Consensus-Based Minimum Requirements for the Adoption of a Computational Pathology Algorithm for Tumor Budding in
Julie E M Swillens1, Iris D Nagtegaal2, Alessandro Lugli3
1Radboud University Medical Center, IQ Health Science Department, Nijmegen, the Netherlands.
JCO Clinical Cancer Informatics
|July 31, 2026
Summary
Consensus requirements for computational pathology algorithms in colorectal cancer tumor budding assessment were established via an international Delphi study. This work aims to improve diagnostic accuracy and patient care by guiding algorithm validation and clinical adoption.
Area of Science:
- Digital pathology
- Oncology
- Medical informatics
Background:
- Tumor budding (TB) is a key prognostic biomarker in colorectal cancer (CRC).
- Current visual scoring of TB is subjective and labor-intensive, limiting clinical use.
- Computational pathology (CPath) offers potential for objective and efficient TB assessment.
Purpose of the Study:
- To establish consensus requirements for validating and implementing CPath algorithms for TB assessment in CRC.
- To guide the development of CPath tools for improved diagnostic accuracy and patient care.
Main Methods:
- An international two-round Delphi process involving 59 experts.
- Experts ranked minimal requirements for CPath algorithm implementation for TB.
- Consensus was sought on technical, organizational, ethical, and legal aspects.
Main Results:
- Consensus was reached on 21 of 29 (72%) minimal requirements for CPath algorithms.
- Agreement was achieved on technical, organizational, ethical, and legal aspects.
- Eight requirements (28%) remained without consensus after the second round.
Conclusions:
- Consensus on most requirements for CPath TB assessment was achieved.
- Context-specific evaluations and continuous stakeholder engagement are crucial for implementation.
- Further work is needed to address areas lacking consensus for widespread adoption.
