Computational Pathology for Colorectal Cancer: Bridging the Gap Between AI Research and the 20-Element Clinical Report
Despite advances in AI-based image analysis, most computational pathology tools for colorectal cancer address only three of the more than 20 elements defined in ICCR reporting guidelines. This review maps the current gaps between AI capabilities and clinical requirements, providing a structured framework for aligning future tool development with the full diagnostic reporting workflow.
The original study
Aligning computational pathology with clinical practice for colorectal cancer.
- Authors
- Baumann E, Carreño-Martínez JF, Frei AL, García-Baroja J, Gwerder M, Khan A, et al.
- Journal
- NPJ precision oncology
- Type
- Journal Article, Review
- PMID
- 41310156
Original abstract
The pathology report in colorectal cancer (CRC) consists of more than 20 elements defined in guidelines such as the International Collaboration on Cancer Reporting (ICCR) guidelines. Recently, computational tools have been proposed to advance the CRC diagnostic routine, yet most lack clinically validated results and focus on only three report elements. This review gives an overview of the current gaps and will contribute to aligning computational pathology with clinical practice.