VTDCE-Net: A time invariant deep neural network for direct estimation of pharmacokinetic parameters from undersampled DCE MRI data.

Researchers

Journal

Modalities

Models

Abstract

To propose a robust time and space invariant deep learning method to directly estimate the pharmacokinetic/tracer kinetic (PK/TK) parameters from undersampled dynamic contrast-enhanced (DCE) Magnetic Resonance Imaging (MRI) data.DCE-MRI consists of 4D (3D-spatial + temporal) data and has been utilized to estimate 3D (spatial) tracer kinetic maps. Existing deep learning architectures for this task needs retraining for variation in temporal and/or spatial dimensions. This work proposes a deep learning algorithm that is invariant to training and testing in both temporal and spatial dimensions. The proposed network was based on 2.5-dimensional Unet architecture, where the encoder consists of 3D convolutional layer and the decoder consists of 2D convolutional layer. The proposed VTDCE-Net was evaluated for solving the ill-posed inverse problem of direct estimating TK parameters from undersampled k – t space data of breast cancer patients, and the results were systematically compared with a Total Variation regularization based direct parameter estimation scheme. In the breast dataset, the training was performed on patients with 32 time samples and testing was carried out on patients with 26 and 32 time samples. Translation of the proposed VTDCE-Net for brain dataset to show the generalizability was also carried out. Undersampling rates (R) of 8×, 12×, 20× were utilized with PSNR and SSIM as the figures of merit in this evaluation. TK parameter maps estimated from fully sampled data was utilized as ground truth.Experiments carried out in this work demonstrate that the proposed VTDCE-Net outperforms the Total Variation scheme on both breast and brain datasets across all undersampling rates. For K trans $\mathbf {K_{trans}}$ and V p $\mathbf {V_{p}}$ maps the improvement over Total Variation is as high as 2 dB and 5 dB respectively using the proposed VTDCE-Net.Temporal points invariant deep learning network that was proposed in this work to estimate the TK-parameters using DCE-MRI data has provided state-of-the-art performance compared to standard image reconstruction methods and is shown to work across all undersampling rates. This article is protected by copyright. All rights reserved.This article is protected by copyright. All rights reserved.