Image quality ranking for breast diffusion-weighted MRI in a multi-center clinical trial

Acquiring breast diffusion-weighted MRI (DWI) with adequate image quality is a challenge in multi-center clinical trials. The most common issues are inadequate fat suppression, image artifacts, and poor signal-to-noise ratio (SNR). We developed and evaluated a DWI quality ranking system for identifying data with quality issues while preserving analyzable data. The identification of artifacts and adequate SNR had the highest and lowest inter-reader agreement, respectively. Removing inferior quality data improved prediction of pathologic outcome following neoadjuvant treatment in a multi-center clinical trial. The quality ranking system standardizes identification of poor quality data and informs recommendations for improvement to clinical sites.