A Data-Driven Sequential Backward Selection Framework to Accelerate Diffusion-Relaxation Prostate Microstructure Mapping

Abstract

Diffusion-Relaxation Correlation Spectrum Imaging (DR-CSI) can provide unique microstructural information for prostate cancer characterization, but requires longer scan times for two-dimensional encoding of TE and b-values. This study developed a data-driven sequential backward selection framework that determined subsampled encoding schemes for DR-CSI, achieving 70% reduction of scan time to 6min while maintaining accurate ex vivo prostate microstructure mapping.