Open-Source AI Model for Predicting Respiratory Mortality in COPD from Chest Radiographs
Jong Hyuk Lee1, Chang-Hoon Lee2, Jayoun Kim3
1Department of Radiology, Seoul National University Hospital, Seoul, Korea.
Artificial intelligence (AI) scores from chest radiographs (CXR-Lung-Risk) effectively predict respiratory mortality in chronic obstructive pulmonary disease (COPD) patients. This AI tool offers superior prognostication compared to existing grading systems.
Area of Science:
- Medical Imaging and Artificial Intelligence
- Pulmonary Medicine and Respiratory Diseases
Background:
- Chronic Obstructive Pulmonary Disease (COPD) poses a significant global health burden.
- Accurate prognostication is crucial for managing COPD patients and improving outcomes.
- Existing risk stratification methods may not fully capture disease severity or progression.
Purpose of the Study:
- To assess the clinical utility of artificial intelligence (AI) derived scores from chest radiographs (CXR-Lung-Risk) in predicting respiratory mortality.
- To evaluate the performance of CXR-Lung-Risk scores against the Global Initiative for Chronic Obstructive Lung Disease (GOLD) grading system.
Main Methods:
- Retrospective analysis of 4226 COPD patients from a tertiary center (2011-2015).
- CXR-Lung-Risk scores generated using an open-source AI algorithm applied to chest radiographs.
- Multivariable Fine-Gray models used to assess the association between CXR-Lung-Risk and respiratory mortality, adjusting for clinical factors and lung function; discrimination evaluated using AUC.
Main Results:
- CXR-Lung-Risk scores significantly predicted respiratory mortality (subdistribution hazard ratio per 5-year increase, 1.16; P < .001), demonstrating added value beyond lung function and clinical factors.
- The area under the receiver operating characteristic curve (AUC) for CXR-Lung-Risk was 0.76, outperforming GOLD grades (AUC = 0.61; P < .001) in predicting mortality up to 10 years.
- Higher CXR-Lung-Risk scores correlated with decreased pulmonary function (P < .001).
Conclusions:
- CXR-Lung-Risk, an open-source AI tool, is a valuable and effective prognosticator for respiratory mortality in COPD patients.
- This AI-driven approach offers improved risk stratification compared to traditional methods, potentially aiding clinical decision-making.
- The findings support the integration of AI-powered image analysis into routine COPD management.
More Related Videos
08:17Generation of a Chronic Obstructive Pulmonary Disease Model in Mice by Repeated Ozone Exposure
Published on: August 25, 2017
02:09Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function
Published on: April 12, 2024
Related Concept Videos
Chronic Obstructive Pulmonary Disease-IV: Assessement and Diagnostic Studies
Medical History
Assessment of Respiration
Subjective Assessment: Nurses interview the patient to gather information directly during the subjective assessment. It includes questions about the individual's medical history, medications, and symptoms, focusing on past respiratory conditions like...
Assessment of Ventilation I: Respiratory Rate
A Ventilation assessment is critical for monitoring a patient's health status. Respiration, one of the most accessible vital signs, provides insights into the function of numerous body systems and can indicate serious health issues, such as brainstem injuries from head trauma.
Critical Guidelines for Assessing Ventilation:
Physical Assessment of the Respiratory Tract II: Inspection
Chest Configuration
The chest configuration...
Chronic Obstructive Pulmonary Disease-I: Introduction
Respiratory Volumes and Capacities
