ChronoMedKG: A Temporally-Grounded Biomedical Knowledge Graph and Benchmark for Clinical Reasoning

2026-05-21

AI Summary

This work addresses the lack of temporal modeling in existing biomedical knowledge graphs, which limits their utility for longitudinal clinical reasoning. The authors propose ChronoMedKG, the first temporally annotated knowledge graph encompassing 6,250 diseases—including 1,657 rare disorders—with 460,000 time-stamped disease–phenotype associations. These triples are extracted via a multi-agent large language model framework that integrates consensus filtering, ontology alignment, and PubMed ID (PMID)-based evidence tracing, yielding traceable, multi-signal confidence scores. Evaluation using the newly introduced ChronoTQA temporal question-answering benchmark demonstrates that retrieval-augmented large language models leveraging ChronoMedKG achieve a failure recovery rate of 47–65% on long-tail tasks—substantially outperforming HPOA-RAG (17–29%)—while maintaining 92.7% consistency with Orphadata.

Abstract

Biomedical knowledge graphs (KGs) treat disease associations as static facts, but temporal information is crucial for clinical reasoning, e.g., a symptom diagnostic of one disease at age 3 may imply a different disease at age 13. Existing KGs such as PrimeKG, Hetionet, and iKraph do not encode when a finding becomes clinically relevant over the course of a disease. This limits their usefulness for longitudinal clinical reasoning and retrieval augmentation. We introduce ChronoMedKG, a temporal biomedical knowledge graph that contains 460,497 evidence-linked triples (filtered from 13M raw extractions) covering 13,431 diseases. Each association is tied to temporal components like onset window or progression stage, which are backed by PMID-traceable evidence and a multi-signal credibility score. The graph is constructed through a disease-autonomous multi-agent pipeline in which multiple frontier LLMs independently extract knowledge from PubMed and PMC literature. Only those relations are kept that are supported by multi-model consensus, survive credibility filtering, as well as ontology alignment. ChronoMedKG scored 92.7% agreement against Orphadata and adds temporal grounding for 6,250 diseases absent from HPOA, Orphadata, and Phenopackets, including 1,657 Orphanet-coded rare diseases. We further introduce ChronoTQA, a benchmark of 3,341 questions across eight task types (six temporal plus two static controls), with a 12-question supplementary probe. Frontier LLMs lose roughly 30 points moving from static to temporal questions; ChronoMedKG retrieval rescues 47-65% of their long-tail failures, against 17-29% for HPOA-RAG. As such, ChronoMedKG provides a crucial temporal axis for retrieval-augmented clinical systems that was previously absent.

Problem

• temporal grounding
• biomedical knowledge graph
• clinical reasoning
• longitudinal reasoning
• disease progression

Innovation

• temporal knowledge graph
• clinical reasoning
• multi-agent LLM extraction
• longitudinal biomedical knowledge
• retrieval-augmented generation