality and motion artifact level; 4) computation time. Results showed that SDRAP-CC achieved the highest DSC of 85.0±3.9% and the least MBE of 0.92±0.25 mm among compare methods. Additionally, SDRAP-CC performed the best by resulting in the lower SD value (28.1±17.6 ms) and higher subjective image quality scores (3.30±0.79 for overall quality and 3.53±0.68 for motion artifact) evaluated by a cardiologist. The proposed SDRAP took only 0.52s to register one slice of MOLLI images, achieving about 7x acceleration than FFD (3.7s/slice).This article is protected by copyright. All rights reserved.