Improved unsupervised physics-informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients

bf Purpose: Earlier work showed that IVIM-NETorig, an unsupervised physics-informed deep neural network, was more accurate than other state-of-the-art intravoxel-incoherent motion (IVIM) fitting approaches to DWI. This study presents an improved version: IVIM-NEToptim, and characterizes its superior performance in pancreatic ductal adenocarcinoma (PDAC) patients. bf Method: In simulations (SNR=20), the accuracy, independence and consistency of IVIM-NET were evaluated for combinations of hyperparameters (fit S0, constraints, network architecture, # hidden layers, dropout, batch normalization, learning rate), by calculating the NRMSE, Spearman's ρ, and the coefficient of variation (CVNET), respectively. The best performing network, IVIM-NEToptim was compared to least squares (LS) and a Bayesian approach at different SNRs. IVIM-NEToptim's performance was evaluated in 23 PDAC patients. 14 of the patients received no treatment between scan sessions and 9 received chemoradiotherapy between sessions. Intersession within-subject standard deviations (wSD) and treatment-induced changes were assessed. bf Results: In simulations, IVIM-NEToptim outperformed IVIM-NETorig in accuracy (NRMSE(D)=0.18 vs 0.20; NMRSE(f)=0.22 vs 0.27; NMRSE(D*)=0.39 vs 0.39), independence (ρ(D*,f)=0.22 vs 0.74) and consistency (CVNET (D)=0.01 vs 0.10; CVNET (f)=0.02 vs 0.05; CVNET (D*)=0.04 vs 0.11). IVIM-NEToptim showed superior performance to the LS and Bayesian approaches at SNRs<50. In vivo, IVIM-NEToptim sshowed significantly less noisy parameter maps with lower wSD for D and f than the alternatives. In the treated cohort, IVIM-NEToptim detected the most individual patients with significant parameter changes compared to day-to-day variations. bf Conclusion: IVIM-NEToptim is recommended for IVIM fitting to DWI data.