Frontal Chest Research Articles

Dual-Ray Net: Automatic Diagnosis of Thoracic Diseases Using Frontal and Lateral Chest X-rays

Computer-aided diagnosis (CAD) is an important work which can improve the working efficiency of physicians. With the availability of large-scale data sets, several methods have been proposed to classify pathology on chest X-ray images. However, most methods report performance based on a frontal chest radiograph, ignoring the effect of the lateral chest radiography on the diagnosis. This paper puts forward a kind of model, Dual-Ray Net, of a deep convolutional neural network which can deal with the front and lateral chest radiography at the same time by referring the method of using lateral chest radiography to assist diagnose during the diagnosis used by radiologists. Firstly, we evaluated the performance of parameter migration to small data after pre-training for large datasets. The data sets for pre-training are chest X-ray 14 and ImageNet respectively. The results showed that pre-training with chest X-ray 14 performed better than with the generic dataset ImageNet. Secondly, We evaluated the performance of the Frontal and lateral chest radiographs in different modes of input model for the diagnosis of assisted chest disease. Finally, by comparing different feature fusion methods of addition and concatenation, we found that the fusion effect of concatenation is better, which average AUC reached 0.778. The comparison results show that whether it is a public or a non-public dataset, our Dual-Ray Net (concatenation) architecture shows improved performance in recognizing findings in CXR images when compared to applying separate baseline frontal and lateral classes.

Cardiomegaly Detection on Chest Radiographs: Segmentation Versus Classification

In this study, we investigate the detection of cardiomegaly on frontal chest radiographs through two alternative deep-learning approaches - via anatomical segmentation and via image-level classification. We used the publicly available ChestX-ray14 dataset, and obtained heart and lung segmentation annotations for 778 chest radiographs for the development of the segmentation-based approach. The classification-based method was trained with 65k standard chest radiographs with image-level labels. For both approaches, the best models were found through hyperparameter searches where architectural, learning, and regularization related parameters were optimized systematically. The resulting models were tested on a set of 367 held-out images for which cardiomegaly annotations were hand-labeled by two independent expert radiologists. Sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiver operating characteristic curve (AUC) were calculated. The performance of the segmentation-based system with an AUC of 0.977 is significantly better for classifying cardiomegaly than the classification-based model which achieved an AUC of 0.941. Only the segmentation-based model achieved comparable performance to an independent expert reader (AUC of 0.978). We conclude that the segmentation-based model requires 100 times fewer annotated chest radiographs to achieve a substantially better performance, while also producing more interpretable results.

Chest Radiograph Interpretation with Deep Learning Models: Assessment with Radiologist-adjudicated Reference Standards and Population-adjusted Evaluation.

BackgroundDeep learning has the potential to augment the use of chest radiography in clinical radiology, but challenges include poor generalizability, spectrum bias, and difficulty comparing across studies.PurposeTo develop and evaluate deep learning models for chest radiograph interpretation by using radiologist-adjudicated reference standards.Materials and MethodsDeep learning models were developed to detect four findings (pneumothorax, opacity, nodule or mass, and fracture) on frontal chest radiographs. This retrospective study used two data sets. Data set 1 (DS1) consisted of 759 611 images from a multicity hospital network and ChestX-ray14 is a publicly available data set with 112 120 images. Natural language processing and expert review of a subset of images provided labels for 657 954 training images. Test sets consisted of 1818 and 1962 images from DS1 and ChestX-ray14, respectively. Reference standards were defined by radiologist-adjudicated image review. Performance was evaluated by area under the receiver operating characteristic curve analysis, sensitivity, specificity, and positive predictive value. Four radiologists reviewed test set images for performance comparison. Inverse probability weighting was applied to DS1 to account for positive radiograph enrichment and estimate population-level performance.ResultsIn DS1, population-adjusted areas under the receiver operating characteristic curve for pneumothorax, nodule or mass, airspace opacity, and fracture were, respectively, 0.95 (95% confidence interval [CI]: 0.91, 0.99), 0.72 (95% CI: 0.66, 0.77), 0.91 (95% CI: 0.88, 0.93), and 0.86 (95% CI: 0.79, 0.92). With ChestX-ray14, areas under the receiver operating characteristic curve were 0.94 (95% CI: 0.93, 0.96), 0.91 (95% CI: 0.89, 0.93), 0.94 (95% CI: 0.93, 0.95), and 0.81 (95% CI: 0.75, 0.86), respectively.ConclusionExpert-level models for detecting clinically relevant chest radiograph findings were developed for this study by using adjudicated reference standards and with population-level performance estimation. Radiologist-adjudicated labels for 2412 ChestX-ray14 validation set images and 1962 test set images are provided.© RSNA, 2019Online supplemental material is available for this article.See also the editorial by Chang in this issue.

Exploring Large-scale Public Medical Image Datasets

Medical artificial intelligence systems are dependent on well characterized large-scale datasets. Recently released public datasets have been of great interest to the field, but pose specific challenges due to the disconnect they cause between data generation and data usage, potentially limiting the utility of these datasets. We visually explore two large public datasets, to determine how accurate the provided labels are and whether other subtle problems exist. The ChestXray14 dataset contains 112,120 frontal chest films, and the Musculoskeletal Radiology (MURA) dataset contains 40,561 upper limb radiographs. A subset of around 700 images from both datasets was reviewed by a board-certified radiologist, and the quality of the original labels was determined. The ChestXray14 labels did not accurately reflect the visual content of the images, with positive predictive values mostly between 10% and 30% lower than the values presented in the original documentation. There were other significant problems, with examples of hidden stratification and label disambiguation failure. The MURA labels were more accurate, but the original normal/abnormal labels were inaccurate for the subset of cases with degenerative joint disease, with a sensitivity of 60% and a specificity of 82%. Visual inspection of images is a necessary component of understanding large image datasets. We recommend that teams producing public datasets should perform this important quality control procedure and include a thorough description of their findings, along with an explanation of the data generating procedures and labeling rules, in the documentation for their datasets.

Factors associated with chest injuries to front seat occupants in frontal impacts

Objective: Frontal impact chest protection in European cars has been highlighted as an area where possible improvements could be made. The chest is particularly vulnerable in older occupants whose numbers are forecast to increase significantly in the coming decades. This study aimed to provide some direction to areas for possible improvements in frontal crash chest protection.Methods: Real-world crash injury data were interrogated, focusing on cars with current restraint components. The research examined belted front seat occupants in frontal impacts where airbags, pretensioners, and load limiters were present.Results: The chest was the most often injured body region at Abbreviated Injury Scale (AIS) 2+, 3+, and 4+ injury levels. The rate of AIS 2+ and AIS 3+ chest injuries was highest among elderly occupants and lowest among young occupants, and elderly occupants sustained proportionally more severe chest injuries in low/moderate-speed impacts compared to young and middle-aged occupants. However, it should be noted that rates of AIS 2 chest injury were also significantly higher for middle-aged occupants compared to the young. The front passenger seat was shown to be more often associated with significant chest injury than the driver seat. The higher proportion of elderly female occupants was postulated as a reason for this. Skeletal injury was the most frequent type of AIS 2+ chest injury, and the rate of injury for elderly occupants with such injuries was higher than that for young and middle-aged occupants. With the increase in the number of rib fractures, the risk of pulmonary complications and organ injuries tended to increase. The major cause of chest injury was identified as restraining loads transmitted to the chest via the seat belt. The absence of intrusion in the majority of cases suggests an opportunity for the restraint system to better manage the crash pulse, not only for elderly occupants but for those who are middle-aged as well.Conclusions: This study shows the necessity for safety interventions, through new vehicle crashworthiness systems, to improve chest protection, especially for middle-aged and elderly car occupants. Deployment of appropriate injury risk criteria, use of an appropriate dummy thorax, development of a low-energy restraint test, and the development of more adaptive restraints have been discussed as possible solutions to the problem.

A Case Report of Mild TRALI and Literature Review

Abstract Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-associated fatalities. The diagnosis of TRALI is challenging due to lack of pathognomonic laboratory tests and the diagnosis is mainly based on clinical and radiographic findings. The criteria for the diagnosis of TRALI include (1) onset: within 6 hours of transfusion; (2) oxygenation: PAO2/FiO2 ≤300 mm Hg regardless of positive end-expiratory pressure level or oxygen saturation of ≤90% on room air; (3) chest x-ray: bilateral infiltrates on frontal chest radiography; and (4) blood pressure: pulmonary artery occlusion pressure ≤18 mm Hg when measured or no evidence of left atrial hypertension. Although the pathogenesis of TRALI is not fully understood, two different hypothetical pathways have been postulated. One of the pathways is an antibody-mediated, one-hit event (immune TRALI, 89% of cases) and mainly related to the antibodies against the human leukocyte antigen (HLA) or human neutrophil antigen (HNA). The other pathway (nonimmune TRALI) is a two-hit event, and the risk depends on certain predisposing factors, such as underlying pulmonary diseases. However, diagnosis of mild TRALI is challenging due to absence of stridor, wheezing, and rales on physical examination and lack of typical chest x-ray features of TRALI. On the other hand, identification of mild TRALI is important in order to initiate appropriate donor investigation and limit the potential for more severe occurrence in other patients. Recently, we diagnosed a rare case of mild TRALI and report it as follows: A 47-year-old African American male was admitted for relapsed acute myeloid leukemia (AML) who developed therapy-related thrombocytopenia requiring platelet transfusion. Within 15 minutes of starting the transfusion, the patient developed acute dyspnea, hypertension, tachycardia, and hypoxemia with PaO2/FiO2 of 78.9 mm Hg and SaO2 of 92.4%. However, the chest x-ray showed no infiltrates. He was treated with IV steroids, antihistamines, and 15 L of oxygen by face mask with improvement of signs and symptoms within 90 minutes of the reaction. The diagnosis of mild TRALI was confirmed by the detection of anti-HLA class I antibodies in the donor’s plasma. We reviewed the literature, and only four cases of mild TRALI were reported. Mild TRALI can be caused by transfusion of fresh-frozen plasma (2/4), RBCs (1/4), and platelets (1/4). The initial symptoms usually include fever (3/4), chills (4/4), dyspnea (4/4), tachycardia (3/4), and blood pressure changes (2/4 increased in both systolic and diastolic BP, 1/4 increased in systolic BP, 1/4 increased in systolic but decreased in diastolic BP, and 1/4 decreased in diastolic BP). On physical examination, no stridor, wheeze, rales, or urticaria were found (4/4). Laboratory test showed a decrease in the WBC (4/4). Chest x-ray did not show obvious bilateral infiltration and therefore was not required. The symptoms improved rapidly in a few minutes (2/4) to several hours (1/4 in 2 hours and 1/4 in 4 hours). HLA testing showed HLA class I (4/4) or class II (2/4) antibodies in the donor’s plasma. In conclusion, identification of mild TRALI is critical but challenging because diagnosis is based principally on the clinical symptoms and there is no specific laboratory test for mild TRALI. We need to have a high index of suspicion for TRALI in any patient who experiences respiratory symptoms related to transfusion in our daily practice.

Consideration of diagnostic reference levels for pediatric chest X-ray examinations.

The use of diagnostic reference levels (DRLs) is currently recommended, and dose evaluation is considered to be important for establishing a Japanese radiological protection system in radiological medicine. Children, in particular, are sensitive to radiation, and their exposure levels must be taken into account. The DRL for the entrance surface dose (ESD) used in pediatric chest X-ray examinations in Japan is 0.2mGy. However, the bodies of infants and young children show major changes with rapidly developing organs. Thus, the possibility that organ development may also be affected by radiation exposure should be taken into account. Therefore, radiological technologists must be conservative in setting radiographic conditions for pediatric examinations. The objective of this study was to evaluate the doses used in pediatric chest X-ray examinations at our hospital and compare them with the current DRLs, considering the assumption that setting conditions individually for different ages and subject thicknesses and performing more detailed dose evaluations will help reduce radiation exposure. The study was carried out to estimate the ESDs in 163 pediatric patients who underwent frontal or lateral chest X-ray examinations at our hospital. All doses were lower than 0.2mGy, the dose recommended in the Japanese DRLs 2015. The doses showed a strong correlation with age, but a weaker correlation with subject thickness. These results suggest that instead of considering a common DRL for all children, the DRL should be evaluated on the basis of age.

تلوين صورة الأشعة السينية الصدرية الأمامية باستخدام معادلات رياضية تجريبية متنوعة

Chest x-ray medical images are widely used for diagnostic purposes for several reasons including it available, attainable and excellent imaging of hard tissues like bones, but there is difficulty in imaging of the soft tissues, moreover the saturated colour of the x-ray image causing intricacy to read it, therefore, turned attention to the colorization process which can be used to transform the grayscale medical images to the colour medical images. In this paper, a new pseudo-coloring method proposed for colorizing the frontal chest x-ray medical images which are very important for disease detection and recognition. The key idea of this method is applying different and various experimental mathematical equations on the proposed medical sample image, which it belongs to the patient with Hydatid cyst parasite disease in the lung. The counterfeit coloring had intense contribute in enhancing of the visual attractiveness of medical images, the output colored images have a better visual description, it has a possibility to describe hidden information better than a conventional monochromatic image, this is an excellent indicator of the algorithm's suitability for the purpose for which it was formulated.

Assessment of Critical Feeding Tube Malpositions on Radiographs Using Deep Learning

Assess the efficacy of deep convolutional neural networks (DCNNs) in detection of critical enteric feeding tube malpositions on radiographs. 5475 de-identified HIPAA compliant frontal view chest and abdominal radiographs were obtained, consisting of 174 x-rays of bronchial insertions and 5301 non-critical radiographs, including normal course, normal chest, and normal abdominal x-rays. The ground-truth classification for enteric feeding tube placement was performed by two board-certified radiologists. Untrained and pretrained deep convolutional neural network models for Inception V3, ResNet50, and DenseNet 121 were each employed. The radiographs were fed into each deep convolutional neural network, which included untrained and pretrained models. The Tensorflow framework was used for Inception V3, ResNet50, and DenseNet. Images were split into training (4745), validation (630), and test (100). Both real-time and preprocessing image augmentation strategies were performed. Receiver operating characteristic (ROC) and area under the curve (AUC) on the test data were used to assess the models. Statistical differences among the AUCs were obtained. p < 0.05 was considered statistically significant. The pretrained Inception V3, which had an AUC of 0.87 (95 CI; 0.80–0.94), performed statistically significantly better (p < .001) than the untrained Inception V3, with an AUC of 0.60 (95 CI; 0.52–0.68). The pretrained Inception V3 also had the highest AUC overall, as compared with ResNet50 and DenseNet121, with AUC values ranging from 0.82 to 0.85. Each pretrained network outperformed its untrained counterpart. (p < 0.05). Deep learning demonstrates promise in differentiating critical vs. non-critical placement with an AUC of 0.87. Pretrained networks outperformed untrained ones in all cases. DCNNs may allow for more rapid identification and communication of critical feeding tube malpositions.

Deep Learning Method for Automated Classification of Anteroposterior and Posteroanterior Chest Radiographs.

Ensuring correct radiograph view labeling is important for machine learning algorithm development and quality control of studies obtained from multiple facilities. The purpose of this study was to develop and test the performance of a deep convolutional neural network (DCNN) for the automated classification of frontal chest radiographs (CXRs) into anteroposterior (AP) or posteroanterior (PA) views. We obtained 112,120 CXRs from the NIH ChestX-ray14 database, a publicly available CXR database performed in adult (106,179 (95%)) and pediatric (5941 (5%)) patients consisting of 44,810 (40%) AP and 67,310 (60%) PA views. CXRs were used to train, validate, and test the ResNet-18 DCNN for classification of radiographs into anteroposterior and posteroanterior views. A second DCNN was developed in the same manner using only the pediatric CXRs (2885 (49%) AP and 3056 (51%) PA). Receiver operating characteristic (ROC) curves with area under the curve (AUC) and standard diagnostic measures were used to evaluate the DCNN's performance on the test dataset. The DCNNs trained on the entire CXR dataset and pediatric CXR dataset had AUCs of 1.0 and 0.997, respectively, and accuracy of 99.6% and 98%, respectively, for distinguishing between AP and PA CXR. Sensitivity and specificity were 99.6% and 99.5%, respectively, for the DCNN trained on the entire dataset and 98% for both sensitivity and specificity for the DCNN trained on the pediatric dataset. The observed difference in performance between the two algorithms was not statistically significant (p = 0.17). Our DCNNs have high accuracy for classifying AP/PA orientation of frontal CXRs, with only slight reduction in performance when the training dataset was reduced by 95%. Rapid classification of CXRs by the DCNN can facilitate annotation of large image datasets for machine learning and quality assurance purposes.

Chest Radiograph Alone Is Sufficient as the Foreign Body Survey for Children Presenting With Coin Ingestion.

Radiographic survey of the entire aerodigestive tract (nares to anus) is common practice in children presenting to the emergency department following coin ingestion. The purpose of our study was to determine the optimal protocol for radiographic survey post-coin ingestion. We hypothesized that for children presenting with a clear history of coin ingestion a frontal chest radiograph including the entire esophagus is adequate to guide treatment. We reviewed the clinical history and radiographic surveys of 134 patients presenting with suspected or witnessed coin ingestion to the emergency department of a tertiary care pediatric hospital between January 2012 and June 2016. Patient demographics, presenting symptoms, anatomic coverage of survey, type, number and location of radiopaque foreign bodies, intervention, and follow-up were recorded. Coins were identified on radiographic surveys in 109 of 134 patients; 25 of 134 patients had no coin. Of those with coins, none of 109 had coin(s) in the airway, 42 of 109 had coin(s) in the esophagus, and 67 of 109 had coin(s) distal to the esophagus. Of those with esophageal coins, 35 of 42 reported symptoms, 7 of 42 were asymptomatic, 40 of 42 underwent endoscopic coin removal, and 2 of 42 had no intervention. Of 92 of 134 surveyed patients with no coin or coin distal to the esophagus, 30 of 92 reported symptoms, 62 of 92 were asymptomatic, 90 of 92 had no further intervention, and 2 of 92 eventually underwent endoscopic coin removal for specific indications (abdominal pain, delayed passage). When there was no coin found in the esophagus, the negative predictive value for intervention was 97.8%. Following a clear history of coin ingestion, a frontal chest radiograph including the entire length of the esophagus provides sufficient information to direct patient management.

Automated detection of moderate and large pneumothorax on frontal chest X-rays using deep convolutional neural networks: A retrospective study.

BackgroundPneumothorax can precipitate a life-threatening emergency due to lung collapse and respiratory or circulatory distress. Pneumothorax is typically detected on chest X-ray; however, treatment is reliant on timely review of radiographs. Since current imaging volumes may result in long worklists of radiographs awaiting review, an automated method of prioritizing X-rays with pneumothorax may reduce time to treatment. Our objective was to create a large human-annotated dataset of chest X-rays containing pneumothorax and to train deep convolutional networks to screen for potentially emergent moderate or large pneumothorax at the time of image acquisition.Methods and findingsIn all, 13,292 frontal chest X-rays (3,107 with pneumothorax) were visually annotated by radiologists. This dataset was used to train and evaluate multiple network architectures. Images showing large- or moderate-sized pneumothorax were considered positive, and those with trace or no pneumothorax were considered negative. Images showing small pneumothorax were excluded from training. Using an internal validation set (n = 1,993), we selected the 2 top-performing models; these models were then evaluated on a held-out internal test set based on area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and positive predictive value (PPV). The final internal test was performed initially on a subset with small pneumothorax excluded (as in training; n = 1,701), then on the full test set (n = 1,990), with small pneumothorax included as positive. External evaluation was performed using the National Institutes of Health (NIH) ChestX-ray14 set, a public dataset labeled for chest pathology based on text reports. All images labeled with pneumothorax were considered positive, because the NIH set does not classify pneumothorax by size. In internal testing, our “high sensitivity model” produced a sensitivity of 0.84 (95% CI 0.78–0.90), specificity of 0.90 (95% CI 0.89–0.92), and AUC of 0.94 for the test subset with small pneumothorax excluded. Our “high specificity model” showed sensitivity of 0.80 (95% CI 0.72–0.86), specificity of 0.97 (95% CI 0.96–0.98), and AUC of 0.96 for this set. PPVs were 0.45 (95% CI 0.39–0.51) and 0.71 (95% CI 0.63–0.77), respectively. Internal testing on the full set showed expected decreased performance (sensitivity 0.55, specificity 0.90, and AUC 0.82 for high sensitivity model and sensitivity 0.45, specificity 0.97, and AUC 0.86 for high specificity model). External testing using the NIH dataset showed some further performance decline (sensitivity 0.28–0.49, specificity 0.85–0.97, and AUC 0.75 for both). Due to labeling differences between internal and external datasets, these findings represent a preliminary step towards external validation.ConclusionsWe trained automated classifiers to detect moderate and large pneumothorax in frontal chest X-rays at high levels of performance on held-out test data. These models may provide a high specificity screening solution to detect moderate or large pneumothorax on images collected when human review might be delayed, such as overnight. They are not intended for unsupervised diagnosis of all pneumothoraces, as many small pneumothoraces (and some larger ones) are not detected by the algorithm. Implementation studies are warranted to develop appropriate, effective clinician alerts for the potentially critical finding of pneumothorax, and to assess their impact on reducing time to treatment.

Assessment of Convolutional Neural Networks for Automated Classification of Chest Radiographs.

Purpose To assess the ability of convolutional neural networks (CNNs) to enable high-performance automated binary classification of chest radiographs. Materials and Methods In a retrospective study, 216 431 frontal chest radiographs obtained between 1998 and 2012 were procured, along with associated text reports and a prospective label from the attending radiologist. This data set was used to train CNNs to classify chest radiographs as normal or abnormal before evaluation on a held-out set of 533 images hand-labeled by expert radiologists. The effects of development set size, training set size, initialization strategy, and network architecture on end performance were assessed by using standard binary classification metrics; detailed error analysis, including visualization of CNN activations, was also performed. Results Average area under the receiver operating characteristic curve (AUC) was 0.96 for a CNN trained with 200 000 images. This AUC value was greater than that observed when the same model was trained with 2000 images (AUC = 0.84, P < .005) but was not significantly different from that observed when the model was trained with 20 000 images (AUC = 0.95, P > .05). Averaging the CNN output score with the binary prospective label yielded the best-performing classifier, with an AUC of 0.98 (P < .005). Analysis of specific radiographs revealed that the model was heavily influenced by clinically relevant spatial regions but did not reliably generalize beyond thoracic disease. Conclusion CNNs trained with a modestly sized collection of prospectively labeled chest radiographs achieved high diagnostic performance in the classification of chest radiographs as normal or abnormal; this function may be useful for automated prioritization of abnormal chest radiographs. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by van Ginneken in this issue.

Chest Radiographs in Congestive Heart Failure: Visualizing Neural Network Learning.

Purpose To examine Generative Visual Rationales (GVRs) as a tool for visualizing neural network learning of chest radiograph features in congestive heart failure (CHF). Materials and Methods A total of 103 489 frontal chest radiographs in 46 712 patients acquired from January 1, 2007, to December 31, 2016, were divided into a labeled data set (with B-type natriuretic peptide [BNP] result as a marker of CHF) and unlabeled data set (without BNP result). A generative model was trained on the unlabeled data set, and a neural network was trained on the encoded representations of the labeled data set to estimate BNP. The model was used to visualize how a radiograph with high estimated BNP would look without disease (a "healthy" radiograph). An overfitted model was developed for comparison, and 100 GVRs were blindly assessed by two experts for features of CHF. Area under the receiver operating characteristic curve (AUC), κ coefficient, and mixed-effects logistic regression were used for statistical analyses. Results At a cutoff BNP of 100 ng/L as a marker of CHF, the correctly trained model achieved an AUC of 0.82. Assessment of GVRs revealed that the correctly trained model highlighted conventional radiographic features of CHF as reasons for an elevated BNP prediction more frequently than the overfitted model, including cardiomegaly (153 [76.5%] of 200 vs 64 [32%] of 200, respectively; P < .001) and pleural effusions (47 [23.5%] of 200 vs 16 [8%] of 200, respectively; P = .003). Conclusion Features of congestive heart failure on chest radiographs learned by neural networks can be identified using Generative Visual Rationales, enabling detection of bias and overfitted models. © RSNA, 2018 See also the editorial by Ngo in this issue.

Deep learning in chest radiography: Detection of findings and presence of change.

BackgroundDeep learning (DL) based solutions have been proposed for interpretation of several imaging modalities including radiography, CT, and MR. For chest radiographs, DL algorithms have found success in the evaluation of abnormalities such as lung nodules, pulmonary tuberculosis, cystic fibrosis, pneumoconiosis, and location of peripherally inserted central catheters. Chest radiography represents the most commonly performed radiological test for a multitude of non-emergent and emergent clinical indications. This study aims to assess accuracy of deep learning (DL) algorithm for detection of abnormalities on routine frontal chest radiographs (CXR), and assessment of stability or change in findings over serial radiographs.Methods and findingsWe processed 874 de-identified frontal CXR from 724 adult patients (> 18 years) with DL (Qure AI). Scores and prediction statistics from DL were generated and recorded for the presence of pulmonary opacities, pleural effusions, hilar prominence, and enlarged cardiac silhouette. To establish a standard of reference (SOR), two thoracic radiologists assessed all CXR for these abnormalities. Four other radiologists (test radiologists), unaware of SOR and DL findings, independently assessed the presence of radiographic abnormalities. A total 724 radiographs were assessed for detection of findings. A subset of 150 radiographs with follow up examinations was used to asses change over time. Data were analyzed with receiver operating characteristics analyses and post-hoc power analysis.ResultsAbout 42% (305/ 724) CXR had no findings according to SOR; single and multiple abnormalities were seen in 23% (168/724) and 35% (251/724) of CXR. There was no statistical difference between DL and SOR for all abnormalities (p = 0.2–0.8). The area under the curve (AUC) for DL and test radiologists ranged between 0.837–0.929 and 0.693–0.923, respectively. DL had lowest AUC (0.758) for assessing changes in pulmonary opacities over follow up CXR. Presence of chest wall implanted devices negatively affected the accuracy of DL algorithm for evaluation of pulmonary and hilar abnormalities.ConclusionsDL algorithm can aid in interpretation of CXR findings and their stability over follow up CXR. However, in its present version, it is unlikely to replace radiologists due to its limited specificity for categorizing specific findings.

Effectiveness of pneumococcal Haemophilus influenzae protein D conjugate vaccine against pneumonia in children: A cluster-randomised trial

BackgroundPneumococcal conjugate vaccines have potential to prevent significant proportion of childhood pneumonia. Finnish Invasive Pneumococcal disease vaccine trial was designed to assess the vaccine effectiveness (VE) of the 10-valent pneumococcal Haemophilus influenzae protein D conjugate vaccine (PHiD-CV10) against several outcomes. We now report results for pneumonia. MethodsIn this nationwide, cluster-randomised, double-blind trial, children younger than 19 months received PHiD-CV10 in 52 clusters or hepatitis vaccines as control in 26 clusters. Infants younger than 7 months at the first vaccination received either 3+1 or 2+1 vaccination schedule, children aged 7–11 months received 2+1, and those 12–18 months of age two-dose schedule. All hospitalizations and outpatient visits to hospital associated with ICD-10 codes compatible with pneumonia were identified through the National Care Register and 1–3 frontal chest X-ray images per event were collected. External readers who were unaware of the patients’ vaccination status retrospectively interpreted the images. The evaluated outcomes were hospital-diagnosed, hospital-treated pneumonia as primary diagnosis, and radiologically confirmed pneumonia during the blinded, intention-to-treat follow-up period from the first vaccination to the end of 2011. Total VE was calculated as 1 minus rate ratio of all pneumonia episodes. Results47 366 children were enrolled from February 2009, to October 2010. VE against all episodes of hospital-diagnosed pneumonia was 27% (95% confidence interval [CI]: 14%, 38%), 32% (95% CI: 3%, 52%), and 23% (95% CI: −5%, 44%) in subjects enrolled at age <7, 7–11, and 12–18 months, respectively. Corresponding rate reductions were 3.4, 4.7, and 2.5 per 1000 person-years. VE estimates against pneumonia with alveolar consolidation or pleural effusion (WHO criteria) in the three cohorts were 45% (95% CI: 26%, 60%), 56% (95% CI: 14%, 77%), and 48% (95% CI: 2%, 73%), respectively. ConclusionPHiD-CV10 vaccination remarkably reduced disease burden due to pneumonia in infants and young children. Clinical trial registrationMain trial NCT00861380, nested carriage and otitis media trial NCT00839254 (ClinicalTrials.gov).

Feature Selection for Automatic Tuberculosis Screening in Frontal Chest Radiographs.

To detect pulmonary abnormalities such as Tuberculosis (TB), an automatic analysis and classification of chest radiographs can be used as a reliable alternative to more sophisticated and technologically demanding methods (e.g. culture or sputum smear analysis). In target areas like Kenya TB is highly prevalent and often co-occurring with HIV combined with low resources and limited medical assistance. In these regions an automatic screening system can provide a cost-effective solution for a large rural population. Our completely automatic TB screening system is processing the incoming CXRs (chest X-ray) by applying image preprocessing techniques to enhance the image quality followed by an adaptive segmentation based on model selection. The delineated lung regions are described by a multitude of image features. These characteristics are than optimized by a feature selection strategy to provide the best description for the classifier, which will later decide if the analyzed image is normal or abnormal. Our goal is to find the optimal feature set from a larger pool of generic image features, -used originally for problems such as object detection, image retrieval, etc. For performance evaluation measures such as under the curve (AUC) and accuracy (ACC) were considered. Using a neural network classifier on two publicly available data collections, -namely the Montgomery and the Shenzhen dataset, we achieved the maximum area under the curve and accuracy of 0.99 and 97.03%, respectively. Further, we compared our results with existing state-of-the-art systems and to radiologists' decision.

Tracheal buckling: A bizarre but normal finding in chest radiograph of a sick preterm infant

Lateral deviation or buckling of the tracheal air column just above the thoracic inlet is infrequently observed on frontal chest X-rays of infants and young children. Tracheal buckling can cause unnecessary concern and alarm to neonatal and pediatric intensivists when dealing with critically ill infants. We report a case of preterm female infant, who was born at 25 weeks of gestation and required intubation at birth and surfactant administration. She was extubated to continuous positive airway pressure support at 5 h of life. On day 24 of life, she had increasing respiratory distress and X-ray of the chest was done. The X-ray showed a significant right lateral deviation of the trachea, raising the concern of mediastinal mass effect or right upper lobe collapse to the attending neonatologist. However, the radiologist diagnosed the lateral deviation of the trachea as benign tracheal buckling. The infant was intubated and ventilated. Two weeks later, a postextubation chest X-ray showed normal trachea.

Diagnostic Utility of Chest Radiography in Predicting Long-Standing Systemic Arterial Hypertension.

To investigate the association between aortic arch width on frontal chest radiography and systemic hypertension. A total of 200 consecutive patients were included. Relationships between aortic arch width measurement on chest radiography and blood pressure measurement were investigated using Student's t -tests and Fisher's exact tests. Twenty-five patients were normotensive (< 130/90 mmHg), and 175 were hypertensive. Using cut-off values, 136 patients had an aortic arch width ≥ 3.5 cm, and 65 had an aortic arch width ≥ 4 cm. We found a significant relationship between aortic arch width and hypertension (p < 0.001) as well between aortic arch width cut-off values of 3.5 cm and 4 cm and hypertension (p < 0.001 and p < 0.005, respectively). An aortic arch width ≥ 3.5 cm was associated with a positive likelihood ratio (LR) of 2.3, negative LR of 0.39, sensitivity of 73, specificity of 68, positive predictive value of 94, negative predictive value of 26.6, pretest odds of 7, posttest odds of 16, and posttest probability of 94%. An aortic arch width ≥ 4 cm was associated with a positive LR of 4.50, negative LR of 0.70, sensitivity of 36, specificity of 92, positive predictive value of 97, negative predictive value of 17, pretest odds of 7, posttest odds of 31.5, and posttest probability of 97%. Aortic arch width measurement on chest radiography can be used to predict the presence of long-standing systemic arterial hypertension.

An evaluation tool for myofascial adhesions in patients after breast cancer (MAP-BC evaluation tool): Development and interrater reliability.

PurposeTo develop a tool to evaluate myofascial adhesions objectively in patients with breast cancer and to investigate its interrater reliability.Methods1) Development of the evaluation tool. Literature was searched, experts in the field of myofascial therapy were consulted and pilot testing was performed. 2) Thirty patients (63% had a mastectomy, 37% breast-conserving surgery and 97% radiotherapy) with myofascial adhesions were evaluated using the developed tool by 2 independent raters. The Weighted Kappa (WK) and the intra-class correlation coefficient (ICC) were calculated.Results1) The evaluation tool for Myofascial Adhesions in Patients with Breast Cancer (MAP-BC evaluation tool) consisted of the assessment of myofascial adhesions at 7 locations: axillary and breast region scars, musculi pectorales region, axilla, frontal chest wall, lateral chest wall and the inframammary fold. At each location the degree of the myofascial adhesion was scored at three levels (skin, superficial and deep) on a 4-points scale (between no adhesions and very stiff adhesions). Additionally, a total score (0–9) was calculated, i.e. the sum of the different levels of each location. 2) Interrater agreement of the different levels separately was moderate for the axillary and mastectomy scar (WK 0.62–0.73) and good for the scar on the breast (WK >0.75). Moderate agreement was reached for almost all levels of the non-scar locations. Interrater reliability of the total scores was the highest for the scars (ICC 0.82–0.99). At non-scar locations good interrater reliability was reached, except for the inframammary fold (ICC = 0.71).ConclusionsThe total scores of all locations of the MAP-BC evaluation tool had good to excellent interrater reliability, except for the inframammary fold which only reached moderate reliability.