From Many to Meaningful: Feature-Guided Zero-Shot Chronic Kidney Disease Screening Using Large Language Models

📄 arXiv: 2607.12260v1 📥 PDF

作者: Muhammad Ashad Kabir, Sirajam Munira

分类: cs.LG

发布日期: 2026-07-14

备注: Author-prepared preprint. The Version of Record was published in AIME 2026, Springer, and is available via the DOI

期刊: Proceedings of the Artificial Intelligence in Medicine (AIME 2026)

DOI: 10.1007/978-3-032-30710-1_5


💡 一句话要点

提出特征引导的零-shot框架以解决慢性肾病筛查问题

🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 慢性肾病 零-shot学习 特征选择 大型语言模型 机器学习 社区医疗 筛查方法

📋 核心要点

  1. 现有的机器学习筛查方法依赖于大量标注数据,难以在资源有限的环境中有效部署。
  2. 论文提出了一种特征引导的零-shot框架,利用临床上有意义的特征进行CKD筛查,无需特定数据集的训练。
  3. 实验结果表明,所选特征集在多个数据集上显著提高了模型的平衡准确性和概率估计,适合筛查应用。

📝 摘要(中文)

慢性肾病(CKD)的早期筛查对于防止不可逆转的病程进展至关重要。然而,许多基于机器学习的筛查方法由于依赖大量标注数据、资源密集型病理测试或高维临床特征,难以在社区和资源有限的环境中部署。本研究探讨了在零-shot设置下使用大型语言模型(LLMs)进行早期CKD筛查的可行性,提出了一种特征引导的零-shot框架,利用一组临床上有意义且易于获取的社区特征进行评估。通过对四个LLM的零-shot性能进行评估,结果显示所选特征集在平衡准确性和概率估计上均有显著提升,表明LLMs能够支持基于最少社区可获取患者特征的临床意义筛查。

🔬 方法详解

问题定义:本研究旨在解决慢性肾病(CKD)早期筛查中现有机器学习方法的局限性,特别是其对大量标注数据和复杂特征的依赖,使得在资源有限的环境中难以实施。

核心思路:论文提出了一种特征引导的零-shot框架,利用经过机器学习分析选出的临床相关特征,避免了对特定数据集的训练,从而实现了高效的CKD筛查。

技术框架:整体流程包括特征选择、数据序列化和零-shot推理三个主要模块。首先,通过机器学习分析选择出一组临床特征;然后,将患者记录序列化为文本格式;最后,利用大型语言模型进行推理。

关键创新:最重要的创新在于引入了特征引导的零-shot框架,允许在没有特定数据集训练的情况下,使用有限的社区可获取特征进行CKD筛查,这与传统方法显著不同。

关键设计:在特征选择过程中,采用了机器学习分析方法以识别出紧凑且临床相关的变量;同时,使用标准化的提示模板将表格数据转换为文本,以便进行零-shot推理。

🖼️ 关键图片

fig_0
img_1
img_2

📊 实验亮点

实验结果显示,所选特征集在四个大型语言模型(LLaMA-3、Qwen-3、Mistral和GPT-4o-mini)上均实现了显著的性能提升,平衡准确性和概率估计均有统计学意义的提高,适合用于CKD筛查。

🎯 应用场景

该研究的潜在应用领域包括社区医疗、初级保健和资源有限的医疗环境。通过使用大型语言模型进行CKD筛查,可以降低对复杂设备和大量标注数据的依赖,从而提高筛查的可及性和效率,具有重要的实际价值和未来影响。

📄 摘要(原文)

Early screening of chronic kidney disease (CKD) is essential for preventing irreversible progression; however, many machine learning (ML)-based screening methods remain difficult to deploy in community and resource-limited screening settings due to their reliance on large labeled datasets, resource-intensive pathology tests, or high-dimensional clinical features, and limited robustness to population and distributional shifts. This study examines the feasibility of using large language models (LLMs) for early-stage CKD screening in a zero-shot setting, without dataset-specific training. We propose a feature-guided zero-shot framework that evaluates LLM performance using a selected set of clinically meaningful, readily available community-based features, rather than exhaustive clinical inputs. Feature selection was guided by ML-based analysis to identify a compact, clinically relevant subset of variables. Tabular patient records were subsequently serialized into text using standardized prompt templates to enable zero-shot inference. The zero-shot performance of four LLMs (LLaMA-3, Qwen-3, Mistral, and GPT-4o-mini) was evaluated using both the full feature set and the selected subset. Generalizability was assessed across three heterogeneous CKD datasets spanning three countries. Across models and datasets, the selected feature set yielded consistent and statistically significant improvements in balanced accuracy and probability estimates, achieving performance levels suitable for screening purposes. These findings suggest that LLMs can support clinically meaningful, training-free CKD screening using minimal community-accessible patient features, offering a practical complement to conventional ML methods in real-world screening contexts.