Deep Learning Reconstruction Enhances Whole-Body Diffusion-Weighted Imaging Quality for Multiple Myeloma Detection: A Preliminary Study

Motivation: Whole-body diffusion-weighted imaging (WB-DWI) is increasingly used for assessing multiple myeloma (MM). However, current WB-DWI techniques face challenges, including slow scanning speeds and suboptimal image quality. Deep-learning reconstruction (DLR) has recently been proposed to address these issues. Goal(s): Investigate the impact of DLR on Whole-body DWI's image quality in routine MM scanning. Approach: Thirty MM patients with original WB-DWI images and DLR WB-DWI images are to be included. Image quality is compared using objective and subjective metrics. Impact: DLR holds the potential to improve WB-DWI routine scanning by accelerating scan times while enhancing image quality, thereby supporting more effective detection and assessment of MM.