Motivation: Conventional breast DWI suffers from limitations in image quality, hindering accurate diagnosis. Advances in deep learning-enhanced DWI techniques offer potential improvements, enabling higher-quality imaging for more reliable diagnosis. Goal(s): To evaluate effectiveness of DL-enhanced RESOLVE (DL-RESOLVE) compared to standard RESOLVE in clinical breast imaging. Approach: DL-RESOLVE and RESOLVE breast images from 9 participants were acquired. Two independent radiologists, blinded to the sequence type, evaluated the image quality. Quantitative metrics — SNR, CNR, and ADC —were measured. Results: DL-RESOLVE demonstrated higher SNR, reduced artifacts, improved lesion visibility and shorter acquisition time compared to RESOLVE, highlighting its potential to enhance diagnostic accuracy in breast MRI. Impact: DL-RESOLVE leverages advanced deep learning techniques to provide contrast-free breast MRI with better image quality, reduced artifacts, and shorter acquisition time. This innovative approach enhances diagnostic accuracy and efficiency, making it a promising tool for clinical breast imaging.
Laraib et al. (Tue,) studied this question.