Large language models surpass domain-specific architectures for antepartum electronic fetal monitoring analysis

2025-09-09

arXiv.org

AI Summary

Traditional domain-specific models for electronic fetal monitoring (EFM/CTG) analysis suffer from poor generalizability and strong data dependency. Method: We propose the first unified benchmark framework encompassing 15 models—including large language models (LLMs), temporal models, and domain-specific models—evaluated on a dataset of over 2,500 twenty-minute, multichannel CTG recordings under varying data conditions. Contribution/Results: Fine-tuned LLMs significantly outperform state-of-the-art domain-specific models in multiclass CTG classification, especially when integrating multimodal signal inputs; however, they incur substantially higher computational overhead, necessitating a trade-off between inference efficiency and accuracy. This work provides the first empirical validation of LLMs’ potential in perinatal intelligent monitoring and establishes a reproducible benchmark and methodological foundation for medical time-series foundation modeling.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Foundation models (FMs) and large language models (LLMs) have demonstrated promising generalization across diverse domains for time-series analysis, yet their potential for electronic fetal monitoring (EFM) and cardiotocography (CTG) analysis remains underexplored. Most existing CTG studies relied on domain-specific models and lack systematic comparisons with modern foundation or language models, limiting our understanding of whether these models can outperform specialized systems in fetal health assessment. In this study, we present the first comprehensive benchmark of state-of-the-art architectures for automated antepartum CTG classification. Over 2,500 20-minutes recordings were used to evaluate over 15 models spanning domain-specific, time-series, foundation, and language-model categories under a unified framework. Fine-tuned LLMs consistently outperformed both foundation and domain-specific models across data-availability scenarios, except when uterine-activity signals were absent, where domain-specific models showed greater robustness. These performance gains, however, required substantially higher computational resources. Our results highlight that while fine-tuned LLMs achieved state-of-the-art performance for CTG classification, practical deployment must balance performance with computational efficiency.

Problem

• Evaluating large language models for electronic fetal monitoring analysis
• Comparing foundation models with domain-specific CTG classification systems
• Assessing computational efficiency trade-offs in fetal health assessment

Innovation

• Fine-tuned LLMs outperform domain-specific CTG models
• Comprehensive benchmark evaluates 15 models systematically
• Balancing performance with computational efficiency for deployment