Chapter 14 - Multiplatform Standardization of Breast DWI Protocols: Quality Control and Test Objects

Abstract

To more fully realize the promise of advanced breast diffusion-weighted imaging (DWI) methods in the clinical environment, repeatability and reproducibility of quantitative diffusion-based imaging results need to be assessed and maximized across vendor platforms and clinical sites. Identification of sources of variability and their possible remediation should precede initiation of costly multicenter trials. In particular, gains in consistency are realized by standardization of DWI scan sequences performed across diverse hardware platforms. Standardization in image acquisition protocol includes not only specification of obvious top-level parameters such as b value set, gradient timing, and scan geometry but also secondary-level settings that impact image quality, such as those related to fat suppression, readout bandwidth, phase-encode echo-spacing, and acceleration options. In practice, vendor-specific terminology and default DWI sequence design vary with platform and present a significant challenge to standardization. Reasonably “harmonized” DWI acquisition protocols should be defined in parallel for each clinical scanner platform. Consideration for standardized image processing pipeline and analysis methods are also important. Use of a “central core lab” for image collection, potential vendor-specific DWI sorting and quantitative diffusion metric map generation, and consistency in downstream analysis support overall standardization. This chapter will address breast DWI performance evaluation for clinical applications, overview test objects for quality control, provide guidance on how to standardize DWI protocols, and control conditions toward accurate and reproducible results.