Recommendations on test datasets for evaluating AI solutions in pathology

TL;DR

This paper provides comprehensive recommendations for creating and reporting test datasets to evaluate AI solutions in pathology, aiming to improve diagnostic accuracy, regulatory approval, and clinical integration.

Contribution

It offers the first detailed, consensus-based guidelines on test dataset collection, reporting, and regulatory considerations for AI in pathology.

Findings

Guidelines on dataset size and composition

Strategies for handling low-prevalence cases

Methods for bias detection and reporting standards

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.