AI model predicts in-hospital stroke risk using PPG data

Researchers have developed a method to classify in-hospital stroke risk states using photoplethysmography (PPG) derived hemodynamic features. By analyzing continuous monitoring data from patients who experienced stroke during hospitalization, the study utilized an LLM-assisted pipeline to identify stroke anchors from clinical notes. A ResNet-1D classifier was trained on PPG data, achieving high F1-scores and AUCs across different prediction horizons in two distinct patient cohorts. The PPG model demonstrated superior performance compared to traditional clinical and EHR comparators, though the findings are retrospective and do not establish a clinical alarm or validated prediction lead time. AI

IMPACT This research demonstrates the potential for AI to enhance early detection of critical medical events by analyzing physiological signals, potentially leading to improved patient outcomes.

RANK_REASON Academic paper detailing a novel classification method using physiological data and machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Jiaming Liu
• LLM
• MIMIC-III
• photoplethysmogram
• ResNet-1D

• paper
• model release
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →