Welcome to METRICS-E3
METRICS-E3 is the official Explanation and Elaboration document of the METhodological RadiomICs Score (METRICS). Developed by the EuSoMII Radiomics Auditing Group, METRICS-E3 is specifically crafted to help researchers deeply understand and evaluate the methodological quality of radiomics research. It provides detailed explanations, real-world positive examples from published literature, and illustrative negative examples, enhancing clarity and comprehension of each item and condition.
Explore the full METRICS-E3 statement here.
What is METRICS?
METRICS is a structured tool developed to assess the quality of radiomics studies. Created through a transparent, expert-led consensus involving a diverse international panel, METRICS was officially endorsed by the European Society of Medical Imaging Informatics (EuSoMII). It was designed as a robust quality evaluation instrument that guides methodological rigor in radiomics. METRICS has 30 items along with 5 conditions, covering both hand-crafted and deep learning-based radiomics, including computer vision.
Access the complete METRICS statement here.
Online METRICS tool can be accessed here.
What to Expect from METRICS-E3?
METRICS-E3 is more than a supplementary document. It offers detailed explanations, elaborations, examples for each METRICS item and condition, and recommendations for appropriate and reproducible scoring. For every item or condition, you will find:
- A concise rationale explaining why it matters,
- Positive examples drawn from published studies,
- Hypothetical negative examples for contrast,
- Specifics discussing positive and negative examples,
- Guidance on how to score the item or select the condition appropriately.
These resources collectively ensure METRICS is applied effectively and consistently.
Maximizing the Value of METRICS-E3
To get the most out of METRICS-E3, we recommend reading the section titled “Recommendations for using METRICS-E3 in conjunction with the METRICS tool” in the official METRICS-E3 publication. This section offers general guidance on methodological scoring and provides specific recommendations for each item and condition, presented in comprehensive summary tables based on content from the METRICS-E3 project website. These can help you apply METRICS-E3 more effectively in your evaluations.
Citing METRICS-E3
If METRICS-E3 has supported your research, please acknowledge its impact by citing the METRICS-E3 publication. Your citations play an essential role in promoting higher methodological standards and quality evaluation practices across radiomics and the wider medical artificial intelligence community.
Contributors
Burak Kocak, Angela Ammirabile, Ilaria Ambrosini, Tugba Akinci D’Antonoli, Alessandra Borgheresi, Armando Ugo Cavallo, Roberto Cannella, Gennaro D’Anna, Oliver Díaz, Fabio Doniselli, Salvatore Claudio Fanni, Samuele Ghezzo, Kevin Groot Lipman, Michail Klontzas, Andrea Ponsiglione, Arnaldo Stanzione, Matthaios Triantafyllou, Federica Vernuccio, Renato Cuocolo