用大模型推理链预测中风患者90天康复情况，效果媲美GPT-4
COPE: Chain-Of-Thought Prediction Engine for Open-Source Large Language Model Based Stroke Outcome Prediction from Clinical Notes
- 分两步推理：先生成临床逻辑，再输出功能评分预测
- 平均误差1.01分，90%准确率达74.4%，优于多个基线模型
- 开源轻量，可解释性强，适合医疗隐私敏感场景
急性缺血性中风（AIS）预后预测对临床决策和资源分配至关重要。临床记录包含丰富上下文信息，但其非结构化特性限制了传统模型应用。我们开发并评估了基于开源LLaMA-3-8B的链式思维（CoT）预后预测引擎（COPE），用于从非结构化临床笔记中预测AIS患者90天功能结局。研究纳入464例有出院小结和90天改良Rankin量表（mRS）评分的患者。COPE采用双阶段CoT框架：第一阶段生成临床推理，第二阶段输出mRS预测。与GPT-4.1、ClinicalBERT、基于结构化变量的机器学习模型（Clinical ML）以及无CoT的单步大模型相比，COPE在平均绝对误差（MAE）、±1 mRS点内准确率和精确准确率上表现优异。其MAE为1.01（95%置信区间0.92-1.11），±1准确率为74.4%（69.9, 78.8%），精确准确率为32.8%（28.0, 37.6%），性能与GPT-4.1相当，并显著优于ClinicalBERT（MAE 1.24）、Clinical ML（MAE 1.28）和单步大模型（MAE 1.20）。亚组分析显示，各性别与年龄组表现一致，仅老年患者、接受取栓治疗者及摘要较长者误差略高。结果表明，COPE是一种轻量、可解释且隐私友好的开源框架，为从非结构化临床文本中进行预后预测提供了精准实用的解决方案。
原文摘要 · Abstract (English)
Predicting outcomes in acute ischemic stroke (AIS) guides clinical decision-making, patient counseling, and resource allocation. Clinical notes contain rich contextual information, but their unstructured nature limits their use in traditional predictive models. We developed and evaluated the Chain-of-Thought (CoT) Outcome Prediction Engine (COPE), a reasoning-enhanced large language model framework, for predicting 90-day functional outcomes after AIS from unstructured clinical notes. This study included 464 AIS patients with discharge summaries and 90-day modified Rankin Scale (mRS) scores. COPE uses a two-step CoT framework based on sequential open-source LLaMA-3-8B models: the first generates clinical reasoning, and the second outputs an mRS prediction. We compared COPE with GPT-4.1, ClinicalBERT, a structured variable-based machine learning model (Clinical ML), and a single-step LLM without CoT. Performance was evaluated using mean absolute error (MAE), accuracy within +/-1 mRS point, and exact accuracy. COPE achieved an MAE of 1.01 (95% CI 0.92-1.11), +/-1 accuracy of 74.4% (69.9, 78.8%), and exact accuracy of 32.8% (28.0, 37.6%), comparable to GPT-4.1 and superior to ClinicalBERT [MAE 1.24 (1.13-1.36)], Clinical ML [1.28 (1.18-1.39)], and the single-step LLM [1.20 (1.09-1.33)]. Subgroup analyses showed consistent performance across sex and age, with slightly higher error among older patients, those undergoing thrombectomy, and those with longer summaries. These findings demonstrate that COPE, a lightweight, interpretable, and privacy-preserving open-source framework, provides an accurate and practical solution for outcome prediction from unstructured clinical text.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。