TY - JOUR
T1 - Improved unsupervised physics-informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients
AU - Kaandorp, Misha P. T.
AU - Barbieri, Sebastiano
AU - Klaassen, Remy
AU - van Laarhoven, Hanneke W. M.
AU - Crezee, Hans
AU - While, Peter T.
AU - Nederveen, Aart J.
AU - Gurney-Champion, Oliver J.
N1 - Funding Information: The KWF Dutch Cancer Society supported this work under Grants No. KWF‐UVA 18410 and KWF‐UVA 2013.5932. M.P.T. Kaandorp and P.T. While gratefully acknowledge support from the Research Council of Norway under FRIPRO Researcher Project 302624. Publisher Copyright: © 2021 The Authors. Magnetic Resonance in Medicine published by Wiley Periodicals LLC on behalf of International Society for Magnetic Resonance in Medicine Copyright: Copyright 2021 Elsevier B.V., All rights reserved.
PY - 2021/10
Y1 - 2021/10
N2 - Purpose: Earlier work showed that IVIM-NET orig, an unsupervised physics-informed deep neural network, was faster and more accurate than other state-of-the-art intravoxel-incoherent motion (IVIM) fitting approaches to diffusion-weighted imaging (DWI). This study presents a substantially improved version, IVIM-NET optim, and characterizes its superior performance in pancreatic cancer patients. Method: In simulations (signal-to-noise ratio [SNR] = 20), the accuracy, independence, and consistency of IVIM-NET were evaluated for combinations of hyperparameters (fit S0, constraints, network architecture, number of hidden layers, dropout, batch normalization, learning rate), by calculating the normalized root-mean-square error (NRMSE), Spearman’s ρ, and the coefficient of variation (CV NET), respectively. The best performing network, IVIM-NET optim was compared to least squares (LS) and a Bayesian approach at different SNRs. IVIM-NET optim’s performance was evaluated in an independent dataset of 23 patients with pancreatic ductal adenocarcinoma. Fourteen of the patients received no treatment between two repeated scan sessions and nine received chemoradiotherapy between the repeated sessions. Intersession within-subject standard deviations (wSD) and treatment-induced changes were assessed. Results: In simulations (SNR = 20), IVIM-NET optim outperformed IVIM-NET orig in accuracy (NRMSE(D) = 0.177 vs 0.196; NMRSE(f) = 0.220 vs 0.267; NMRSE(D*) = 0.386 vs 0.393), independence (ρ(D*, f) = 0.22 vs 0.74), and consistency (CV NET(D) = 0.013 vs 0.104; CV NET(f) = 0.020 vs 0.054; CV NET(D*) = 0.036 vs 0.110). IVIM-NET optim showed superior performance to the LS and Bayesian approaches at SNRs < 50. In vivo, IVIM-NET optim showed significantly less noisy parameter maps with lower wSD for D and f than the alternatives. In the treated cohort, IVIM-NET optim detected the most individual patients with significant parameter changes compared to day-to-day variations. Conclusion: IVIM-NET optim is recommended for accurate, informative, and consistent IVIM fitting to DWI data.
AB - Purpose: Earlier work showed that IVIM-NET orig, an unsupervised physics-informed deep neural network, was faster and more accurate than other state-of-the-art intravoxel-incoherent motion (IVIM) fitting approaches to diffusion-weighted imaging (DWI). This study presents a substantially improved version, IVIM-NET optim, and characterizes its superior performance in pancreatic cancer patients. Method: In simulations (signal-to-noise ratio [SNR] = 20), the accuracy, independence, and consistency of IVIM-NET were evaluated for combinations of hyperparameters (fit S0, constraints, network architecture, number of hidden layers, dropout, batch normalization, learning rate), by calculating the normalized root-mean-square error (NRMSE), Spearman’s ρ, and the coefficient of variation (CV NET), respectively. The best performing network, IVIM-NET optim was compared to least squares (LS) and a Bayesian approach at different SNRs. IVIM-NET optim’s performance was evaluated in an independent dataset of 23 patients with pancreatic ductal adenocarcinoma. Fourteen of the patients received no treatment between two repeated scan sessions and nine received chemoradiotherapy between the repeated sessions. Intersession within-subject standard deviations (wSD) and treatment-induced changes were assessed. Results: In simulations (SNR = 20), IVIM-NET optim outperformed IVIM-NET orig in accuracy (NRMSE(D) = 0.177 vs 0.196; NMRSE(f) = 0.220 vs 0.267; NMRSE(D*) = 0.386 vs 0.393), independence (ρ(D*, f) = 0.22 vs 0.74), and consistency (CV NET(D) = 0.013 vs 0.104; CV NET(f) = 0.020 vs 0.054; CV NET(D*) = 0.036 vs 0.110). IVIM-NET optim showed superior performance to the LS and Bayesian approaches at SNRs < 50. In vivo, IVIM-NET optim showed significantly less noisy parameter maps with lower wSD for D and f than the alternatives. In the treated cohort, IVIM-NET optim detected the most individual patients with significant parameter changes compared to day-to-day variations. Conclusion: IVIM-NET optim is recommended for accurate, informative, and consistent IVIM fitting to DWI data.
KW - IVIM
KW - deep neural network
KW - diffusion-weighted magnetic resonance imaging
KW - intravoxel incoherent motion
KW - pancreatic cancer
KW - unsupervised physics-informed deep learning
UR - http://www.scopus.com/inward/record.url?scp=85107574829&partnerID=8YFLogxK
U2 - https://doi.org/10.1002/mrm.28852
DO - https://doi.org/10.1002/mrm.28852
M3 - Article
C2 - 34105184
SN - 0740-3194
VL - 86
SP - 2250
EP - 2265
JO - Magnetic resonance in medicine
JF - Magnetic resonance in medicine
IS - 4
ER -