Deep learning to detect acute respiratory distress syndrome on chest radiographs: a retrospective study with external validation.
Ontology highlight
ABSTRACT: Background
Acute respiratory distress syndrome (ARDS) is a common, but under-recognised, critical illness syndrome associated with high mortality. An important factor in its under-recognition is the variability in chest radiograph interpretation for ARDS. We sought to train a deep convolutional neural network (CNN) to detect ARDS findings on chest radiographs.Methods
CNNs were pretrained on 595 506 radiographs from two centres to identify common chest findings (eg, opacity and effusion), and then trained on 8072 radiographs annotated for ARDS by multiple physicians using various transfer learning approaches. The best performing CNN was tested on chest radiographs in an internal and external cohort, including a subset reviewed by six physicians, including a chest radiologist and physicians trained in intensive care medicine. Chest radiograph data were acquired from four US hospitals.Findings
In an internal test set of 1560 chest radiographs from 455 patients with acute hypoxaemic respiratory failure, a CNN could detect ARDS with an area under the receiver operator characteristics curve (AUROC) of 0·92 (95% CI 0·89-0·94). In the subgroup of 413 images reviewed by at least six physicians, its AUROC was 0·93 (95% CI 0·88-0·96), sensitivity 83·0% (95% CI 74·0-91·1), and specificity 88·3% (95% CI 83·1-92·8). Among images with zero of six ARDS annotations (n=155), the median CNN probability was 11%, with six (4%) assigned a probability above 50%. Among images with six of six ARDS annotations (n=27), the median CNN probability was 91%, with two (7%) assigned a probability below 50%. In an external cohort of 958 chest radiographs from 431 patients with sepsis, the AUROC was 0·88 (95% CI 0·85-0·91). When radiographs annotated as equivocal were excluded, the AUROC was 0·93 (0·92-0·95).Interpretation
A CNN can be trained to achieve expert physician-level performance in ARDS detection on chest radiographs. Further research is needed to evaluate the use of these algorithms to support real-time identification of ARDS patients to ensure fidelity with evidence-based care or to support ongoing ARDS research.Funding
National Institutes of Health, Department of Defense, and Department of Veterans Affairs.
SUBMITTER: Sjoding MW
PROVIDER: S-EPMC8182690 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Purpose: Manual interpretation of chest radiographs is a challenging task and is prone to errors. An automated system capable of categorizing chest radiographs based on the pathologies identified could aid in the timely and efficient diagnosis of chest pathologies. Method: For this retrospective study, 4476 chest radiographs were collected between January and April 2021 from two tertiary care hospitals. Three expert radiologists established the ground truth, and all radiographs were analyzed using a deep-learning AI model to detect suspicious ROIs in the lungs, pleura, and cardiac regions. Three test readers (different from the radiologists who established the ground truth) independently reviewed all radiographs in two sessions (unaided and AI-aided mode) with a washout period of one month. Results: The model demonstrated an aggregate AUROC of 91.2% and a sensitivity of 88.4% in detecting suspicious ROIs in the lungs, pleura, and cardiac regions. These results outperform unaided human readers, who achieved an aggregate AUROC of 84.2% and sensitivity of 74.5% for the same task. When using AI, the aided readers obtained an aggregate AUROC of 87.9% and a sensitivity of 85.1%. The average time taken by the test readers to read a chest radiograph decreased by 21% (p < 0.01) when using AI. Conclusion: The model outperformed all three human readers and demonstrated high AUROC and sensitivity across two independent datasets. When compared to unaided interpretations, AI-aided interpretations were associated with significant improvements in reader performance and chest radiograph interpretation time.
| S-EPMC9914339 | biostudies-literature
Project description:BackgroundCardiothoracic ratio (CTR) is the ratio of the diameter of the heart to the diameter of the thorax. An abnormal CTR (>0.55) is often an indicator of an underlying pathological condition. The accurate prediction of an abnormal CTR chest X-rays (CXRs) aids in the early diagnosis of clinical conditions.PurposeWe propose a deep learning (DL)-based model for automatic CTR calculation to assist radiologists with rapid diagnosis of cardiomegaly and thus optimise the radiology flow.Material and methodsThe study population included 1012 posteroanterior CXRs from a single institution. The Attention U-Net DL architecture was used for the automatic calculation of CTR. An observer performance test was conducted to assess the radiologist's performance in diagnosing cardiomegaly with and without artificial intelligence assistance.ResultsU-Net model exhibited a sensitivity of 0.80 [95% CI: 0.75, 0.85], specificity >99%, precision of 0.99 [95% CI: 0.98, 1], and a F1 score of 0.88 [95% CI: 0.85, 0.91]. Furthermore, the sensitivity of the reviewing radiologist in identifying cardiomegaly increased from 40.50% to 88.4% when aided by the AI-generated CTR.ConclusionOur segmentation-based AI model demonstrated high specificity (>99%) and sensitivity (80%) for CTR calculation. The performance of the radiologist on the observer performance test improved significantly with provision of AI assistance. A DL-based segmentation model for rapid quantification of CTR can therefore have significant potential to be used in clinical workflows by reducing radiologists' burden and alerting to an abnormal enlarged heart early on.
| S-EPMC9309780 | biostudies-literature
Project description:PurposeTo develop a machine learning model to classify the severity grades of pulmonary edema on chest radiographs.Materials and methodsIn this retrospective study, 369 071 chest radiographs and associated radiology reports from 64 581 patients (mean age, 51.71 years; 54.51% women) from the MIMIC-CXR chest radiograph dataset were included. This dataset was split into patients with and without congestive heart failure (CHF). Pulmonary edema severity labels from the associated radiology reports were extracted from patients with CHF as four different ordinal levels: 0, no edema; 1, vascular congestion; 2, interstitial edema; and 3, alveolar edema. Deep learning models were developed using two approaches: a semisupervised model using a variational autoencoder and a pretrained supervised learning model using a dense neural network. Receiver operating characteristic curve analysis was performed on both models.ResultsThe area under the receiver operating characteristic curve (AUC) for differentiating alveolar edema from no edema was 0.99 for the semisupervised model and 0.87 for the pretrained models. Performance of the algorithm was inversely related to the difficulty in categorizing milder states of pulmonary edema (shown as AUCs for semisupervised model and pretrained model, respectively): 2 versus 0, 0.88 and 0.81; 1 versus 0, 0.79 and 0.66; 3 versus 1, 0.93 and 0.82; 2 versus 1, 0.69 and 0.73; and 3 versus 2, 0.88 and 0.63.ConclusionDeep learning models were trained on a large chest radiograph dataset and could grade the severity of pulmonary edema on chest radiographs with high performance.Supplemental material is available for this article.See also the commentary by Auffermann in this issue.© RSNA, 2021.
| S-EPMC8043362 | biostudies-literature
Project description:We propose a deep learning system to automatically detect four explainable emphysema signs on frontal and lateral chest radiographs. Frontal and lateral chest radiographs from 3000 studies were retrospectively collected. Two radiologists annotated these with 4 radiological signs of pulmonary emphysema identified from the literature. A patient with ≥2 of these signs present is considered emphysema positive. Using separate deep learning systems for frontal and lateral images we predict the presence of each of the four visual signs and use these to determine emphysema positivity. The ROC and AUC results on a set of 422 held-out cases, labeled by both radiologists, are reported. Comparison with a black-box model which predicts emphysema without the use of explainable visual features is made on the annotations from both radiologists, as well as the subset that they agreed on. DeLong's test is used to compare with the black-box model ROC and McNemar's test to compare with radiologist performance. In 422 test cases, emphysema positivity was predicted with AUCs of 0.924 and 0.946 using the reference standard from each radiologist separately. Setting model sensitivity equivalent to that of the second radiologist, our model has a comparable specificity (p = 0.880 and p = 0.143 for each radiologist respectively). Our method is comparable with the black-box model with AUCs of 0.915 (p = 0.407) and 0.935 (p = 0.291), respectively. On the 370 cases where both radiologists agreed (53 positives), our model achieves an AUC of 0.981, again comparable to the black-box model AUC of 0.972 (p = 0.289). Our proposed method can predict emphysema positivity on chest radiographs as well as a radiologist or a comparable black-box method. It additionally produces labels for four visual signs to ensure the explainability of the result. The dataset is publicly available at https://doi.org/10.5281/zenodo.6373392.
| S-EPMC9333227 | biostudies-literature
Project description:ObjectiveTo validate a novel artificial-intelligence electrocardiogram algorithm (AI-ECG) to detect left ventricular systolic dysfunction (LVSD) in an external population.BackgroundLVSD, even when asymptomatic, confers increased morbidity and mortality. We recently derived AI-ECG to detect LVSD using ECGs based on a large sample of patients treated at the Mayo Clinic.MethodsWe performed an external validation study with subjects from the Know Your Heart Study, a cross-sectional study of adults aged 35-69 years residing in two cities in Russia, who had undergone both ECG and transthoracic echocardiography. LVSD was defined as left ventricular ejection fraction ≤ 35%. We assessed the performance of the AI-ECG to identify LVSD in this distinct patient population.ResultsAmong 4277 subjects in this external population-based validation study, 0.6% had LVSD (compared to 7.8% of the original clinical derivation study). The overall performance of the AI-ECG to detect LVSD was robust with an area under the receiver operating curve of 0.82. When using the LVSD probability cut-off of 0.256 from the original derivation study, the sensitivity, specificity, and accuracy in this population were 26.9%, 97.4%, 97.0%, respectively. Other probability cut-offs were analysed for different sensitivity values.ConclusionsThe AI-ECG detected LVSD with robust test performance in a population that was very different from that used to develop the algorithm. Population-specific cut-offs may be necessary for clinical implementation. Differences in population characteristics, ECG and echocardiographic data quality may affect test performance.
| S-EPMC7955278 | biostudies-literature
Project description:BackgroundThere is interest in using convolutional neural networks (CNNs) to analyze medical imaging to provide computer-aided diagnosis (CAD). Recent work has suggested that image classification CNNs may not generalize to new data as well as previously believed. We assessed how well CNNs generalized across three hospital systems for a simulated pneumonia screening task.Methods and findingsA cross-sectional design with multiple model training cohorts was used to evaluate model generalizability to external sites using split-sample validation. A total of 158,323 chest radiographs were drawn from three institutions: National Institutes of Health Clinical Center (NIH; 112,120 from 30,805 patients), Mount Sinai Hospital (MSH; 42,396 from 12,904 patients), and Indiana University Network for Patient Care (IU; 3,807 from 3,683 patients). These patient populations had an age mean (SD) of 46.9 years (16.6), 63.2 years (16.5), and 49.6 years (17) with a female percentage of 43.5%, 44.8%, and 57.3%, respectively. We assessed individual models using the area under the receiver operating characteristic curve (AUC) for radiographic findings consistent with pneumonia and compared performance on different test sets with DeLong's test. The prevalence of pneumonia was high enough at MSH (34.2%) relative to NIH and IU (1.2% and 1.0%) that merely sorting by hospital system achieved an AUC of 0.861 (95% CI 0.855-0.866) on the joint MSH-NIH dataset. Models trained on data from either NIH or MSH had equivalent performance on IU (P values 0.580 and 0.273, respectively) and inferior performance on data from each other relative to an internal test set (i.e., new data from within the hospital system used for training data; P values both <0.001). The highest internal performance was achieved by combining training and test data from MSH and NIH (AUC 0.931, 95% CI 0.927-0.936), but this model demonstrated significantly lower external performance at IU (AUC 0.815, 95% CI 0.745-0.885, P = 0.001). To test the effect of pooling data from sites with disparate pneumonia prevalence, we used stratified subsampling to generate MSH-NIH cohorts that only differed in disease prevalence between training data sites. When both training data sites had the same pneumonia prevalence, the model performed consistently on external IU data (P = 0.88). When a 10-fold difference in pneumonia rate was introduced between sites, internal test performance improved compared to the balanced model (10× MSH risk P < 0.001; 10× NIH P = 0.002), but this outperformance failed to generalize to IU (MSH 10× P < 0.001; NIH 10× P = 0.027). CNNs were able to directly detect hospital system of a radiograph for 99.95% NIH (22,050/22,062) and 99.98% MSH (8,386/8,388) radiographs. The primary limitation of our approach and the available public data is that we cannot fully assess what other factors might be contributing to hospital system-specific biases.ConclusionPneumonia-screening CNNs achieved better internal than external performance in 3 out of 5 natural comparisons. When models were trained on pooled data from sites with different pneumonia prevalence, they performed better on new pooled data from these sites but not on external data. CNNs robustly identified hospital system and department within a hospital, which can have large differences in disease burden and may confound predictions.
| S-EPMC6219764 | biostudies-literature
Project description:BackgroundAn artificial intelligence (AI) model using chest radiography (CXR) may provide good performance in making prognoses for COVID-19.ObjectiveWe aimed to develop and validate a prediction model using CXR based on an AI model and clinical variables to predict clinical outcomes in patients with COVID-19.MethodsThis retrospective longitudinal study included patients hospitalized for COVID-19 at multiple COVID-19 medical centers between February 2020 and October 2020. Patients at Boramae Medical Center were randomly classified into training, validation, and internal testing sets (at a ratio of 8:1:1, respectively). An AI model using initial CXR images as input, a logistic regression model using clinical information, and a combined model using the output of the AI model (as CXR score) and clinical information were developed and trained to predict hospital length of stay (LOS) ≤2 weeks, need for oxygen supplementation, and acute respiratory distress syndrome (ARDS). The models were externally validated in the Korean Imaging Cohort of COVID-19 data set for discrimination and calibration.ResultsThe AI model using CXR and the logistic regression model using clinical variables were suboptimal to predict hospital LOS ≤2 weeks or the need for oxygen supplementation but performed acceptably in the prediction of ARDS (AI model area under the curve [AUC] 0.782, 95% CI 0.720-0.845; logistic regression model AUC 0.878, 95% CI 0.838-0.919). The combined model performed better in predicting the need for oxygen supplementation (AUC 0.704, 95% CI 0.646-0.762) and ARDS (AUC 0.890, 95% CI 0.853-0.928) compared to the CXR score alone. Both the AI and combined models showed good calibration for predicting ARDS (P=.079 and P=.859).ConclusionsThe combined prediction model, comprising the CXR score and clinical information, was externally validated as having acceptable performance in predicting severe illness and excellent performance in predicting ARDS in patients with COVID-19.
| S-EPMC9937110 | biostudies-literature
Project description:BackgroundCoronavirus disease 2019 (COVID-19) is a pandemic disease. Fast and accurate diagnosis of COVID-19 from chest radiography may enable more efficient allocation of scarce medical resources and hence improved patient outcomes. Deep learning classification of chest radiographs may be a plausible step towards this. We hypothesize that bone suppression of chest radiographs may improve the performance of deep learning classification of COVID-19 phenomena in chest radiographs.MethodsTwo bone suppression methods (Gusarev et al. and Rajaraman et al.) were implemented. The Gusarev and Rajaraman methods were trained on 217 pairs of normal and bone-suppressed chest radiographs from the X-ray Bone Shadow Suppression dataset (https://www.kaggle.com/hmchuong/xray-bone-shadow-supression). Two classifier methods with different network architectures were implemented. Binary classifier models were trained on the public RICORD-1c and RSNA Pneumonia Challenge datasets. An external test dataset was created retrospectively from a set of 320 COVID-19 positive patients from Queen Elizabeth Hospital (Hong Kong, China) and a set of 518 non-COVID-19 patients from Pamela Youde Nethersole Eastern Hospital (Hong Kong, China), and used to evaluate the effect of bone suppression on classifier performance. Classification performance, quantified by sensitivity, specificity, negative predictive value (NPV), accuracy and area under the receiver operating curve (AUC), for non-suppressed radiographs was compared to that for bone suppressed radiographs. Some of the pre-trained models used in this study are published at (https://github.com/danielnflam).ResultsBone suppression of external test data was found to significantly (P<0.05) improve AUC for one classifier architecture [from 0.698 (non-suppressed) to 0.732 (Rajaraman-suppressed)]. For the other classifier architecture, suppression did not significantly (P>0.05) improve or worsen classifier performance.ConclusionsRajaraman suppression significantly improved classification performance in one classification architecture, and did not significantly worsen classifier performance in the other classifier architecture. This research could be extended to explore the impact of bone suppression on classification of different lung pathologies, and the effect of other image enhancement techniques on classifier performance.
| S-EPMC9246721 | biostudies-literature
Project description:ImportanceThe improvement of pulmonary nodule detection, which is a challenging task when using chest radiographs, may help to elevate the role of chest radiographs for the diagnosis of lung cancer.ObjectiveTo assess the performance of a deep learning-based nodule detection algorithm for the detection of lung cancer on chest radiographs from participants in the National Lung Screening Trial (NLST).Design, setting, and participantsThis diagnostic study used data from participants in the NLST ro assess the performance of a deep learning-based artificial intelligence (AI) algorithm for the detection of pulmonary nodules and lung cancer on chest radiographs using separate training (in-house) and validation (NLST) data sets. Baseline (T0) posteroanterior chest radiographs from 5485 participants (full T0 data set) were used to assess lung cancer detection performance, and a subset of 577 of these images (nodule data set) were used to assess nodule detection performance. Participants aged 55 to 74 years who currently or formerly (ie, quit within the past 15 years) smoked cigarettes for 30 pack-years or more were enrolled in the NLST at 23 US centers between August 2002 and April 2004. Information on lung cancer diagnoses was collected through December 31, 2009. Analyses were performed between August 20, 2019, and February 14, 2020.ExposuresAbnormality scores produced by the AI algorithm.Main outcomes and measuresThe performance of an AI algorithm for the detection of lung nodules and lung cancer on radiographs, with lung cancer incidence and mortality as primary end points.ResultsA total of 5485 participants (mean [SD] age, 61.7 [5.0] years; 3030 men [55.2%]) were included, with a median follow-up duration of 6.5 years (interquartile range, 6.1-6.9 years). For the nodule data set, the sensitivity and specificity of the AI algorithm for the detection of pulmonary nodules were 86.2% (95% CI, 77.8%-94.6%) and 85.0% (95% CI, 81.9%-88.1%), respectively. For the detection of all cancers, the sensitivity was 75.0% (95% CI, 62.8%-87.2%), the specificity was 83.3% (95% CI, 82.3%-84.3%), the positive predictive value was 3.8% (95% CI, 2.6%-5.0%), and the negative predictive value was 99.8% (95% CI, 99.6%-99.9%). For the detection of malignant pulmonary nodules in all images of the full T0 data set, the sensitivity was 94.1% (95% CI, 86.2%-100.0%), the specificity was 83.3% (95% CI, 82.3%-84.3%), the positive predictive value was 3.4% (95% CI, 2.2%-4.5%), and the negative predictive value was 100.0% (95% CI, 99.9%-100.0%). In digital radiographs of the nodule data set, the AI algorithm had higher sensitivity (96.0% [95% CI, 88.3%-100.0%] vs 88.0% [95% CI, 75.3%-100.0%]; P = .32) and higher specificity (93.2% [95% CI, 89.9%-96.5%] vs 82.8% [95% CI, 77.8%-87.8%]; P = .001) for nodule detection compared with the NLST radiologists. For malignant pulmonary nodule detection on digital radiographs of the full T0 data set, the sensitivity of the AI algorithm was higher (100.0% [95% CI, 100.0%-100.0%] vs 94.1% [95% CI, 82.9%-100.0%]; P = .32) compared with the NLST radiologists, and the specificity (90.9% [95% CI, 89.6%-92.1%] vs 91.0% [95% CI, 89.7%-92.2%]; P = .91), positive predictive value (8.2% [95% CI, 4.4%-11.9%] vs 7.8% [95% CI, 4.1%-11.5%]; P = .65), and negative predictive value (100.0% [95% CI, 100.0%-100.0%] vs 99.9% [95% CI, 99.8%-100.0%]; P = .32) were similar to those of NLST radiologists.Conclusions and relevanceIn this study, the AI algorithm performed better than NLST radiologists for the detection of pulmonary nodules on digital radiographs. When used as a second reader, the AI algorithm may help to detect lung cancer.
| S-EPMC7516603 | biostudies-literature
Project description:ImportanceChest radiography is the most common diagnostic imaging test in medicine and may also provide information about longevity and prognosis.ObjectiveTo develop and test a convolutional neural network (CNN) (named CXR-risk) to predict long-term mortality, including noncancer death, from chest radiographs.Design, setting, and participantsIn this prognostic study, CXR-risk CNN development (n = 41 856) and testing (n = 10 464) used data from the screening radiography arm of the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO) (n = 52 320), a community cohort of asymptomatic nonsmokers and smokers (aged 55-74 years) enrolled at 10 US sites from November 8, 1993, through July 2, 2001. External testing used data from the screening radiography arm of the National Lung Screening Trial (NLST) (n = 5493), a community cohort of heavy smokers (aged 55-74 years) enrolled at 21 US sites from August 2002, through April 2004. Data analysis was performed from January 1, 2018, to May 23, 2019.ExposureDeep learning CXR-risk score (very low, low, moderate, high, and very high) based on CNN analysis of the enrollment radiograph.Main outcomes and measuresAll-cause mortality. Prognostic value was assessed in the context of radiologists' diagnostic findings (eg, lung nodule) and standard risk factors (eg, age, sex, and diabetes) and for cause-specific mortality.ResultsAmong 10 464 PLCO participants (mean [SD] age, 62.4 [5.4] years; 5405 men [51.6%]; median follow-up, 12.2 years [interquartile range, 10.5-12.9 years]) and 5493 NLST test participants (mean [SD] age, 61.7 [5.0] years; 3037 men [55.3%]; median follow-up, 6.3 years [interquartile range, 6.0-6.7 years]), there was a graded association between CXR-risk score and mortality. The very high-risk group had mortality of 53.0% (PLCO) and 33.9% (NLST), which was higher compared with the very low-risk group (PLCO: unadjusted hazard ratio [HR], 18.3 [95% CI, 14.5-23.2]; NLST: unadjusted HR, 15.2 [95% CI, 9.2-25.3]; both P < .001). This association was robust to adjustment for radiologists' findings and risk factors (PLCO: adjusted HR [aHR], 4.8 [95% CI, 3.6-6.4]; NLST: aHR, 7.0 [95% CI, 4.0-12.1]; both P < .001). Comparable results were seen for lung cancer death (PLCO: aHR, 11.1 [95% CI, 4.4-27.8]; NLST: aHR, 8.4 [95% CI, 2.5-28.0]; both P ≤ .001) and for noncancer cardiovascular death (PLCO: aHR, 3.6 [95% CI, 2.1-6.2]; NLST: aHR, 47.8 [95% CI, 6.1-374.9]; both P < .001) and respiratory death (PLCO: aHR, 27.5 [95% CI, 7.7-97.8]; NLST: aHR, 31.9 [95% CI, 3.9-263.5]; both P ≤ .001).Conclusions and relevanceIn this study, the deep learning CXR-risk score stratified the risk of long-term mortality based on a single chest radiograph. Individuals at high risk of mortality may benefit from prevention, screening, and lifestyle interventions.
| S-EPMC6646994 | biostudies-literature