Bridging the Clinical-Computational Transparency Gap in Digital Pathology.

Area of Science:

• Computational pathology
• Medical informatics
• Digital pathology

Background:

• Computational pathology integrates computational analysis with clinical pathology to enhance diagnostics and productivity.
• Communication barriers between pathologists and developers impede the full potential of computational pathology tools.
• Existing computer-aided diagnostic (CAD) tools show promise, with some gaining regulatory approval.

Purpose of the Study:

• To propose a standardized framework to improve mutual understanding between pathologists and developers.
• To enhance the development and clinical application of computer-aided diagnostic (CAD) tools.
• To foster better comprehension of clinical objectives and computational methodologies.

Main Methods:

• Defining pivotal roles for pathologists and computer scientists in the CAD development lifecycle.
• Advocating for pathologists to understand computational terminologies, processes, and limitations.
• Encouraging computer scientists to comprehend clinical use cases and avoid clinically irrelevant metrics.

Main Results:

• Improved understanding of machine learning models among pathologists is essential to prevent misuse and misinterpretation.
• There is a need for more accurate performance representation of algorithms compared to pathologist benchmarks.
• Addressing communication gaps can lead to more effective CAD tool development and implementation.

Conclusions:

• A comprehensive grasp of both computational and clinical paradigms is vital to bridge the translational gap in computational pathology.
• Mutual comprehension between disciplines will enhance disease diagnosis accuracy and efficiency.
• Improved collaboration ultimately benefits patient care through advanced diagnostic capabilities.