Consensus Between Radiologists, Specialists in Internal Medicine, and AI Software on Chest X-Rays in a...

Area of Science:

• Medical Imaging
• Telemedicine
• Healthcare Delivery Models

Background:

• Home hospitalization (HAH) is an expanding global care model.
• Telemedicine-enabled HAH services offer alternatives to traditional hospital care for select patients.
• Chest x-rays are a primary diagnostic tool in HAH settings.

Purpose of the Study:

• To describe the implementation and analysis of chest x-rays within HAH services.
• To evaluate the clinical assimilation of chest x-ray interpretation in HAH.
• To introduce novel findings on HAH chest x-ray interpretation globally.

Main Methods:

• Prospective follow-up and analysis of HAH patients undergoing chest x-rays.
• Comparative analysis of interpretation agreement among radiologists, internal medicine specialists, and AI.
• Evaluation of 300 chest radiographs from 260 HAH patients (median age 78).

Main Results:

• Cardiovascular disease was the most common comorbidity (71.2%).
• High agreement (>90%) observed across interpretation modalities.
• Substantial agreement (κ=0.65) between internal medicine specialists and radiologists; moderate agreement (κ=0.49) with AI.

Conclusions:

• Chest x-rays are vital for HAH patient management and determining hospital transfer.
• Internal medicine specialist interpretation shows strong consensus with radiologists.
• AI interpretation requires further development and validation for clinical use in HAH.