Patient safety in AI-powered diagnostic pathology

Area of Science:

• Pathology
• Medical Informatics
• Artificial Intelligence

Background:

• Diagnostic pathology integrates traditional histology with artificial intelligence (AI) for enhanced accuracy.
• AI in pathology involves digital image generation, algorithm training, dataset construction, validation, and output monitoring.
• Current evidence indicates AI complements, but does not autonomously replace, human diagnostic capabilities.

Purpose of the Study:

• To critically review current AI applications in diagnostic pathology.
• To emphasize patient-centered safety considerations in AI implementation.
• To highlight the need for collaborative regulatory measures for AI in pathology.

Main Methods:

• Review of current scientific evidence on AI in diagnostic pathology.
• Analysis of key steps in AI-powered diagnostic pathology workflow.
• Examination of international healthcare recommendations for AI implementation.

Main Results:

• AI can improve diagnostic accuracy but requires human supervision.
• Generative intelligence presents new opportunities in pathology.
• Clear definition of application domains and safety monitoring are crucial for clinical AI use.

Conclusions:

• Patient safety is paramount in AI-powered diagnostic pathology.
• Collaborative efforts are essential for developing safety-oriented regulatory measures.
• International cooperation among stakeholders is necessary for responsible AI deployment in pathology.