PhysVENeT: a physiologically-informed deep learning-based framework for the synthesis of 3D hyperpolarized gas MRI ventilation

Abstract

Functional lung imaging modalities such as hyperpolarized gas MRI ventilation enable visualization and quantification of regional lung ventilation; however, these techniques require specialized equipment and exogenous contrast, limiting clinical adoption. Physiologically-informed techniques to map proton (¹H)-MRI ventilation have been proposed. These approaches have demonstrated moderate correlation with hyperpolarized gas MRI. Recently, deep learning (DL) has been used for image synthesis applications, including functional lung image synthesis. Here, we propose a 3D multi-channel convolutional neural network that employs physiologically-informed ventilation mapping and multi-inflation structural ¹H-MRI to synthesize 3D ventilation surrogates (PhysVENeT). The dataset comprised paired inspiratory and expiratory ¹H-MRI scans and corresponding hyperpolarized gas MRI scans from 170 participants with various pulmonary pathologies. We performed fivefold cross-validation on 150 of these participants and used 20 participants with a previously unseen pathology (post COVID-19) for external validation. Synthetic ventilation surrogates were evaluated using voxel-wise correlation and structural similarity metrics; the proposed PhysVENeT framework significantly outperformed conventional ¹H-MRI ventilation mapping and other DL approaches which did not utilize structural imaging and ventilation mapping. PhysVENeT can accurately reflect ventilation defects and exhibits minimal overfitting on external validation data compared to DL approaches that do not integrate physiologically-informed mapping.

Figure 1

Registration workflow for generating ¹H-MRI specific ventilation (SV) maps.

Figure 2

PhysVENeT architecture and training strategy.

Figure 3

Example coronal slices of TLC and RV ¹H-MRI, ¹²⁹Xe-MRI, DL-based synthetic ventilation scans and the ¹H-MRI SV map for five participants in the dataset. Voxel-wise Spearman’s rs and SSIM values are given for each DL approach and the ¹H-MRI SV map. Green arrows indicate ventilation defects in hyperpolarized gas MRI scans which are replicated in synthetic ventilation scans.

Figure 4

Comparison of performance for DL methods and ¹H-MRI SV map using the voxel-wise Spearman’s rs (left) and SSIM (right) metrics. Significant p-values are provided.

Figure 5

Comparison of performance stratified by participant pathology using Spearman’s rs (left) and SSIM (right) metrics for the proposed PhysVENeT framework.

Figure 6

Comparison of performance on external validation data using the five trained models generated by the PhysVENeT during cross-validation in terms of Spearman’s rs (left) and SSIM (right). Significant p-values are provided.