DART: A Structured Dataset of Regulatory Drug Documents in Italian for Clinical NLP

2025-10-21

AI Summary

To address the scarcity of non-English clinical NLP resources, this paper introduces DART—the first structured dataset of Italian pharmaceutical regulatory texts, covering key pharmacological information including indications, adverse reactions, and drug–drug interactions (DDIs). Methodologically, DART is constructed via web crawling, semantic segmentation, and low-temperature large language model (LLM) decoding to generate high-quality clinical summaries; instruction tuning further enables DDI reasoning. Contributions include: (1) the first publicly available Italian-language structured corpus for pharmaceutical regulation; and (2) a few-shot fine-tuning framework that enhances LLM performance in clinical summarization and DDI inference under resource constraints. Experiments demonstrate that models trained on DART accurately identify potential DDIs and their clinical implications, significantly improving the feasibility and reliability of AI-assisted clinical decision support.

Technology Category

• Natural Language Processing: Summarization
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

The extraction of pharmacological knowledge from regulatory documents has become a key focus in biomedical natural language processing, with applications ranging from adverse event monitoring to AI-assisted clinical decision support. However, research in this field has predominantly relied on English-language corpora such as DrugBank, leaving a significant gap in resources tailored to other healthcare systems. To address this limitation, we introduce DART (Drug Annotation from Regulatory Texts), the first structured corpus of Italian Summaries of Product Characteristics derived from the official repository of the Italian Medicines Agency (AIFA). The dataset was built through a reproducible pipeline encompassing web-scale document retrieval, semantic segmentation of regulatory sections, and clinical summarization using a few-shot-tuned large language model with low-temperature decoding. DART provides structured information on key pharmacological domains such as indications, adverse drug reactions, and drug-drug interactions. To validate its utility, we implemented an LLM-based drug interaction checker that leverages the dataset to infer clinically meaningful interactions. Experimental results show that instruction-tuned LLMs can accurately infer potential interactions and their clinical implications when grounded in the structured textual fields of DART. We publicly release our code on GitHub: https://github.com/PRAISELab-PicusLab/DART.

Problem

• Addressing the lack of Italian-language pharmacological datasets for clinical NLP
• Creating structured corpus from Italian drug regulatory documents for analysis
• Enabling drug interaction checking using Italian regulatory text data

Innovation

• Created first structured Italian drug document corpus
• Built reproducible pipeline with semantic segmentation
• Used few-shot-tuned LLM for clinical summarization