MIDRC-MetricTree.

An interactive Decision Support Tool for Evaluating Machine Learning Algorithm Performance in Medical Image Analysis

Brought to you by MIDRC Technology and Development Projects 3c and 3d.

Last updated June 5, 2024

Also check out the FDA MIC-MET tree that is based on the classification branch of our MIDRC-MetricTree!

Please cite: Drukker, Karen, Berkman Sahiner, Tingting Hu, Grace Hyun Kim, Heather M. Whitney, Natalie Baughan, Kyle J. Myers, Maryellen L. Giger, Michael McNitt-Gray, MIDRC-MetricTree: a decision tree-based tool for recommending performance metrics in artificial intelligence-assisted medical image analysis, J. Med. Imag. 11(2), 024504 (2024), doi: 10.1117/1.JMI.11.2.024504.

Not sure where to start? Check out example oncology tasks and how these relate to the MIDRC MetricTree!


Note that our decision tree focuses on the recommendation of performance metrics for different medical imaging AI tasks. For other aspects of AI model development, training, and testing, please consult a recent AAPM task group report that involved multiple co-authors from MIDRC:

AAPM task group report 273: Recommendations on best practices for AI and machine learning for computer-aided diagnosis in medical imaging

Lubomir Hadjiiski,  Kenny Cha,  Heang-Ping Chan,  Karen Drukker,  Lia Morra,  Janne J. Näppi,  Berkman Sahiner,  Hiroyuki Yoshida,  Quan Chen,  Thomas M. Deserno,  Hayit Greenspan,  Henkjan Huisman,  Zhimin Huo,  Richard Mazurchuk,  Nicholas Petrick,  Daniele Regge,  Ravi Samala,  Ronald M. Summers,  Kenji Suzuki,  Georgia Tourassi,  Daniel Vergara,  Samuel G. Armato III,, Med Phys. 2023; 50: e1– e24. https://doi.org/10.1002/mp.16188