An evaluation of automatic coding of surgical pathology reports

Area of Science:

• Medical Informatics
• Pathology
• Computational Biology

Background:

• Histopathological data requires efficient indexing for research and clinical use.
• The Systematised Nomenclature of Pathology (SNOP) provides a standardized framework for pathological diagnoses.
• Automating the coding of free-text pathology reports is a challenge in medical informatics.

Purpose of the Study:

• To outline the application of the Systematised Nomenclature of Pathology for indexing histopathological data.
• To compare two distinct computer programs designed for automated coding of pathology reports.
• To evaluate the feasibility and benefits of a constrained coding system for pathologists.

Main Methods:

• Development of an indexing system based on the Systematised Nomenclature of Pathology.
• Implementation and comparison of two distinct computer programs for natural language processing of pathology summaries.
• Assessment of coding accuracy and efficiency of the automated systems.

Main Results:

• The study successfully outlines a method for using SNOP to index histopathological data.
• Comparison of two computer programs revealed differences in their ability to generate codes from free-text summaries.
• A simpler coding system demonstrated potential utility for pathologists.

Conclusions:

• The Systematised Nomenclature of Pathology can serve as a foundation for effective histopathological data indexing.
• Automated coding systems, particularly simpler ones, offer advantages for pathologists.
• Pathologists' acceptance of system constraints is key to realizing the benefits of automated histopathological data management.