A hybrid deep image prior and compressed sensing reconstruction method for highly accelerated 3D coronary magnetic resonance angiography

TL;DR

This paper introduces a new method for fast 3D coronary MRI scans using a combination of deep learning and compressed sensing, reducing scan time without sacrificing image quality.

Contribution

A novel unsupervised 3D DIP-CS reconstruction method for CMRA that avoids the need for large training datasets and improves image quality.

Findings

The 3D DIP-CS method outperformed existing reconstruction techniques in accuracy and image quality.

Combining DIP with 3D TV regularization significantly improved image quality over pure DIP methods.

The method achieved high-quality 3D CMRA from a five-minute scan without requiring fully-sampled training data.

Abstract

High-resolution whole-heart coronary magnetic resonance angiography (CMRA) often suffers from unreasonably long scan times, rendering imaging acceleration highly desirable. Traditional reconstruction methods used in CMRA rely on either hand-crafted priors or supervised learning models. Although the latter often yield superior reconstruction quality, they require a large amount of training data and memory resources, and may encounter generalization issues when dealing with out-of-distribution datasets.

To address these challenges, we introduce an unsupervised reconstruction method that combines deep image prior (DIP) with compressed sensing (CS) to accelerate 3D CMRA. This method incorporates a slice-by-slice DIP reconstruction and 3D total variation (TV) regularization, enabling high-quality reconstruction under a significant acceleration while enforcing continuity in the slice direction. We evaluated our method by comparing it to iterative SENSE, CS-TV, CS-wavelet, and other DIP-based variants, using both retrospectively and prospectively undersampled datasets.

The results demonstrate the superiority of our 3D DIP-CS approach, which improved the reconstruction accuracy relative to the other approaches across both datasets. Ablation studies further reveal the benefits of combining DIP with 3D TV regularization, which leads to significant improvements of image quality over pure DIP-based methods. Evaluation of vessel sharpness and image quality scores shows that DIP-CS improves the quality of reformatted coronary arteries.

The proposed method enables scan-specific reconstruction of high-quality 3D CMRA from a five-minute acquisition, without relying on fully-sampled training data or placing a heavy burden on memory resources.

Figures