Three-Dimensional Shear-Wave Viscoelastographic Estimation by System Identification for Prostate Cancer Localization.

Objective: This study aims to estimate viscosity in vivo in the prostate using shear-wave elastography (SWE) and to evaluate its potential, alongside elasticity, for prostate cancer (PCa) localization.

Methods: A cohort of 197 patients referred for radical prostatectomy at four Dutch hospitals underwent three-dimensional (3D) SWE scans. From the acquired data, voxel-based viscosity and shear-wave (SW) velocity were computed using a system identification (SI) approach. SW velocity was also calculated using the standard SW cross-correlation for comparison. Elasticity can be derived from SW velocity. Following a registration procedure, the 3D histopathological results served as ground truth. We then trained a convolutional neural network for each individual feature combined with the prostate mask. The performance was evaluated using cross-validated receiver operating characteristic (ROC) curves.

Results: Both elasticity and viscosity were significantly higher in malignant prostate regions compared to benign regions (Mann-Whitney U test). The area under the ROC curve (AUC) and 95% confidence interval (CI) for voxel-wise PCa detection by viscosity from SI, SW velocity from SI, and SW velocity from cross-correlation were 0.68 (95% CI: 0.66, 0.70), 0.67 (95% CI: 0.65, 0.70), and 0.67 (95% CI: 0.65, 0.69), respectively.

Conclusion: We are the first to report in-vivo voxel-based viscosity of the prostate using SWE. Combining viscosity with elasticity did not significantly improve PCa detection, likely due to system limitations, acquisition noise, and modeling simplifications. However, in 197 prostates, viscosity alone achieved an AUC comparable to that of elasticity for PCa localization, indicating its potential as a powerful adjunctive biomarker.