APSR2026-318

Yes
Late-Breaking
Submitted
Oral Presentation
AO20-002
 
 
https://storage.unitedwebnetwork.com/files/1285/1218206-46903-APSR26_Oral+dischargeAC_433881.pptx
Format: PPTX Only
 
Format: PDF Only
15. Pulmonary Circulation
5

No more than 15 authors can be listed.
Please ensure the authors are listed in the right order.
If there are multiple affiliations, please use the "+" button next to "Select Country/Region" to add additional affiliation fields.

Guohui Fan * China-Japan Friendship Hospital Department of Clinical research and Data management, Center of Respiratory Medicine Beijing China
Dingyi Wang - China-Japan Friendship Hospital Department of Clinical research and Data management, Center of Respiratory Medicine Beijing China
Jing Ma - China-Japan Friendship Hospital Center of Respiratory Medicine Beijing China
Feiya Xu - China-Japan Friendship Hospital Center of Respiratory Medicine Beijing China
Zhenguo Zhai - China-Japan Friendship Hospital Center of Respiratory Medicine Beijing China
 
 
 
 
 
 
 
 
 
 
Presenting Author Details
Guohui
Fan
fanguohui08@163.com
 
31-40
China
Original Research
Differences between Physicians’ Decision and AI-Driven MDT Recommendations in Anticoagulant Choices at Discharge for Pulmonary Embolism Patients - An Analysis based on a nation-wide cohort
Maximum 25 words
Abstract Content
Large language models (LLMs) have shown promising capabilities across diverse tasks, yet their application to pulmonary embolism (PE) management remains scare. An AI tool named Multi-disciplinary Agent-based Resolution Simulator for PE (MARS-PE), was applied to compare the differences in anticoagulant choice at discharge and the impact on long-term prognoses between physicians and MARS-PE.
Data from the multi-center China Pulmonary Thromboembolism Registry Study (CURES) were analyzed, which enrolled PE patients hospitalized from 2016 to 2021, and prospectively followed up for at least 1 year. Clinical features, temporal lab examinations and medications were recorded by pre-designed electronic data capture (EDC) system. MARS-PE was developed based on Qwen3-next-80B and simulated multidisciplinary team (MDT) by auto selecting pre-designed built-in medical specialty agents according to specific medical scenarios, including designated 3 (pulmonologists, nephrologist, pharmacist) and auto selected 2, to generate anticoagulant choices to compare with the actual prescriptions (DOACs, LMWH or VKA) by physicians. Differences in anticoagulant choice and corresponding 30-, 90-, 180-day, 1-year bleeding and all-cause mortality rates between MARS-PE and physicians were analyzed using propensity score weighting under a target trial emulation (TTE) framework.
A total of 7225 PE patients were analyzed. The AI-human consistency rate in anticoagulant choice at discharge was 56.4% (4076 consistent vs 3149 inconsistent). Among patients prescribed LMWHs or VKA by physicians at discharge, 93.9% and 77.2% were recommended by MARS-PE to use DOACs, respectively. After inverse probability of treatment weighting, COX regression showed human-AI consistent group contributed to lower 30-, 90-, 180-day, 1-year bleeding (HR [95%CI] 0.55 [0.41-0.73], 0.63 [0.50-0.80], 0.64 [0.52-0.79], 0.63 [0.52-0.76], all P<0.001) and lower all-cause mortality rates (HR [95%CI]0.55 [0.40-0.76], 0.67 [0.54-0.83], 0.62 [0.52-0.75], 0.64 [0.54-0.75], all P<0.001) compared with inconsistent group.
MARS-PE as LLM-based MDT tool may provide effective decision support for anticoagulant choice at discharge for PE patients.
 
 
 
 
 
No
Yes
https://storage.unitedwebnetwork.com/files/1285/1218206-37089-1_376044.jpg
JPG or JPEG or PNG format ONLY
AI-MDT, anticoagulant choice at discharge, PE management
maximum 3 keywords and no abbreviations, separated by commas
none
330