Recommendations on test datasets for evaluating AI solutions in pathology
Abstract
Artificial intelligence (AI) solutions that automatically extract information from digital histology images have shown great promise for improving pathological diagnosis. Prior to routine use, it is important to evaluate their predictive performance and obtain regulatory approval. This assessment requires appropriate test datasets. However, compiling such datasets is challenging and specific recommendations are missing. A committee of various stakeholders, including commercial AI developers, pathologists, and researchers, discussed key aspects and conducted extensive literature reviews on test datasets in pathology. Here, we summarize the results and derive general recommendations for the collection of test datasets. We address several questions: Which and how many images are needed? How to deal with low-prevalence subsets? How can potential bias be detected? How should datasets be reported? What are the regulatory requirements in different countries? The recommendations are intended to help AI developers demonstrate the utility of their products and to help regulatory agencies and end users verify reported performance measures. Further research is needed to formulate criteria for sufficiently representative test datasets so that AI solutions can operate with less user intervention and better support diagnostic workflows in the future.
Cite
@article{arxiv.2204.14226,
title = {Recommendations on test datasets for evaluating AI solutions in pathology},
author = {André Homeyer and Christian Geißler and Lars Ole Schwen and Falk Zakrzewski and Theodore Evans and Klaus Strohmenger and Max Westphal and Roman David Bülow and Michaela Kargl and Aray Karjauv and Isidre Munné-Bertran and Carl Orge Retzlaff and Adrià Romero-López and Tomasz Sołtysiński and Markus Plass and Rita Carvalho and Peter Steinbach and Yu-Chia Lan and Nassim Bouteldja and David Haber and Mateo Rojas-Carulla and Alireza Vafaei Sadr and Matthias Kraft and Daniel Krüger and Rutger Fick and Tobias Lang and Peter Boor and Heimo Müller and Peter Hufnagl and Norman Zerbe},
journal= {arXiv preprint arXiv:2204.14226},
year = {2022}
}