Clinical Scorecard: AI in Pathology Fails Without Pathologists

At a Glance

Key Highlights

• Pathologists provide essential clinical expertise for AI development.
• Collaboration between developers and pathologists is critical for meaningful outcomes.
• Misconceptions about pathology's variability can hinder AI tool design.
• Multidisciplinary teams enhance the effectiveness of computational pathology projects.
• Sustained investment in digital pathology infrastructure is necessary for progress.

Guideline-Based Recommendations

Diagnosis

• Involve pathologists in the development of AI tools to ensure clinical relevance.

Management

• Foster continuous collaboration and communication between developers and pathologists.

Monitoring & Follow-up

• Implement iterative testing and feedback cycles throughout the development process.

Risks

• Failure to include pathologists may lead to tools that do not meet clinical needs.

Patient & Prescribing Data

Patients requiring diagnostic pathology services.

Tools that enable quantitative analysis and improve consistency are more likely to deliver clinical value.

Clinical Best Practices

• Engage multidisciplinary teams in all stages of computational pathology projects.
• Ensure clear communication between developers and pathologists to align tools with clinical practice.
• Prioritize the development of prognostic and predictive biomarkers for greater clinical impact.

References

• AI in Health Care