Retrieval-Augmented Generation Assistant for Anatomical Pathology Laboratories
#Retrieval-Augmented Generation #Anatomical Pathology #Laboratory Protocols #Medical AI #Healthcare Technology #RAG Systems #Biomedical AI #Knowledge Management
📌 Key Takeaways
- Researchers developed a RAG assistant for pathology laboratories
- 70% of medical decisions depend on laboratory diagnoses
- The system was tested with 99 protocols and 323 question-answer pairs
- Biomedical-specific embedding models improved performance metrics
- Single top-ranked chunk retrieval maximized efficiency and accuracy
📖 Full Retelling
Researchers Diogo Pires, Yuriy Perezhohin, and Mauro Castelli have developed and evaluated a Retrieval-Augmented Generation (RAG) assistant specifically designed for Anatomical Pathology (AP) laboratories, addressing critical documentation challenges in Portuguese healthcare settings. Their study, submitted to arXiv on December 8, 2025, introduces an AI-powered solution to transform how laboratory technicians access protocol information, which is particularly crucial given that up to 70% of medical decisions depend on laboratory diagnoses. The research team curated a novel corpus of 99 AP protocols and constructed 323 question-answer pairs to systematically evaluate their assistant's performance across ten experiments with varying chunking strategies, retrieval methods, and embedding models. The findings demonstrate that specialized AI systems can significantly improve workflow efficiency and support patient safety by converting static documentation into dynamic knowledge resources. The study highlights the importance of domain-specific approaches in healthcare AI, showing that biomedical-specific embedding models and proper chunking strategies are essential for achieving high performance in medical settings.
🏷️ Themes
Healthcare AI, Medical Documentation, Information Retrieval
📚 Related People & Topics
Anatomical pathology
Medical specialty
Anatomical pathology (Commonwealth) or anatomic pathology (U.S.) is a medical specialty that is concerned with the diagnosis of disease based on the macroscopic, microscopic, biochemical, immunologic and molecular examination of organs and tissues. Over the 20th century, surgical pathology has evolv...
Entity Intersection Graph
No entity connections available yet for this article.
Original Source
--> Computer Science > Information Retrieval arXiv:2602.22216 [Submitted on 8 Dec 2025] Title: Retrieval-Augmented Generation Assistant for Anatomical Pathology Laboratories Authors: Diogo Pires , Yuriy Perezhohin , Mauro Castelli View a PDF of the paper titled Retrieval-Augmented Generation Assistant for Anatomical Pathology Laboratories, by Diogo Pires and 2 other authors View PDF HTML Abstract: Accurate and efficient access to laboratory protocols is essential in Anatomical Pathology , where up to 70% of medical decisions depend on laboratory diagnoses. However, static documentation such as printed manuals or PDFs is often outdated, fragmented, and difficult to search, creating risks of workflow errors and diagnostic delays. This study proposes and evaluates a Retrieval-Augmented Generation assistant tailored to AP laboratories, designed to provide technicians with context-grounded answers to protocol-related queries. We curated a novel corpus of 99 AP protocols from a Portuguese healthcare institution and constructed 323 question-answer pairs for systematic evaluation. Ten experiments were conducted, varying chunking strategies, retrieval methods, and embedding models. Performance was assessed using the RAGAS framework (faithfulness, answer relevance, context recall) alongside top-k retrieval metrics. Results show that recursive chunking and hybrid retrieval delivered the strongest baseline performance. Incorporating a biomedical-specific embedding model further improved answer relevance (0.74), faithfulness (0.70), and context recall (0.77), showing the importance of domain-specialised embeddings. Top-k analysis revealed that retrieving a single top-ranked chunk 1) maximized efficiency and accuracy, reflecting the modular structure of AP protocols. These findings highlight critical design considerations for deploying RAG systems in healthcare and demonstrate their potential to transform static documentation into dynamic, reliable knowledge assistants, thus impr...
Read full article at source