Scholarlink

Drug-drug interactions (DDIs) pose a significant clinical risk, especially as polypharmacy becomes more common in ageing populations. Traditional interaction checkers rely on static databases and often fail to infer mechanisms for unrecorded drug combinations. This contribution introduces INFERMed, a retrieval-augmented generation (RAG) system that integrates multi-source knowledge with a pharmacokinetic/pharmacodynamic (PK/PD) reasoning layer to predict and explain DDIs. The architecture combines a large knowledge graph (PubChemRDF accessed via QLever), tabular clinical and risk data (via DuckDB), and real-world adverse event reports (via OpenFDA) to ground a local large language model in factual drug information. The system dynamically retrieves enzyme, pathway, signal, and risk annotations and uses custom prompts to guide the model towards mechanistic reasoning. INFERMed was evaluated on 50 drug pairs with known interactions using an automated rubric-based inspector. The system achieved higher rubric scores for enzyme-mediated inhibition and induction cases than for absorption-based or herbal interactions. Case studies illustrate both high-accuracy explanations (e.g., metformin + iodinated contrast, rifampicin + oral contraceptive) and challenging edge cases (e.g., lithium + ACE inhibitor, warfarin + Ginkgo biloba). The results indicate that the integration of structured PK/PD knowledge with an LLM improves the explanations of DDI and supports mechanism-focused pharmacovigilance. The limitations of the current system, including sparse coverage of supplements and induction mechanisms, are discussed, along with directions for extending data sources and evaluation.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

INFERMed: A PK/PD-Aware Retrieval-Augmented System for Explainable Drug-Drug Interaction Analysis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Pranjul Mishra,
  • Maria Ganzha,
  • Dariusz Plewczynski,
  • Marcin Paprzycki

摘要

Drug-drug interactions (DDIs) pose a significant clinical risk, especially as polypharmacy becomes more common in ageing populations. Traditional interaction checkers rely on static databases and often fail to infer mechanisms for unrecorded drug combinations. This contribution introduces INFERMed, a retrieval-augmented generation (RAG) system that integrates multi-source knowledge with a pharmacokinetic/pharmacodynamic (PK/PD) reasoning layer to predict and explain DDIs. The architecture combines a large knowledge graph (PubChemRDF accessed via QLever), tabular clinical and risk data (via DuckDB), and real-world adverse event reports (via OpenFDA) to ground a local large language model in factual drug information. The system dynamically retrieves enzyme, pathway, signal, and risk annotations and uses custom prompts to guide the model towards mechanistic reasoning. INFERMed was evaluated on 50 drug pairs with known interactions using an automated rubric-based inspector. The system achieved higher rubric scores for enzyme-mediated inhibition and induction cases than for absorption-based or herbal interactions. Case studies illustrate both high-accuracy explanations (e.g., metformin + iodinated contrast, rifampicin + oral contraceptive) and challenging edge cases (e.g., lithium + ACE inhibitor, warfarin + Ginkgo biloba). The results indicate that the integration of structured PK/PD knowledge with an LLM improves the explanations of DDI and supports mechanism-focused pharmacovigilance. The limitations of the current system, including sparse coverage of supplements and induction mechanisms, are discussed, along with directions for extending data sources and evaluation.