Retrieval-Augmented Generation Assistant for Anatomical Pathology Laboratories

Researchers developed a Retrieval-Augmented Generation (RAG) assistant for anatomical pathology laboratories to replace outdated static documentation with dynamic, searchable protocol guidance. The system achieved strong performance using biomedical-specific embeddings and could transform healthcare laboratory workflows by providing technicians with accurate, context-grounded answers to protocol queries.

• →Up to 70% of medical decisions depend on laboratory diagnoses, making accurate protocol access critical for patient safety.
• →The RAG system used 99 anatomical pathology protocols and 323 question-answer pairs to demonstrate improved workflow efficiency.
• →Biomedical-specific embedding models (MedEmbed) significantly improved answer relevance, faithfulness, and context recall metrics.
• →Single top-ranked chunk retrieval (k=1) maximized both efficiency and accuracy for the modular structure of medical protocols.
• →The system transforms static healthcare documentation into dynamic, reliable knowledge assistants for laboratory technicians.