Abstract:
|
Multi-reader studies are required to evaluate imaging strategies. Traditional study designs (e.g., multi-reader, multi-case, MRMC) are expensive and time-consuming due to an innumerable number of plausible configurations. The overarching goal of the method is to support better decision making through informed selection of altered imaging strategies. The screening protocol utilizes a novel application of Simon's two-stage design for identifying and prioritizing candidate imaging strategies. This approach addresses challenges faced in MRMC studies-- large reader burden, which often limits the number of candidate strategies that can be considered, and preliminary data requirements to determine the MRMC sample size. Early milestones involving relatively few cases (Stage I) must be met in terms of diagnostic performance to warrant further evaluation in Stage II, which has additional cases but fewer candidate strategies. Imaging strategies that pass through this two-stage screening process are prioritized for further testing using standard MRMC methodology. This design will accelerate screening of candidate imaging strategies while assisting in managing resources and imaging priorities.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.