Graph Neural Networks Fail to Improve Pulmonary Embolism Risk Stratification

Researchers have investigated the effectiveness of graph neural networks (GNNs) and other models for pulmonary embolism (PE) risk stratification using CTPA images and medical records. Their findings indicate that medical records and cardiac biomarkers are the most significant predictors, while vascular biomarkers extracted from CTPA images do not improve stratification accuracy. Surprisingly, even GNNs applied to vascular graphs did not outperform strong tabular baselines, suggesting that vascular graphs may not contain discriminative information for PE risk stratification. AI

IMPACT Investigates the limitations of current graph neural network applications in medical diagnostics.

RANK_REASON Academic paper detailing a research study and its findings. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• graph neural networks
• pulmonary embolism

• paper
• other

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