content top

Minicourse: Computer-assisted Decision Systems in Radiology, CAD evaluation issues.


08:30 AM – 10:00 AM


Nico Karssemeijer, Robert M. Nishikawa, and Nicholas Petrick

Code: RC223


This session includes three sections. The first is Clinical and Technical Evaluation Issues, which reviews methods to measure and compare standalone performance. Attendees will learn how the effect of CAD on reader performance can be predicted and measured. They will understand limitations of observer studies with CAD. The second section is The Statistical Perspective: The Limitations of Clinical CAD Studies, in which attendees will understand the strengths and weaknesses of the different methodologies used to evaluate CADe clinically. They will learn how to interpret seemingly disparate study results. This section highlights the pitfalls and important innovations in the different study designs. The third section, The FDA’s Perspective, reviews the basic principles and techniques for assessing computer assisted decision devices, including a discussion of standalone and reader performance testing.  This section also reviews the latest FDA thinking on the evaluation of various types of computer assisted decision devices.


Leave a Reply