Saved in:
Bibliographic Details
Main Authors: Qiu, Pengcheng, Wu, Chaoyi, Liu, Junwei, Zheng, Qiaoyu, Liao, Yusheng, Wang, Haowen, Yue, Yun, Fan, Qianrui, Zhen, Shuai, Wang, Jian, Gu, Jinjie, Wang, Yanfeng, Zhang, Ya, Xie, Weidi
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.24654
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912892914237440
author Qiu, Pengcheng
Wu, Chaoyi
Liu, Junwei
Zheng, Qiaoyu
Liao, Yusheng
Wang, Haowen
Yue, Yun
Fan, Qianrui
Zhen, Shuai
Wang, Jian
Gu, Jinjie
Wang, Yanfeng
Zhang, Ya
Xie, Weidi
author_facet Qiu, Pengcheng
Wu, Chaoyi
Liu, Junwei
Zheng, Qiaoyu
Liao, Yusheng
Wang, Haowen
Yue, Yun
Fan, Qianrui
Zhen, Shuai
Wang, Jian
Gu, Jinjie
Wang, Yanfeng
Zhang, Ya
Xie, Weidi
contents We present a framework for training large language models (LLMs) as diagnostic agents with reinforcement learning, enabling them to manage multi-turn interactive diagnostic processes, adaptively select examinations, and commit to final diagnoses. Unlike instruction-tuned models trained on static data, our method acquires diagnostic strategies through dynamic exploration and outcome-based feedback, mapping evolving patient states to the next optimal examination and subsequent diagnosis. Our contributions include: (i) DiagGym, a diagnostics world model trained with electronic health records, serving as a virtual clinical environment to support closed-loop in-silico training and evaluation for interactive diagnosis; (ii) DiagAgent, trained via end-to-end multi-turn RL to learn dynamic diagnostic policies that optimize both interactive effectiveness and final accuracy; (iii) DiagBench, a multi-center diagnostic benchmark designed to evaluate multi-turn diagnostic interaction trajectories. The benchmark comprises 2.2K physician-validated cases sourced from 4 distinct distributions, alongside 3.3K physician-written rubrics for granular process-oriented evaluation. (iv) Extensive evaluations demonstrate DiagAgent's superior performance across both in-domain and out-of-domain (OOD) settings. DiagAgent significantly outperforms 11 SOTA LLMs and 2 prompt-engineered agents. In the end-to-end setting, it delivers a 11.20% increase in diagnostic accuracy and a 17.58% boost in examination recommendation F1 score, while consistently maintaining SOTA performance across all three external centers. Furthermore, in rubric-based evaluations, it surpasses the next-best model by 7.1% in weighted rubric score. These findings indicate that learning policies in interactive clinical environments confers long-term diagnostic management abilities unattainable through passive training.
format Preprint
id arxiv_https___arxiv_org_abs_2510_24654
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Evolving Interactive Diagnostic Agents in a Virtual Clinical Environment
Qiu, Pengcheng
Wu, Chaoyi
Liu, Junwei
Zheng, Qiaoyu
Liao, Yusheng
Wang, Haowen
Yue, Yun
Fan, Qianrui
Zhen, Shuai
Wang, Jian
Gu, Jinjie
Wang, Yanfeng
Zhang, Ya
Xie, Weidi
Computation and Language
We present a framework for training large language models (LLMs) as diagnostic agents with reinforcement learning, enabling them to manage multi-turn interactive diagnostic processes, adaptively select examinations, and commit to final diagnoses. Unlike instruction-tuned models trained on static data, our method acquires diagnostic strategies through dynamic exploration and outcome-based feedback, mapping evolving patient states to the next optimal examination and subsequent diagnosis. Our contributions include: (i) DiagGym, a diagnostics world model trained with electronic health records, serving as a virtual clinical environment to support closed-loop in-silico training and evaluation for interactive diagnosis; (ii) DiagAgent, trained via end-to-end multi-turn RL to learn dynamic diagnostic policies that optimize both interactive effectiveness and final accuracy; (iii) DiagBench, a multi-center diagnostic benchmark designed to evaluate multi-turn diagnostic interaction trajectories. The benchmark comprises 2.2K physician-validated cases sourced from 4 distinct distributions, alongside 3.3K physician-written rubrics for granular process-oriented evaluation. (iv) Extensive evaluations demonstrate DiagAgent's superior performance across both in-domain and out-of-domain (OOD) settings. DiagAgent significantly outperforms 11 SOTA LLMs and 2 prompt-engineered agents. In the end-to-end setting, it delivers a 11.20% increase in diagnostic accuracy and a 17.58% boost in examination recommendation F1 score, while consistently maintaining SOTA performance across all three external centers. Furthermore, in rubric-based evaluations, it surpasses the next-best model by 7.1% in weighted rubric score. These findings indicate that learning policies in interactive clinical environments confers long-term diagnostic management abilities unattainable through passive training.
title Evolving Interactive Diagnostic Agents in a Virtual Clinical Environment
topic Computation and Language
url https://arxiv.org/abs/2510.24654