Trade-offs in human observer performance, image quality metrics, and patient dose

点击翻译按钮获取中文摘要

The lack of a coherent, evidence-based approach for determining the "right" CT imaging protocol explains to a large degree the considerable variation in doses and techniques observed in clinical CT scanning and the recent injuries to patients undergoing CT brain imaging. A comprehensive, time-efficient, and translatable approach is needed to drive CT doses to the lowest levels that maintain acceptable diagnostic accuracy and to do so while keeping up with ever-changing technology. Our long term objective is to develop and validate quantitative methodsthat can quantitatively determine, for a specific diagnostic task, CT protocols that deliver the needed diagnostic accuracy at the lowest patient dose. The obiective of this project is to determine the tradeoffs between CT image quality and patient radiation dose in a way that moves beyond current simple models (e.g. standard deviation and CTDIvol) by investigating the relationships between task-based metrics of image quality, observer performance, patient size, and patient dose. This project was previously the 4th project in a POl proposal entitled Quantitative systematic development of dose-optimized CT imaging protocols {^ POl EB014221-01).