The MRI Scanner as a Diagnostic: Image-less Active Sampling

TL;DR

This work proposes an ML-based framework that learns an active sampling strategy, via reinforcement learning, at a patient-level to directly infer disease from undersampled k-space, and achieves diagnostic performance comparable with ML-based diagnosis, using fully sampled k-space data.

Abstract

Despite the high diagnostic accuracy of Magnetic Resonance Imaging (MRI), using MRI as a Point-of-Care (POC) disease identification tool poses significant accessibility challenges due to the use of high magnetic field strength and lengthy acquisition times. We ask a simple question: Can we dynamically optimise acquired samples, at the patient level, according to an (automated) downstream decision task, while discounting image reconstruction? We propose an ML-based framework that learns an active sampling strategy, via reinforcement learning, at a patient-level to directly infer disease from undersampled k-space. We validate our approach by inferring Meniscus Tear in undersampled knee MRI data, where we achieve diagnostic performance comparable with ML-based diagnosis, using fully sampled k-space data. We analyse task-specific sampling policies, showcasing the adaptability of our active sampling approach. The introduced frugal sampling strategies have the potential to reduce high field strength requirements that in turn strengthen the viability of MRI-based POC disease identification and associated preliminary screening tools.

Figures (6)

• Figure 1: Conventional vs. ML-based Diagnostic Processes.
• Figure 2: Our framework for active MRI sampling for point-of-care diagnosis.
• Figure 3: Classification performance varying sample rates (horizontal axis). More metrics are reported in the Supplementary.
• Figure 4: $k$-space behaviour of the two policies. The horizontal axis indicates the cumulative lines acquired while sampling.The initial 16 lines are randomly sampled.
• Figure 5: $k$-space preference as a function of sample rate (vertical axis downward) and for 3 center fraction scenarios. Colorbar indicates the possibility of being sampled.