Causal Tree Extraction from Medical Case Reports: A Novel Task for Experts-like Text Comprehension

2025-03-03

AI Summary

Existing medical relation extraction methods struggle to model the inherent multi-layered causal reasoning structure in clinical diagnosis. This paper introduces Causal Tree Extraction (CTE), a novel task that constructs interpretable, hierarchical causal trees rooted at the primary disease to precisely capture diagnostic causal logic. Key contributions include: (1) a formal definition of the CTE task and its evaluation paradigm; (2) the first Japanese clinical case causal tree dataset, J-Casemap; (3) a generative modeling approach aligned with clinical reasoning, integrating structured prompt engineering and physician-preference-oriented evaluation metrics; and (4) a human-verification-driven evaluation protocol and a cross-task transfer experimentation framework. Human evaluation demonstrates a +20.2-point improvement over baselines. Moreover, fine-tuning on J-Casemap significantly enhances performance on downstream medical QA tasks.

Technology Category

• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Causal Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Community question answering

Abstract

Extracting causal relationships from a medical case report is essential for comprehending the case, particularly its diagnostic process. Since the diagnostic process is regarded as a bottom-up inference, causal relationships in cases naturally form a multi-layered tree structure. The existing tasks, such as medical relation extraction, are insufficient for capturing the causal relationships of an entire case, as they treat all relations equally without considering the hierarchical structure inherent in the diagnostic process. Thus, we propose a novel task, Causal Tree Extraction (CTE), which receives a case report and generates a causal tree with the primary disease as the root, providing an intuitive understanding of a case's diagnostic process. Subsequently, we construct a Japanese case report CTE dataset, J-Casemap, propose a generation-based CTE method that outperforms the baseline by 20.2 points in the human evaluation, and introduce evaluation metrics that reflect clinician preferences. Further experiments also show that J-Casemap enhances the performance of solving other medical tasks, such as question answering.

Problem

• Extract causal relationships from medical case reports.
• Generate a causal tree to understand diagnostic processes.
• Improve performance on medical tasks using causal trees.

Innovation

• Causal Tree Extraction for medical diagnostics
• Generation-based method improves baseline by 20.2%
• J-Casemap dataset enhances medical task performance