Digital image analysis and artificial intelligence in pathology diagnostics-the Swiss view

Area of Science:

• Pathology
• Medical Informatics
• Artificial Intelligence

Background:

• Digital pathology (DP) is becoming standard in clinical diagnostics.
• The increasing digitization of pathology workloads makes digital image analysis (DIA) and artificial intelligence (AI) tools desirable for daily sign-out.
• The Swiss Digital Pathology Consortium (SDiPath) recognized the need for standardized practices.

Purpose of the Study:

• To generate best-practice recommendations for the implementation of digital image analysis (DIA) and artificial intelligence (AI) in digital pathology.
• To address the specific needs and perspectives of practicing pathologists within the Swiss healthcare system.
• To provide guidance on integrating DIA/AI into routine pathology workflows.

Main Methods:

• A Delphi process was employed to reach consensus on recommendations.
• Recommendations covered scanners, quality assurance, system integration, digital workflow, and DIA/AI.
• The focus of this article is on the DIA/AI-related recommendations, verified by SDiPath members.

Main Results:

• The study generated specific recommendations for the use of DIA/AI in digital pathology.
• These recommendations reflect the practical needs and views of pathologists from various healthcare settings (academic, cantonal, private).
• The consensus-based approach ensures the recommendations are relevant and actionable for the Swiss context.

Conclusions:

• Best-practice recommendations for DIA/AI in digital pathology have been established through a consensus process.
• These guidelines are crucial for the effective and efficient integration of advanced computational tools into routine pathology diagnostics.
• The recommendations aim to support pathologists in leveraging DIA/AI to enhance diagnostic accuracy and workflow efficiency.