Related Experiment Video
Updated: Jul 2, 2026
08:14
MicroRNA Based Liquid Biopsy: The Experience of the Plasma miRNA Signature Classifier MSC for Lung Cancer Screening
Published on: October 26, 2017
16.1K
ESMO basic requirements for AI-based biomarkers in oncology (EBAI).
M Aldea1, M Salto-Tellez2, A Marra3
1Department of Cancer Medicine, Gustave Roussy, Villejuif, France; Faculty of Medicine, Paris-Saclay University, Kremlin Bicêtre, France; Lowe Center for Thoracic Oncology, Dana-Farber Cancer Institute, Boston, USA.
Summary
The European Society for Medical Oncology (ESMO) developed the Basic Requirements for AI-based Biomarkers In Oncology (EBAI) framework. EBAI provides criteria for adopting artificial intelligence biomarkers in routine oncology care.
Area of Science:
- Oncology
- Medical Informatics
- Biomarker Discovery
Background:
- Artificial intelligence (AI) is increasingly generating novel biomarkers in oncology.
- A need exists to bridge the gap between oncology and computer science for clinical AI biomarker implementation.
- The European Society for Medical Oncology (ESMO) proposed the Basic Requirements for AI-based Biomarkers In Oncology (EBAI) framework.
Purpose of the Study:
- To establish recommendations for AI-based biomarkers suitable for routine clinical use in oncology.
- To define a common language and criteria for the adoption of AI biomarkers.
Main Methods:
- A modified Delphi methodology was employed.
- A multidisciplinary panel of 37 experts participated in four consensus rounds.
- AI biomarkers were classified into three classes (A, B, C) with defined validation requirements.
Main Results:
- AI biomarkers classified as Class A (quantification), Class B (pre-screening), and Class C (novel, C1 prognostic, C2 predictive).
- Essential criteria include ground truth, performance, and generalizability; fairness is recommended.
- Specific validation requirements detailed for each class, emphasizing concordance, analytical, and clinical validation.
Conclusions:
- The EBAI framework provides essential criteria for the routine clinical adoption of AI-based biomarkers.
- EBAI facilitates a shared understanding among clinicians, AI developers, and researchers.
- Standardized criteria ensure robust validation and appropriate application of AI biomarkers in oncology.
