Methodology and real-world applications of dynamic uncertain causality graph for clinical diagnosis with explainability and invariance

作者全名：Zhang, Zhan; Zhang, Qin; Jiao, Yang; Lu, Lin; Ma, Lin; Liu, Aihua; Liu, Xiao; Zhao, Juan; Xue, Yajun; Wei, Bing; Zhang, Mingxia; Gao, Ru; Zhao, Hong; Lu, Jie; Li, Fan; Zhang, Yang; Wang, Yiming; Zhang, Lei; Tian, Fengwei; Hu, Jie; Gou, Xin

作者地址：[Zhang, Zhan] Tsinghua Univ, Dept Comp Sci & Technol, Beijing, Peoples R China; [Zhang, Qin] Tsinghua Univ, Inst Nucl & New Energy Technol, Dept Comp Sci & Technol, Beijing, Peoples R China; [Jiao, Yang] Chinese Acad Med Sci & Peking Union Med Coll, Peking Union Med Coll Hosp, Dept Gen Internal Med, Beijing, Peoples R China; [Lu, Lin] Chinese Acad Med Sci & Peking Union Med Coll, Union Med Coll Hosp, Dept Endocrinol, Key Lab Endocrinol Natl Hlth Commiss, Beijing, Peoples R China; [Ma, Lin] Chinese Acad Med Sci & Peking Union Med Coll, Peking Union Med Coll Hosp, Dept Rheumatol, Beijing, Peoples R China; [Liu, Aihua] Capital Med Univ, Beijing Tiantan Hosp, Beijing Neurosurg Inst, Beijing, Peoples R China; [Liu, Xiao] Chinese Acad Med Sci, Beijing Hosp, Inst Geriatr Med, Natl Ctr Gerontol,Dept Gastroenterol, Beijing, Peoples R China; [Zhao, Juan] Capital Med Univ, Beijing YouAn Hosp, Beijing, Peoples R China; [Xue, Yajun] Tsinghua Univ, Beijing Tsinghua Changgung Hosp, Sch Clin Med, Dept Cardiol, Beijing, Peoples R China; [Wei, Bing; Zhang, Mingxia; Zhao, Hong] Capital Med Univ, Xuan Wu Hosp, Dept Cardiol, Beijing, Peoples R China; [Gao, Ru] Capital Med Univ, Beijing Chao Yang Hosp, Dept Gastroenterol, Beijing, Peoples R China; [Lu, Jie] Chinese Acad Med Sci & Peking Union Med Coll, Fuwai Hosp, Dept Special Med Treatment Ctr, Natl Ctr Cardiovasc Dis, Beijing, Peoples R China; [Li, Fan] Peking Univ, Hosp 1, Dept Gastroenterol, Beijing, Peoples R China; [Zhang, Yang] Peking Univ, Peoples Hosp, Beijing, Peoples R China; [Wang, Yiming] China Rehabil Res Ctr, Dept Urol, Beijing, Peoples R China; [Zhang, Lei] Capital Inst Pediat, Beijing, Peoples R China; [Tian, Fengwei] Chongqing Tradit Chinese Med Hosp, Chongqing, Peoples R China; [Hu, Jie] Suining Cent Hosp, Suining, Sichuan, Peoples R China; [Gou, Xin] Chongqing Med Univ, Affiliated Hosp 1, Dept Urol, Chongqing, Peoples R China

通信作者：Zhang, Q (通讯作者)，Tsinghua Univ, Inst Nucl & New Energy Technol, Dept Comp Sci & Technol, Beijing, Peoples R China.

来源：ARTIFICIAL INTELLIGENCE REVIEW

ESI学科分类：COMPUTER SCIENCE

WOS号：WOS:001230220300001

JCR分区：Q1

影响因子：10.7

年份：2024

卷号：57

期号：6

开始页：　

结束页：　

文献类型：Article

关键词：Diagnosis; Causality; Probabilistic reasoning; Explainability; Counterfactual inference

摘要：AI-aided clinical diagnosis is desired in medical care. Existing deep learning models lack explainability and mainly focus on image analysis. The recently developed Dynamic Uncertain Causality Graph (DUCG) approach is causality-driven, explainable, and invariant across different application scenarios, without problems of data collection, labeling, fitting, privacy, bias, generalization, high cost and high energy consumption. Through close collaboration between clinical experts and DUCG technicians, 46 DUCG models covering 54 chief complaints were constructed. Over 1,000 diseases can be diagnosed without triage. Before being applied in real-world, the 46 DUCG models were retrospectively verified by third-party hospitals. The verified diagnostic precisions were no less than 95%, in which the diagnostic precision for every disease including uncommon ones was no less than 80%. After verifications, the 46 DUCG models were applied in the real-world in China. Over one million real diagnosis cases have been performed, with only 17 incorrect diagnoses identified. Due to DUCG's transparency, the mistakes causing the incorrect diagnoses were found and corrected. The diagnostic abilities of the clinicians who applied DUCG frequently were improved significantly. Following the introduction to the earlier presented DUCG methodology, the recommendation algorithm for potential medical checks is presented and the key idea of DUCG is extracted.

基金机构：Institute for Guo Qiang, Tsinghua University [2020GQG0001]; Beijing Yutong Intelligence Technology Co., Ltd. [cstc2018jscx-mszdx0106, cstc2019jscx-dxwtBX0018]; Chongqing Science and Technology Bureau [2022-PUMCH-A-017]; National High Level Hospital Clinical Research Funding

基金资助正文：This research was supported by Beijing Yutong Intelligence Technology Co., Ltd., Institute for Guo Qiang, Tsinghua University (project number: 2020GQG0001), Chongqing Science and Technology Bureau (project numbers: cstc2018jscx-mszdx0106 and cstc2019jscx-dxwtBX0018), National High Level Hospital Clinical Research Funding (2022-PUMCH-A-017).