Interpretation Of Chest Radiographs Research Articles

The impact of artificial intelligence on the reading times of radiologists for chest radiographs

Whether the utilization of artificial intelligence (AI) during the interpretation of chest radiographs (CXRs) would affect the radiologists’ workload is of particular interest. Therefore, this prospective observational study aimed to observe how AI affected the reading times of radiologists in the daily interpretation of CXRs. Radiologists who agreed to have the reading times of their CXR interpretations collected from September to December 2021 were recruited. Reading time was defined as the duration in seconds from opening CXRs to transcribing the image by the same radiologist. As commercial AI software was integrated for all CXRs, the radiologists could refer to AI results for 2 months (AI-aided period). During the other 2 months, the radiologists were automatically blinded to the AI results (AI-unaided period). A total of 11 radiologists participated, and 18,680 CXRs were included. Total reading times were significantly shortened with AI use, compared to no use (13.3 s vs. 14.8 s, p < 0.001). When there was no abnormality detected by AI, reading times were shorter with AI use (mean 10.8 s vs. 13.1 s, p < 0.001). However, if any abnormality was detected by AI, reading times did not differ according to AI use (mean 18.6 s vs. 18.4 s, p = 0.452). Reading times increased as abnormality scores increased, and a more significant increase was observed with AI use (coefficient 0.09 vs. 0.06, p < 0.001). Therefore, the reading times of CXRs among radiologists were influenced by the availability of AI. Overall reading times shortened when radiologists referred to AI; however, abnormalities detected by AI could lengthen reading times.

Interrater Reliability of the 2015 Pediatric Acute Lung Injury Consensus Conference Criteria for Pediatric ARDS

Diagnostic guidelines for pediatric acute respiratory distress syndrome (PARDS) were developed at the 2015 Pediatric Acute Lung Injury Consensus Conference (PALICC). Although this was an improvement in creating pediatric-specific diagnostic criteria, there remains potential for variability in identification of PARDS. What is the inter-rater reliability of the 2015 PALICC criteria for diagnosing moderate-severe PARDS? What clinical criteria and patient factors are associated with diagnostic disagreements? Patients with acute hypoxic respiratory failure admitted from 2016-2021 who received invasive mechanical ventilation were retrospectively reviewed by two PICU physicians. Reviewers evaluated whether the patient met the 2015 PALICC definition of moderate-severe PARDS and rated their diagnostic confidence. Inter-rater reliability was measured using Gwet's Agreement Coefficient (AC1). 37 of 191 encounters had a diagnostic disagreement. Inter-rater reliability was "substantial" (AC1=0.74, 95% CI [0.65 - 0.83]). Disagreements were due to different interpretations of chest radiographs (56.8%), ambiguity in origin of pulmonary edema (37.8%), or lack of clarity if patient's current condition was significantly different from baseline (27.0%). Disagreement was more likely in patients who were chronically ventilated (OR 4.66, 95% CI [2.16-10.08], p<0.001), had a primary cardiac admission diagnosis (OR 3.36, 95% CI [1.18 - 9.53], p=0.02), or underwent cardiothoracic surgery during the admission (OR 4.90, 95% CI [1.60 - 15.00], p=0.005). Reviewers were at least moderately confident in their decision 73% of the time, however were less likely to be confident if the patient had cardiac disease or chronic respiratory failure. The inter-rater reliability of the 2015 PALICC criteria for diagnosing moderate-severe PARDS in this cohort was substantial, with diagnostic disagreements commonly due to differences in chest radiograph interpretations. Patients with cardiac disease or chronic respiratory failure were more vulnerable to diagnostic disagreements. More guidance is needed on interpreting chest radiographs and diagnosing PARDS in these subgroups.

Assessing and comparing chest radiograph interpretation in the Department of Internal Medicine at the University of the Witwatersrand medical school, according to seniority.

Chest radiographs are a common diagnostic tool in the internal medicine department, and correct interpretation is imperative for adequate patient management. To determine the diagnostic accuracy of common pathologies in South Africa that are evident on chest radiographs, and to determine whether there are discrepancies according to different levels of qualification of doctors rotating through the internal medicine department, and which factors contribute to an accurate diagnosis. Fifteen chest radiographs with common pathologies were given to all doctors rotating through the Department of Internal Medicine at Chris Hani Baragwanath Academic Hospital, and they were asked to interpret them. Information pertaining to their experience, designation and confidence in chest radiograph interpretation was also obtained. Diagnostic accuracy according to years of experience was as follows: 0 - 5 years 27.0%, 6 - 10 years 43.0%, and >10 years 47.9%. For different designations, accuracy was as follows: consultants 50.5%, registrars 40.9%, medical officers 36.4%, and interns 19.5%. Participants who were confident obtained a mean score of 39.4% and those who were not, a mean score of 31.6%. Chest radiographs are readily accessible and used daily in clinical practice in numerous facilities. An accurate diagnosis is important to provide quality healthcare. Improved training in interpretation for all, but especially for junior doctors, should be a priority in our training facilities. What the study adds. This study tested the diagnostic accuracy with regard to common pathologies present on chest X ray by doctors rotating through, or stationed at the internal medicine department at an academic hospital. Implications of the findings. Interpretation of chest X-rays was generally poor but the study did find that this improves with experience and confidence in diagnostic ability. These findings are significant in that they indicate a need to implement improved teaching programs in radiological interpretation, especially at an undergraduate level.

Methods of Visualizing the Results of an Artificial-Intelligence-Based Computer-Aided Detection System for Chest Radiographs: Effect on the Diagnostic Performance of Radiologists.

It is unclear whether the visualization methods for artificial-intelligence-based computer-aided detection (AI-CAD) of chest radiographs influence the accuracy of readers' interpretation. We aimed to evaluate the accuracy of radiologists' interpretations of chest radiographs using different visualization methods for the same AI-CAD. Initial chest radiographs of patients with acute respiratory symptoms were retrospectively collected. A commercialized AI-CAD using three different methods of visualizing was applied: (a) closed-line method, (b) heat map method, and (c) combined method. A reader test was conducted with five trainee radiologists over three interpretation sessions. In each session, the chest radiographs were interpreted using AI-CAD with one of the three visualization methods in random order. Examination-level sensitivity and accuracy, and lesion-level detection rates for clinically significant abnormalities were evaluated for the three visualization methods. The sensitivity (p = 0.007) and accuracy (p = 0.037) of the combined method are significantly higher than that of the closed-line method. Detection rates using the heat map method (p = 0.043) and the combined method (p = 0.004) are significantly higher than those using the closed-line method. The methods for visualizing AI-CAD results for chest radiographs influenced the performance of radiologists' interpretations. Combining the closed-line and heat map methods for visualizing AI-CAD results led to the highest sensitivity and accuracy of radiologists.

Autonomous Chest Radiograph Reporting Using AI: Estimation of Clinical Impact.

Background Automated interpretation of normal chest radiographs could alleviate the workload of radiologists. However, the performance of such an artificial intelligence (AI) tool compared with clinical radiology reports has not been established. Purpose To perform an external evaluation of a commercially available AI tool for (a) the number of chest radiographs autonomously reported, (b) the sensitivity for AI detection of abnormal chest radiographs, and (c) the performance of AI compared with that of the clinical radiology reports. Materials and Methods In this retrospective study, consecutive posteroanterior chest radiographs from adult patients in four hospitals in the capital region of Denmark were obtained in January 2020, including images from emergency department patients, in-hospital patients, and outpatients. Three thoracic radiologists labeled chest radiographs in a reference standard based on chest radiograph findings into the following categories: critical, other remarkable, unremarkable, or normal (no abnormalities). AI classified chest radiographs as high confidence normal (normal) or not high confidence normal (abnormal). Results A total of 1529 patients were included for analysis (median age, 69 years [IQR, 55-69 years]; 776 women), with 1100 (72%) classified by the reference standard as having abnormal radiographs, 617 (40%) as having critical abnormal radiographs, and 429 (28%) as having normal radiographs. For comparison, clinical radiology reports were classified based on the text and insufficient reports excluded (n = 22). The sensitivity of AI was 99.1% (95% CI: 98.3, 99.6; 1090 of 1100 patients) for abnormal radiographs and 99.8% (95% CI: 99.1, 99.9; 616 of 617 patients) for critical radiographs. Corresponding sensitivities for radiologist reports were 72.3% (95% CI: 69.5, 74.9; 779 of 1078 patients) and 93.5% (95% CI: 91.2, 95.3; 558 of 597 patients), respectively. Specificity of AI, and hence the potential autonomous reporting rate, was 28.0% of all normal posteroanterior chest radiographs (95% CI: 23.8, 32.5; 120 of 429 patients), or 7.8% (120 of 1529 patients) of all posteroanterior chest radiographs. Conclusion Of all normal posteroanterior chest radiographs, 28% were autonomously reported by AI with a sensitivity for any abnormalities higher than 99%. This corresponded to 7.8% of the entire posteroanterior chest radiograph production. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Park in this issue.

The accuracy of intensive care nurses' interpretation of chest radiographs to recognise misplacement of endotracheal and nasogastric tubes after a single training session and comparison with residents' interpretation

BackgroundMisplacements of endotracheal and nasogastric tubes are frequent encounters in critically ill patients. ObjectivesThe purpose of this study was to assess the effectiveness of a single standardised training session on the ability of intensive care registered nurses (RNs) to recognise the misplacement of endotracheal and nasogastric tubes on bedside chest radiographs of patients in intensive care units (ICUs). MethodsIn eight French ICUs, RNs received a 110-min standardised teaching on the position of endotracheal and nasogastric tubes on chest radiographs. Their knowledge was evaluated within the subsequent weeks. For 20 chest radiographs, each with an endotracheal and nasogastric tube, RNs had to indicate whether each tube was in the proper or incorrect position. Training success was defined as >90% for the lower bound of the 95% confidence interval (95% CI) of the mean correct response rate (CRR). Residents of the participating ICUs underwent the same evaluation (without prior specific training). ResultsIn total, 181 RNs were trained and evaluated and 110 residents were evaluated. The global mean CRR for RNs was 84.6% (95% CI: 83.3–85.9), significantly higher than for residents (81.4% [95% CI: 79.7–83.2]) (P < 0.0001). The mean CRR for RNs and residents was 95.9% (93.9–98.0) and 97.0% (94.7–99.3) for misplaced nasogastric tubes (P = 0.54), 86.8% (85.2–88.5) and 82.6% (79.4–85.7) (P = 0.07) for nasogastric tubes in the correct position, 86.6% (83.8–89.3) and 62.7% (57.9–67.5) for misplaced endotracheal tubes (P < 0.0001), and 79.1% (76.6–81.6) and 84.7% (82.1–87.2) for endotracheal tubes in the correct position (P = 0.01), respectively. ConclusionsThe ability of trained RNs to detect tube misplacement did not reach the predetermined arbitrary level, indicating training success. Their mean CRR was higher than that for residents and was considered satisfactory for detecting misplaced nasogastric tubes. This finding is encouraging but insufficient to ensure patient safety. Transferring responsibility for reading radiographs to detect the misplacement of endotracheal tubes to intensive care RNs will need a more advanced or more in-depth teaching method.

ACCURACY OF PATIENTS AGE ASSESSMENT FROM FRONTAL CHEST RADIOGRAPHS

Background: Chest radiograph interpretation can be aided by knowing the patient'sage. Patients’ age is often assessed by radiologists from costal cartilage calcification assessment ofchest radiographs with anteroposterior views. Objectives: Therefore, we performed this research todetermine the precision of these radiologists in age estimation from CXRs. Methods: Ten radiologyexperts were selected to evaluate 3500 chest digital radiography with posterior-anterior images, inNational Institute of Child Health Karachi/JSMU from January 2022 to 2023. The most importantinclusion criteria were selecting normal or nearly normal radiographs in the study. Radiologists wereblind to patients actual age and were requested to determine patient’s age to closest decade from CXR.The respondents entered their responses in separate Excel spreadsheets. Results: A sum of 3,500CXRs was interpreted by radiologists, out of which 32.14% CXRs were correctly interpreteddetermining age at SD + 3 years, whereas 2375 CXRs were either overrated or underestimated(p&lt;0.05) and misinterpreted the age of patients (2375/3500; 67.85%). Conclusion: Overall ageassessment from a frontal CXR was only 32.14% in our study; considerable disparities were identifiedin age estimation using CXRs. Yet, it would be fascinating to expand this academic endeavor throughartificial intelligence tools and possibly improving the precision of patient age prediction from CXR.This approach for determining the age will be useful for screening tests in the prospect.

Cardiovascular abnormalities in chest radiographs of children with pneumonia, Uganda.

To describe chest radiograph findings among children hospitalized with clinically diagnosed severe pneumonia and hypoxaemia at three tertiary facilities in Uganda. The study involved clinical and radiograph data on a random sample of 375 children aged 28days to 12years enrolled in the Children's Oxygen Administration Strategies Trial in 2017. Children were hospitalized with a history of respiratory illness and respiratory distress complicated by hypoxaemia, defined as a peripheral oxygen saturation (SpO2) < 92%. Radiologists blinded to clinical findings interpreted chest radiographs using standardized World Health Organization method for paediatric chest radiograph reporting. We report clinical and chest radiograph findings using descriptive statistics. Overall, 45.9% (172/375) of children had radiological pneumonia, 36.3% (136/375) had a normal chest radiograph and 32.8% (123/375) had other radiograph abnormalities, with or without pneumonia. In addition, 28.3% (106/375) had a cardiovascular abnormality, including 14.9% (56/375) with both pneumonia and another abnormality. There was no significant difference in the prevalence of radiological pneumonia or of cardiovascular abnormalities or in 28-day mortality between children with severe hypoxaemia (SpO2: < 80%) and those with mild hypoxaemia (SpO2:80 to < 92%). Cardiovascular abnormalities were relatively common among children hospitalized with severe pneumonia in Uganda. The standard clinical criteria used to identify pneumonia among children in resource-poor settings were sensitive but lacked specificity. Chest radiographs should be performed routinely for all children with clinical signs of severe pneumonia because it provides useful information on both cardiovascular and respiratory systems.

Validation of a Deep Learning Model for Detecting Chest Pathologies from Digital Chest Radiographs.

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.

The prevalence, risk factors, and public health consequences of peripheral lymph node–associated clinical and subclinical pulmonary tuberculosis

ObjectivesRelatively little is known about the prevalence, risk factors, and public health consequences of peripheral lymph node (PLN)–associated pulmonary tuberculosis (PTB). MethodsWe developed a 10-year (2010-2019) population-based cohort of PLNTB patients in Canada. We used systematically collected primary source data and expert reader chest radiograph interpretations in a multivariable logistic regression to determine associations between sputum culture positivity and demographic, clinical, and radiographic features. Public health risks were estimated among contacts of PLNTB patients. ResultsThere were 306 patients with PLNTB, among whom 283 (92.5%) were 15-64 years of age, 159 (52.0%) were female, and 293 (95.8%) were foreign-born. Respiratory symptoms were present in 21.6%, and abnormal chest radiograph in 23.2%. Sputum culture positivity ranged from 12.9% in patients with no symptoms and normal lung parenchyma to 66.7% in patients with both. Respiratory symptoms, abnormal lung parenchyma, and HIV-coinfection (borderline) were independent predictors of sputum culture positivity (odds ratio [OR] 2.24 [95% confidence interval [CI] 1.15-4.39], P = 0.01, OR 4.78 [95% CI 2.41-9.48], P < 0.001, and OR 2.54 [95% CI 0.99-6.52], P = 0.05), respectively. Among contacts of sputum culture-positive PLNTB patients, one secondary case and 16 new infections were identified. ConclusionIsochronous PTB is common in PLNTB patients. Routine screening of PLNTB patients for PTB is strongly recommended.

Image augmentation and automated measurement of endotracheal-tube-to-carina distance on chest radiographs in intensive care unit using a deep learning model with external validation

BackgroundChest radiographs are routinely performed in intensive care unit (ICU) to confirm the correct position of an endotracheal tube (ETT) relative to the carina. However, their interpretation is often challenging and requires substantial time and expertise. The aim of this study was to propose an externally validated deep learning model with uncertainty quantification and image segmentation for the automated assessment of ETT placement on ICU chest radiographs.MethodsThe CarinaNet model was constructed by applying transfer learning to the RetinaNet model using an internal dataset of ICU chest radiographs. The accuracy of the model in predicting the position of the ETT tip and carina was externally validated using a dataset of 200 images extracted from the MIMIC-CXR database. Uncertainty quantification was performed using the level of confidence in the ETT–carina distance prediction. Segmentation of the ETT was carried out using edge detection and pixel clustering.ResultsThe interrater agreement was 0.18 cm for the ETT tip position, 0.58 cm for the carina position, and 0.60 cm for the ETT–carina distance. The mean absolute error of the model on the external test set was 0.51 cm for the ETT tip position prediction, 0.61 cm for the carina position prediction, and 0.89 cm for the ETT–carina distance prediction. The assessment of ETT placement was improved by complementing the human interpretation of chest radiographs with the CarinaNet model.ConclusionsThe CarinaNet model is an efficient and generalizable deep learning algorithm for the automated assessment of ETT placement on ICU chest radiographs. Uncertainty quantification can bring the attention of intensivists to chest radiographs that require an experienced human interpretation. Image segmentation provides intensivists with chest radiographs that are quickly interpretable and allows them to immediately assess the validity of model predictions. The CarinaNet model is ready to be evaluated in clinical studies.Graphical

Development of a novel artificial intelligence algorithm to detect pulmonary nodules on chest radiography.

In this study, we aimed to develop a novel artificial intelligence (AI) algorithm to support pulmonary nodule detection, which will enable physicians to efficiently interpret chest radiographs for lung cancer diagnosis. We analyzed chest X-ray images obtained from a health examination center in Fukushima and the National Institutes of Health (NIH) Chest X-ray 14 dataset. We categorized these data into two types: type A included both Fukushima and NIH datasets, and type B included only the Fukushima dataset. We also demonstrated pulmonary nodules in the form of a heatmap display on each chest radiograph and calculated the positive probability score as an index value. Our novel AI algorithms had a receiver operating characteristic (ROC) area under the curve (AUC) of 0.74, a sensitivity of 0.75, and a specificity of 0.60 for the type A dataset. For the type B dataset, the respective values were 0.79, 0.72, and 0.74. The algorithms in both the type A and B datasets were superior to the accuracy of radiologists and similar to previous studies. The proprietary AI algorithms had a similar accuracy for interpreting chest radiographs when compared with previous studies and radiologists. Especially, we could train a high quality AI algorithm, even with our small type B data set. However, further studies are needed to improve and further validate the accuracy of our AI algorithm.

Effects of Implementing Artificial Intelligence-Based Computer-Aided Detection for Chest Radiographs in Daily Practice on the Rate of Referral to Chest Computed Tomography in Pulmonology Outpatient Clinic.

The clinical impact of artificial intelligence-based computer-aided detection (AI-CAD) beyond diagnostic accuracy remains uncertain. We aimed to investigate the influence of the clinical implementation of AI-CAD for chest radiograph (CR) interpretation in daily practice on the rate of referral for chest computed tomography (CT). AI-CAD was implemented in clinical practice at the Seoul National University Hospital. CRs obtained from patients who visited the pulmonology outpatient clinics before (January-December 2019) and after (January-December 2020) implementation were included in this study. After implementation, the referring pulmonologist requested CRs with or without AI-CAD analysis. We conducted multivariable logistic regression analyses to evaluate the associations between using AI-CAD and the following study outcomes: the rate of chest CT referral, defined as request and actual acquisition of chest CT within 30 days after CR acquisition, and the CT referral rates separately for subsequent positive and negative CT results. Multivariable analyses included various covariates such as patient age and sex, time of CR acquisition (before versus after AI-CAD implementation), referring pulmonologist, nature of the CR examination (baseline versus follow-up examination), and radiology reports presence at the time of the pulmonology visit. A total of 28546 CRs from 14565 patients (mean age: 67 years; 7130 males) and 25888 CRs from 12929 patients (mean age: 67 years; 6435 males) before and after AI-CAD implementation were included. The use of AI-CAD was independently associated with increased chest CT referrals (odds ratio [OR], 1.33; P = 0.008) and referrals with subsequent negative chest CT results (OR, 1.46; P = 0.005). Meanwhile, referrals with positive chest CT results were not significantly associated with AI-CAD use (OR, 1.08; P = 0.647). The use of AI-CAD for CR interpretation in pulmonology outpatients was independently associated with an increased frequency of overall referrals for chest CT scans and referrals with subsequent negative results.

A Performance-Based Competency Assessment of Pediatric Chest Radiograph Interpretation Among Practicing Physicians.

There is limited knowledge on pediatric chest radiograph (pCXR) interpretation skill among practicing physicians. We systematically determined baseline interpretation skill, the number of pCXR cases physicians required complete to achieve a performance benchmark, and which diagnoses posed the greatest diagnostic challenge. Physicians interpreted 434 pCXR cases via a web-based platform until they achieved a performance benchmark of 85% accuracy, sensitivity, and specificity. Interpretation difficulty scores for each case were derived by applying one-parameter item response theory to participant data. We compared interpretation difficulty scores across diagnostic categories and described the diagnoses of the 30% most difficult-to-interpret cases. 240 physicians who practice in one of three geographic areas interpreted cases, yielding 56,833 pCXR case interpretations. The initial diagnostic performance (first 50 cases) of our participants demonstrated an accuracy of 68.9%, sensitivity of 69.4%, and a specificity of 68.4%. The median number of cases completed to achieve the performance benchmark was 102 (interquartile range 69, 176; min, max, 54, 431). Among the 30% most difficult-to-interpret cases, 39.2% were normal pCXR and 32.3% were cases of lobar pneumonia. Cases with a single trauma-related imaging finding, cardiac, hilar, and diaphragmatic pathologies were also among the most challenging. At baseline, practicing physicians misdiagnosed about one-third of pCXR and there was up to an eight-fold difference between participants in number of cases completed to achieve the standardized performance benchmark. We also identified the diagnoses with the greatest potential for educational intervention.

Artificial Intelligence for Clinical Interpretation of Bedside Chest Radiographs.

Background Supine chest radiography for bedridden patients in intensive care units (ICUs) is one of the most frequently ordered imaging studies worldwide. Purpose To evaluate the diagnostic performance of a neural network-based model that is trained on structured semiquantitative radiologic reports of bedside chest radiographs. Materials and Methods For this retrospective single-center study, children and adults in the ICU of a university hospital who had been imaged using bedside chest radiography from January 2009 to December 2020 were reported by using a structured and itemized template. Ninety-eight radiologists rated the radiographs semiquantitatively for the severity of disease patterns. These data were used to train a neural network to identify cardiomegaly, pulmonary congestion, pleural effusion, pulmonary opacities, and atelectasis. A held-out internal test set (100 radiographs from 100 patients) that was assessed independently by an expert panel of six radiologists provided the ground truth. Individual assessments by each of these six radiologists, by two nonradiologist physicians in the ICU, and by the neural network were compared with the ground truth. Separately, the nonradiologist physicians assessed the images without and with preliminary readings provided by the neural network. The weighted Cohen κ coefficient was used to measure agreement between the readers and the ground truth. Results A total of 193 566 radiographs in 45 016 patients (mean age, 66 years ± 16 [SD]; 61% men) were included and divided into training (n = 122 294; 64%), validation (n = 31 243; 16%), and test (n = 40 029; 20%) sets. The neural network exhibited higher agreement with a majority vote of the expert panel (κ = 0.86) than each individual radiologist compared with the majority vote of the expert panel (κ = 0.81 to ≤0.84). When the neural network provided preliminary readings, the reports of the nonradiologist physicians improved considerably (aided vs unaided, κ = 0.87 vs 0.79, respectively; P < .001). Conclusion A neural network trained with structured semiquantitative bedside chest radiography reports allowed nonradiologist physicians improved interpretations compared with the consensus reading of expert radiologists. © RSNA, 2022 Supplemental material is available for this article. See also the editorial by Wielpütz in this issue.

Bulla or pneumothorax? - A radiological dilemma

Diagnosing and treating pneumothorax can be an emergency life-saving situation. However, differentiating pneumothorax from bulla may be difficult in some circumstances and requires careful attention to details on the chest radiograph. We report a case of loculated hydropneumothorax which initially presented like a bulla. We wish to highlight the atypical radiographic presentation of pneumothorax empasising the importance of careful interpretation of chest radiograph. Follow up radiologic imaging or computed tomogram/magnetic resonance imaging would be required in such cases if the clinical and radiographic diagnosis do not match.

Deep learning to detect significant coronary artery disease from plain chest radiographs AI4CAD

BackgroundThe predictive role of chest radiographs in patients with suspected coronary artery disease (CAD) is underestimated and may benefit from artificial intelligence (AI) applications. ObjectivesTo train, test, and validate a deep learning (DL) solution for detecting significant CAD based on chest radiographs. MethodsData of patients referred for angina and undergoing chest radiography and coronary angiography were analysed retrospectively. A deep convolutional neural network (DCNN) was designed to detect significant CAD from posteroanterior/anteroposterior chest radiographs. The DCNN was trained for severe CAD binary classification (absence/presence). Coronary angiography reports were the ground truth. Stenosis severity of ≥70% for non–left main vessels and ≥ 50% for left main defined severe CAD. ResultsInformation of 7728 patients was reviewed. Severe CAD was present in 4091 (53%). Patients were randomly divided for algorithm training (70%; n = 5454) and fine-tuning/model validation (10%; n = 773). Internal clinical validation (model testing) was performed with the remaining patients (20%; n = 1501). At binary logistic regression, DCNN prediction was the strongest severe CAD predictor (p < 0.0001; OR: 1.040; CI: 1.032–1.048). Using a high sensitivity operating cut-point, the DCNN had a sensitivity of 0.90 to detect significant CAD (specificity 0.31; AUC 0.73; 95% CI DeLong, 0.69–0.76). Adding to the AI chest radiograph interpretation angina status improved the prediction (AUC 0.77; 95% CI DeLong, 0.74–0.80). ConclusionAI-read chest radiographs could be used to pre-test significant CAD probability in patients referred for suspected angina. Further studies are required to externally validate our algorithm, develop a clinically applicable tool, and support CAD screening in broader settings.

Deep learning to detect significant coronary artery disease from plain chest radiographs

Abstract Background Coronary artery disease (CAD) is the single leading cause of mortality, premature death, and morbidity worldwide. Artificial intelligence (AI) could help identify markers present within first-line diagnostic imaging routinely performed in patients referred for suspected angina, such as chest radiographs. Purpose To train, test, and validate a deep learning (DL) algorithm for detecting the presence of significant CAD based on chest radiographs. Methods Data of patients undergoing chest radiography and coronary angiography were retrospectively analysed. A deep convolutional neural network (DCNN) was designed to detect significant CAD from the patient posteroanterior/anteroposterior chest radiograph. The DCNN was trained for binary classification of severe CAD absence/presence (at least one diseased coronary vessel with ≥70% stenosis). Coronary angiography reports were used as the ground truth. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) of the DCNN were calculated. Multivariate analysis was performed to identify independent correlation among the presence of significant CAD (dependent variable), DCNN prediction, and CAD risk factors. Results Information of 7728 patients referred for suspected angina was reviewed. Severe CAD was present in 4482 patients (58%; 1% left main, 28% one vessel, 16% two vessels, and 12% 3 vessels). Patients were randomly divided for training (70%; n=5454) and fine-tuning/testing (10%; n=773) of the algorithm. Internal validation was performed with the remaining patients (20%; n=1501). At binary logistic regression, the DCNN prediction was the strongest independent determinant of severe CAD (p&amp;lt;0.0001; OR: 52.8; CI: 25.1–110.9). Age (p=0.008; OR: 1.01; CI: 1.0–1.02) and Diamond-Forrester score (p&amp;lt;0.0001; OR: 1.022; CI: 1.018–1.026) were also independently related to CAD, although with lower significance and odds-ratios. Using an operating cut-point with high sensitivity, the DCNN had a sensitivity of 0.90 and specificity of 0.31 to detect significant CAD in the internal validation group (AUC 0.73; 95% CI DeLong, 0.69–0.76). Adding to the AI chest radiograph interpretation, patient age and angina status improved the prediction (AUC 0.77; 95% CI DeLong, 0.74–0.80). Conclusion The chest radiograph is ubiquitous and carries a plethora of information concerning the patient's health status, including direct and indirect signs of CAD. Our DL algorithm can predict, with high sensitivity, the presence of severe CAD in patients referred for suspected angina. It could be used to pre-test significant CAD probability in outpatient clinics, emergency room settings, and CAD screening in more extensive settings. Further studies are required to externally validate the algorithm and develop a clinically applicable tool. Funding Acknowledgement Type of funding sources: None.

Deep Learning Detection of Active Pulmonary Tuberculosis at Chest Radiography Matched the Clinical Performance of Radiologists.

Background The World Health Organization (WHO) recommends chest radiography to facilitate tuberculosis (TB) screening. However, chest radiograph interpretation expertise remains limited in many regions. Purpose To develop a deep learning system (DLS) to detect active pulmonary TB on chest radiographs and compare its performance to that of radiologists. Materials and Methods A DLS was trained and tested using retrospective chest radiographs (acquired between 1996 and 2020) from 10 countries. To improve generalization, large-scale chest radiograph pretraining, attention pooling, and semisupervised learning ("noisy-student") were incorporated. The DLS was evaluated in a four-country test set (China, India, the United States, and Zambia) and in a mining population in South Africa, with positive TB confirmed with microbiological tests or nucleic acid amplification testing (NAAT). The performance of the DLS was compared with that of 14 radiologists. The authors studied the efficacy of the DLS compared with that of nine radiologists using the Obuchowski-Rockette-Hillis procedure. Given WHO targets of 90% sensitivity and 70% specificity, the operating point of the DLS (0.45) was prespecified to favor sensitivity. Results A total of 165 754 images in 22 284 subjects (mean age, 45 years; 21% female) were used for model development and testing. In the four-country test set (1236 subjects, 17% with active TB), the receiver operating characteristic (ROC) curve of the DLS was higher than those for all nine India-based radiologists, with an area under the ROC curve of 0.89 (95% CI: 0.87, 0.91). Compared with these radiologists, at the prespecified operating point, the DLS sensitivity was higher (88% vs 75%, P < .001) and specificity was noninferior (79% vs 84%, P = .004). Trends were similar within other patient subgroups, in the South Africa data set, and across various TB-specific chest radiograph findings. In simulations, the use of the DLS to identify likely TB-positive chest radiographs for NAAT confirmation reduced the cost by 40%-80% per TB-positive patient detected. Conclusion A deep learning method was found to be noninferior to radiologists for the determination of active tuberculosis on digital chest radiographs. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by van Ginneken in this issue.

Association of Artificial Intelligence–Aided Chest Radiograph Interpretation With Reader Performance and Efficiency

The efficient and accurate interpretation of radiologic images is paramount. To evaluate whether a deep learning-based artificial intelligence (AI) engine used concurrently can improve reader performance and efficiency in interpreting chest radiograph abnormalities. This multicenter cohort study was conducted from April to November 2021 and involved radiologists, including attending radiologists, thoracic radiology fellows, and residents, who independently participated in 2 observer performance test sessions. The sessions included a reading session with AI and a session without AI, in a randomized crossover manner with a 4-week washout period in between. The AI produced a heat map and the image-level probability of the presence of the referrable lesion. The data used were collected at 2 quaternary academic hospitals in Boston, Massachusetts: Beth Israel Deaconess Medical Center (The Medical Information Mart for Intensive Care Chest X-Ray [MIMIC-CXR]) and Massachusetts General Hospital (MGH). The ground truths for the labels were created via consensual reading by 2 thoracic radiologists. Each reader documented their findings in a customized report template, in which the 4 target chest radiograph findings and the reader confidence of the presence of each finding was recorded. The time taken for reporting each chest radiograph was also recorded. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC) were calculated for each target finding. A total of 6 radiologists (2 attending radiologists, 2 thoracic radiology fellows, and 2 residents) participated in the study. The study involved a total of 497 frontal chest radiographs-247 from the MIMIC-CXR data set (demographic data for patients were not available) and 250 chest radiographs from MGH (mean [SD] age, 63 [16] years; 133 men [53.2%])-from adult patients with and without 4 target findings (pneumonia, nodule, pneumothorax, and pleural effusion). The target findings were found in 351 of 497 chest radiographs. The AI was associated with higher sensitivity for all findings compared with the readers (nodule, 0.816 [95% CI, 0.732-0.882] vs 0.567 [95% CI, 0.524-0.611]; pneumonia, 0.887 [95% CI, 0.834-0.928] vs 0.673 [95% CI, 0.632-0.714]; pleural effusion, 0.872 [95% CI, 0.808-0.921] vs 0.889 [95% CI, 0.862-0.917]; pneumothorax, 0.988 [95% CI, 0.932-1.000] vs 0.792 [95% CI, 0.756-0.827]). AI-aided interpretation was associated with significantly improved reader sensitivities for all target findings, without negative impacts on the specificity. Overall, the AUROCs of readers improved for all 4 target findings, with significant improvements in detection of pneumothorax and nodule. The reporting time with AI was 10% lower than without AI (40.8 vs 36.9 seconds; difference, 3.9 seconds; 95% CI, 2.9-5.2 seconds; P < .001). These findings suggest that AI-aided interpretation was associated with improved reader performance and efficiency for identifying major thoracic findings on a chest radiograph.