Research output: Contribution to journal › Article › Academic › peer-review
Self-supervised neural network improves tri-exponential intravoxel incoherent motion model fitting compared to least-squares fitting in non-alcoholic fatty liver disease. / Troelstra, Marian A.; van Dijk, Anne-Marieke; Witjes, Julia J. et al.
In: Frontiers in physiology, Vol. 13, 942495, 06.09.2022.Research output: Contribution to journal › Article › Academic › peer-review
}
TY - JOUR
T1 - Self-supervised neural network improves tri-exponential intravoxel incoherent motion model fitting compared to least-squares fitting in non-alcoholic fatty liver disease
AU - Troelstra, Marian A.
AU - van Dijk, Anne-Marieke
AU - Witjes, Julia J.
AU - Mak, Anne Linde
AU - Zwirs, Diona
AU - Runge, Jurgen H.
AU - Verheij, Joanne
AU - Beuers, Ulrich H.
AU - Nieuwdorp, Max
AU - Holleboom, Adriaan G.
AU - Nederveen, Aart J.
AU - Gurney-Champion, Oliver J.
N1 - Publisher Copyright: Copyright © 2022 Troelstra, Van Dijk, Witjes, Mak, Zwirs, Runge, Verheij, Beuers, Nieuwdorp, Holleboom, Nederveen and Gurney-Champion.
PY - 2022/9/6
Y1 - 2022/9/6
N2 - Recent literature suggests that tri-exponential models may provide additional information and fit liver intravoxel incoherent motion (IVIM) data more accurately than conventional bi-exponential models. However, voxel-wise fitting of IVIM results in noisy and unreliable parameter maps. For bi-exponential IVIM, neural networks (NN) were able to produce superior parameter maps than conventional least-squares (LSQ) generated images. Hence, to improve parameter map quality of tri-exponential IVIM, we developed an unsupervised physics-informed deep neural network (IVIM3-NET). We assessed its performance in simulations and in patients with non-alcoholic fatty liver disease (NAFLD) and compared outcomes with bi-exponential LSQ and NN fits and tri-exponential LSQ fits. Scanning was performed using a 3.0T free-breathing multi-slice diffusion-weighted single-shot echo-planar imaging sequence with 18 b-values. Images were analysed for visual quality, comparing the bi- and tri-exponential IVIM models for LSQ fits and NN fits using parameter-map signal-to-noise ratios (SNR) and adjusted R2. IVIM parameters were compared to histological fibrosis, disease activity and steatosis grades. Parameter map quality improved with bi- and tri-exponential NN approaches, with a significant increase in average parameter-map SNR from 3.38 to 5.59 and 2.45 to 4.01 for bi- and tri-exponential LSQ and NN models respectively. In 33 out of 36 patients, the tri-exponential model exhibited higher adjusted R2 values than the bi-exponential model. Correlating IVIM data to liver histology showed that the bi- and tri-exponential NN outperformed both LSQ models for the majority of IVIM parameters (10 out of 15 significant correlations). Overall, our results support the use of a tri-exponential IVIM model in NAFLD. We show that the IVIM3-NET can be used to improve image quality compared to a tri-exponential LSQ fit and provides promising correlations with histopathology similar to the bi-exponential neural network fit, while generating potentially complementary additional parameters.
AB - Recent literature suggests that tri-exponential models may provide additional information and fit liver intravoxel incoherent motion (IVIM) data more accurately than conventional bi-exponential models. However, voxel-wise fitting of IVIM results in noisy and unreliable parameter maps. For bi-exponential IVIM, neural networks (NN) were able to produce superior parameter maps than conventional least-squares (LSQ) generated images. Hence, to improve parameter map quality of tri-exponential IVIM, we developed an unsupervised physics-informed deep neural network (IVIM3-NET). We assessed its performance in simulations and in patients with non-alcoholic fatty liver disease (NAFLD) and compared outcomes with bi-exponential LSQ and NN fits and tri-exponential LSQ fits. Scanning was performed using a 3.0T free-breathing multi-slice diffusion-weighted single-shot echo-planar imaging sequence with 18 b-values. Images were analysed for visual quality, comparing the bi- and tri-exponential IVIM models for LSQ fits and NN fits using parameter-map signal-to-noise ratios (SNR) and adjusted R2. IVIM parameters were compared to histological fibrosis, disease activity and steatosis grades. Parameter map quality improved with bi- and tri-exponential NN approaches, with a significant increase in average parameter-map SNR from 3.38 to 5.59 and 2.45 to 4.01 for bi- and tri-exponential LSQ and NN models respectively. In 33 out of 36 patients, the tri-exponential model exhibited higher adjusted R2 values than the bi-exponential model. Correlating IVIM data to liver histology showed that the bi- and tri-exponential NN outperformed both LSQ models for the majority of IVIM parameters (10 out of 15 significant correlations). Overall, our results support the use of a tri-exponential IVIM model in NAFLD. We show that the IVIM3-NET can be used to improve image quality compared to a tri-exponential LSQ fit and provides promising correlations with histopathology similar to the bi-exponential neural network fit, while generating potentially complementary additional parameters.
KW - deep learning
KW - diffusion magnetic resonance imaging
KW - intravoxel incoherent motion (IVIM)
KW - magnetic resonance imaging
KW - non-alcoholic fatty liver disease
KW - tri-exponential
KW - unsupervised learning
UR - http://www.scopus.com/inward/record.url?scp=85138340793&partnerID=8YFLogxK
U2 - 10.3389/fphys.2022.942495
DO - 10.3389/fphys.2022.942495
M3 - Article
C2 - 36148303
VL - 13
JO - Frontiers in physiology
JF - Frontiers in physiology
SN - 1664-042X
M1 - 942495
ER -
ID: 26188209