Dataset Information

The deep learning algorithm estimates chest radiograph-based sex and age as independent risk factors for future cardiovascular outcomes.

ABSTRACT:

Objectives

Chest X-rays (CXRs) convey much illegible physiological information that deep learning model (DLM) has been reported interpreting successfully. Since the electrocardiogram age established by DLM was revealed as a heart biological marker, we hypothesize that CXR age has similar potential to describe the heart and lung states. Therefore, we developed a DLM to predict sex and age through CXR and analyzed its relation with future cardiovascular diseases (CVD).

Methods

A total of 90,396 CXRs aged 20 to 90 were collected and separated into a development set with 53,102 CXRs and demographic information pairs, a tuning set with 7073 pairs, an internal validation set with 17,364 pairs, and an external validation set with 12,857 pairs. The study trained DLM with development set for estimating age and sex and compared them to actual information.

Results

The mean absolute errors of predicted age were 4.803 and 4.313 years in the internal and external validation sets, respectively. The area under the curve of sex analysis was 0.9993 and 0.9988 in the internal and external validation sets, respectively. Patients whose CXR age was 5 years older than chronologic age lead to higher risk of all-cause mortality (hazard ratio (HR): 2.42, 95% confidence interval (CI): 2.00-2.92), cardiovascular (CV)-cause mortality (HR: 7.57, 95% CI: 4.55-12.60), new-onset heart failure (HR: 2.07, 95% CI: 1.56-2.76), new-onset chronic kidney disease (HR: 1.73, 95% CI: 1.46-2.05), new-onset acute myocardial infarction (HR: 1.80, 95% CI: 1.12-2.92), new-onset stroke (HR: 1.45, 95% CI: 1.10-1.90), new-onset coronary artery disease (HR: 1.26, 95% CI: 1.04-1.52), and new-onset atrial fibrillation (HR: 1.43, 95% CI: 1.01-2.02).

Conclusions

Using DLM to predict CXR age provided additional information for future CVDs. Older CXR age is an accessible risk classification tool for clinician use.

SUBMITTER: Liao HC

PROVIDER: S-EPMC10388631 | biostudies-literature | 2023 Jan-Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

The deep learning algorithm estimates chest radiograph-based sex and age as independent risk factors for future cardiovascular outcomes.

Liao Hao-Chun HC Lin Chin C Wang Chih-Hung CH Fang Wen-Hui WH

Digital health 20230101

<h4>Objectives</h4>Chest X-rays (CXRs) convey much illegible physiological information that deep learning model (DLM) has been reported interpreting successfully. Since the electrocardiogram age established by DLM was revealed as a heart biological marker, we hypothesize that CXR age has similar potential to describe the heart and lung states. Therefore, we developed a DLM to predict sex and age through CXR and analyzed its relation with future cardiovascular diseases (CVD).<h4>Methods</h4>A tot ...[more]

PMID: 37529539

Similar Datasets

Project description:BackgroundChest radiograph interpretation is critical for the detection of thoracic diseases, including tuberculosis and lung cancer, which affect millions of people worldwide each year. This time-consuming task typically requires expert radiologists to read the images, leading to fatigue-based diagnostic error and lack of diagnostic expertise in areas of the world where radiologists are not available. Recently, deep learning approaches have been able to achieve expert-level performance in medical image interpretation tasks, powered by large network architectures and fueled by the emergence of large labeled datasets. The purpose of this study is to investigate the performance of a deep learning algorithm on the detection of pathologies in chest radiographs compared with practicing radiologists.Methods and findingsWe developed CheXNeXt, a convolutional neural network to concurrently detect the presence of 14 different pathologies, including pneumonia, pleural effusion, pulmonary masses, and nodules in frontal-view chest radiographs. CheXNeXt was trained and internally validated on the ChestX-ray8 dataset, with a held-out validation set consisting of 420 images, sampled to contain at least 50 cases of each of the original pathology labels. On this validation set, the majority vote of a panel of 3 board-certified cardiothoracic specialist radiologists served as reference standard. We compared CheXNeXt's discriminative performance on the validation set to the performance of 9 radiologists using the area under the receiver operating characteristic curve (AUC). The radiologists included 6 board-certified radiologists (average experience 12 years, range 4-28 years) and 3 senior radiology residents, from 3 academic institutions. We found that CheXNeXt achieved radiologist-level performance on 11 pathologies and did not achieve radiologist-level performance on 3 pathologies. The radiologists achieved statistically significantly higher AUC performance on cardiomegaly, emphysema, and hiatal hernia, with AUCs of 0.888 (95% confidence interval [CI] 0.863-0.910), 0.911 (95% CI 0.866-0.947), and 0.985 (95% CI 0.974-0.991), respectively, whereas CheXNeXt's AUCs were 0.831 (95% CI 0.790-0.870), 0.704 (95% CI 0.567-0.833), and 0.851 (95% CI 0.785-0.909), respectively. CheXNeXt performed better than radiologists in detecting atelectasis, with an AUC of 0.862 (95% CI 0.825-0.895), statistically significantly higher than radiologists' AUC of 0.808 (95% CI 0.777-0.838); there were no statistically significant differences in AUCs for the other 10 pathologies. The average time to interpret the 420 images in the validation set was substantially longer for the radiologists (240 minutes) than for CheXNeXt (1.5 minutes). The main limitations of our study are that neither CheXNeXt nor the radiologists were permitted to use patient history or review prior examinations and that evaluation was limited to a dataset from a single institution.ConclusionsIn this study, we developed and validated a deep learning algorithm that classified clinically important abnormalities in chest radiographs at a performance level comparable to practicing radiologists. Once tested prospectively in clinical settings, the algorithm could have the potential to expand patient access to chest radiograph diagnostics.

Project description:Chest radiography is a frequently used imaging modality in children. However, only fair to moderate inter-observer agreement has been reported between chest radiograph interpreters. Most studies were not performed in real-world clinical settings. Our aims were to examine the agreement between emergency department pediatricians and board-certified radiologists in a pediatric real-life setting and to identify clinical risk factors for the discrepancies. Included were children aged 3 months to 18 years who underwent chest radiography in the emergency department not during the regular hours of radiologist interpretation. Every case was reviewed by an expert panel. Inter-observer agreement between emergency department pediatricians and board-certified radiologists was assessed by Cohen's kappa; risk factors for disagreement were analyzed. Among 1373 cases, the level of agreement between emergency department pediatricians and board-certified radiologists was "moderate" (k = 0.505). For radiographs performed after midnight, agreement was only "fair" (k = 0.391). The expert panel identified clinically relevant disagreements in 260 (18.9%) of the radiographs. Over-treatment of antibiotics was identified in 121 (8.9%) of the cases and under-treatment in 79 (5.8%). In a multivariable logistic regression, the following parameters were found to be significantly associated with disagreements: neurological background (p = 0.046), fever (p = 0.001), dyspnea (p = 0.014), and radiographs performed after midnight (p = 0.007).ConclusionsModerate agreement was found between emergency department pediatricians and board-certified radiologists in interpreting chest radiographs. Neurological background, fever, dyspnea, and radiographs performed after midnight were identified as risk factors for disagreement. Implementing these findings could facilitate the use of radiologist expertise, save time and resources, and potentially improve patient care.What is known• Only fair to moderate inter-observer agreement has been reported between chest radiograph interpreters. • Most studies were not performed in real-world clinical settings. Clinical risk factors for disagreements have not been reported.What is new• In this study, which included 1373 cases at the emergency department, the level of agreement between interpreters was only "moderate." • The major clinical parameters associated with interpretation discrepancies were neurological background, fever, dyspnea, and interpretations conducted during the night shift.

Project description:ObjectivesThe goal of this study was to assess whether a deep learning estimate of age from a chest radiograph image (CXR-Age) can predict longevity beyond chronological age.BackgroundChronological age is an imperfect measure of longevity. Biological age, a measure of overall health, may improve personalized care. This paper proposes a new way to estimate biological age using a convolutional neural network that takes as input a CXR image and outputs a chest x-ray age (in years) as a measure of long-term mortality risk.MethodsCXR-Age was developed using CXR from 116,035 individuals and validated in 2 held-out testing sets: 1) 75% of the CXR arm of PLCO (Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial) (N = 40,967); and 2) the CXR arm of NLST (National Lung Screening Trial) (N = 5,414). CXR-Age was compared to chronological age and a multivariable regression model of chronological age, risk factors, and radiograph findings to predict all-cause and cardiovascular mortality with a maximum 23 years and 13 years of follow-up, respectively. The primary outcome was observed mortality; results are provided for the testing datasets only.ResultsIn the PLCO testing dataset, a 5-year increase in CXR-Age carried a higher risk of all-cause mortality than a 5-year increase in chronological age (CXR-Age hazard ratio [HR]: 2.26 [95% confidence interval (CI): 2.24 to 2.29] vs. chronological age HR: 1.77 [95% CI: 1.75 to 1.78]; p < 0.001). A similar pattern was found for cardiovascular mortality (CXR-Age cause-specific HR: 2.45 per 5 years [95% CI: 2.34 to 2.56] vs. chronological age HR: 1.82 per 5 years [95% CI: 1.74 to 1.90]). Similar results were seen for both outcomes in the NLST external testing dataset. Adding CXR-Age to the multivariable model resulted in significant improvements for predicting both outcomes in both testing datasets (p < 0.001 for all comparisons).ConclusionsBased on a CXR image, CXR-Age predicted long-term all-cause and cardiovascular mortality.

Dataset Information

The deep learning algorithm estimates chest radiograph-based sex and age as independent risk factors for future cardiovascular outcomes.

Objectives

Methods

Results

Conclusions

Publications

The deep learning algorithm estimates chest radiograph-based sex and age as independent risk factors for future cardiovascular outcomes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets