# Recommendations on test datasets for evaluating AI solutions in pathology

arXiv: https://arxiv.org/abs/2204.14226 · HTML: https://mlsanity.com/paper/2204.14226/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 158 references that are in the map.
- First posted on arXiv 21 Apr 2022.

## Builds on (most cited first, 15 of 15)

- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias](https://mlsanity.com/paper/1810.01943/) (2018)
- [Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imaging](https://mlsanity.com/paper/1909.12475/) (2019)
- [Quantifying the effects of data augmentation and stain color normalization in convolutional neural networks for computational pathology](https://mlsanity.com/paper/1902.06543/) (2019)
- [Aequitas: A Bias and Fairness Audit Toolkit](https://mlsanity.com/paper/1811.05577/) (2018)
- [PanNuke Dataset Extension, Insights and Baselines](https://mlsanity.com/paper/2003.10778/) (2020)
- [Lizard: A Large-Scale Dataset for Colonic Nuclear Instance Segmentation and Classification](https://mlsanity.com/paper/2108.11195/) (2021)
- [Does Your Dermatology Classifier Know What It Doesn't Know? Detecting the Long-Tail of Unseen Conditions](https://mlsanity.com/paper/2104.03829/) (2021)
- [Negative Data Augmentation](https://mlsanity.com/paper/2102.05113/) (2021)
- [A Benchmark of Medical Out of Distribution Detection](https://mlsanity.com/paper/2007.04250/) (2020)
- [Development and Validation of a Deep Learning Algorithm for Improving Gleason Scoring of Prostate Cancer](https://mlsanity.com/paper/1811.06497/) (2018)
- [Confidence-based Out-of-Distribution Detection: A Comparative Study and Analysis](https://mlsanity.com/paper/2107.02568/) (2021)
- [PathologyGAN: Learning deep representations of cancer tissue](https://mlsanity.com/paper/1907.02644/) (2019)
- [FRODO: Free rejection of out-of-distribution samples: application to chest x-ray analysis](https://mlsanity.com/paper/1907.01253/) (2019)
- [A Pathologist-Annotated Dataset for Validating Artificial Intelligence: A Project Description and Pilot Study](https://mlsanity.com/paper/2010.06995/) (2020)

## Built on by (most cited first, 8 of 8)

- [Domain shifts in dermoscopic skin cancer datasets: Evaluation of essential limitations for clinical translation](https://mlsanity.com/paper/2304.06968/) (2023)
- [The NCI Imaging Data Commons as a platform for reproducible research in computational pathology](https://mlsanity.com/paper/2303.09354/) (2023)
- [Statistical Challenges with Dataset Construction: Why You Will Never Have Enough Images](https://mlsanity.com/paper/2408.11160/) (2024)
- [Seeing the random forest through the decision trees. Supporting learning health systems from histopathology with machine learning models: Challenges and opportunities](https://mlsanity.com/paper/2312.03812/) (2023)
- [Performance of externally validated machine learning models based on histopathology images for the diagnosis, classification, prognosis, or treatment outcome prediction in female breast cancer: A systematic review](https://mlsanity.com/paper/2312.06697/) (2023)
- [Joining Forces for Pathology Diagnostics with AI Assistance: The EMPAIA Initiative](https://mlsanity.com/paper/2401.09450/) (2023)
- [Explainable histomorphology-based survival prediction of glioblastoma, IDH-wildtype](https://mlsanity.com/paper/2601.11691/) (2026)
- [Performance evaluation of deep learning models for image analysis: considerations for visual control and statistical metrics](https://mlsanity.com/paper/2603.13557/) (2026)

## Abstract (arXiv)

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.