Using LLMs for Multilingual Clinical Entity Linking to ICD-10

2025-09-05

AI Summary

This study addresses the challenge of automatically linking entities in multilingual clinical texts to ICD-10 codes. We propose a novel multi-stage hybrid approach that integrates precise clinical dictionary matching with large language model (LLM)-based in-context learning. Specifically, we leverage GPT-4.1 under few-shot settings to handle out-of-vocabulary terms, enabling accurate ICD-10 code recommendations for low-resource languages such as Spanish and Greek. Our method balances rule-based interpretability with LLM-driven semantic generalization, significantly improving cross-lingual clinical information structuring efficiency and consistency. Evaluated on the CodiEsp dataset, our approach achieves 0.89 category-level F1 and 0.78 subcategory-level F1; on ElCardioCC, it attains 0.85 F1—outperforming existing baselines. The framework offers a scalable, robust paradigm for multilingual clinical coding automation, advancing interoperability and standardization in global health informatics.

Technology Category

• Natural Language Processing: Machine Translation, Multilinguality, Cross-Lingual NLP
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

The linking of clinical entities is a crucial part of extracting structured information from clinical texts. It is the process of assigning a code from a medical ontology or classification to a phrase in the text. The International Classification of Diseases - 10th revision (ICD-10) is an international standard for classifying diseases for statistical and insurance purposes. Automatically assigning the correct ICD-10 code to terms in discharge summaries will simplify the work of healthcare professionals and ensure consistent coding in hospitals. Our paper proposes an approach for linking clinical terms to ICD-10 codes in different languages using Large Language Models (LLMs). The approach consists of a multistage pipeline that uses clinical dictionaries to match unambiguous terms in the text and then applies in-context learning with GPT-4.1 to predict the ICD-10 code for the terms that do not match the dictionary. Our system shows promising results in predicting ICD-10 codes on different benchmark datasets in Spanish - 0.89 F1 for categories and 0.78 F1 on subcategories on CodiEsp, and Greek - 0.85 F1 on ElCardioCC.

Problem

• Linking clinical terms to ICD-10 codes across languages
• Automating disease classification from medical discharge summaries
• Applying LLMs for multilingual clinical entity recognition

Innovation

• Using clinical dictionaries for unambiguous term matching
• Applying in-context learning with GPT-4.1
• Multistage pipeline for multilingual ICD-10 coding