scholarly journals Encoder-decoder models for chest X-ray report generation perform no better than unconditioned baselines

PLoS ONE ◽  
2021 ◽  
Vol 16 (11) ◽  
pp. e0259639
Author(s):  
Zaheer Babar ◽  
Twan van Laarhoven ◽  
Elena Marchiori
Keyword(s):  
Diagnostic Accuracy ◽  
Input Image ◽  
Radiology Reporting ◽  
High Quality ◽  
Report Generation ◽  
X Ray ◽  
Input And Output ◽  
Patient Diagnosis ◽  
Chest X Ray ◽  
Better Than

High quality radiology reporting of chest X-ray images is of core importance for high-quality patient diagnosis and care. Automatically generated reports can assist radiologists by reducing their workload and even may prevent errors. Machine Learning (ML) models for this task take an X-ray image as input and output a sequence of words. In this work, we show that ML models for this task based on the popular encoder-decoder approach, like ‘Show, Attend and Tell’ (SA&T) have similar or worse performance than models that do not use the input image, called unconditioned baseline. An unconditioned model achieved diagnostic accuracy of 0.91 on the IU chest X-ray dataset, and significantly outperformed SA&T (0.877) and other popular ML models (p-value < 0.001). This unconditioned model also outperformed SA&T and similar ML methods on the BLEU-4 and METEOR metrics. Also, an unconditioned version of SA&T obtained by permuting the reports generated from images of the test set, achieved diagnostic accuracy of 0.862, comparable to that of SA&T (p-value ≥ 0.05).

scholarly journals Artificial Intelligence Framework for Efficient Detection and Classification of Pneumonia Using Chest Radiography Images

2021 ◽  
Author(s):  
Ali Mohammad Alqudah ◽  
Shoroq Qazan ◽  
Ihssan S. Masad
Keyword(s):  
Artificial Intelligence ◽  
Input Image ◽  
Image Size ◽  
X Ray ◽  
Softmax Classifier ◽  
Non Invasive ◽  
Efficient Detection ◽  
X Ray Imaging ◽  
Chest X Ray

Abstract BackgroundChest diseases are serious health problems that threaten the lives of people. The early and accurate diagnosis of such diseases is very crucial in the success of their treatment and cure. Pneumonia is one of the most widely occurred chest diseases responsible for a high percentage of deaths especially among children. So, detection and classification of pneumonia using the non-invasive chest x-ray imaging would have a great advantage of reducing the mortality rates.ResultsThe results showed that the best input image size in this framework was 64 64 based on comparison between different sizes. Using CNN as a deep features extractor and utilizing the 10-fold methodology the propose artificial intelligence framework achieved an accuracy of 94% for SVM and 93.9% for KNN, a sensitivity of 93.33% for SVM and 93.19% for KNN and a specificity of 96.68% for SVM and 96.60% for KNN.ConclusionsIn this study, an artificial intelligence framework has been proposed for the detection and classification of pneumonia based on chest x-ray imaging with different sizes of input images. The proposed methodology used CNN for features extraction that were fed to two different types of classifiers, namely, SVM and KNN; in addition to the SoftMax classifier which is the default CNN classifier. The proposed CNN has been trained, validated, and tested using a large dataset of chest x-ray images contains in total 5852 images.

scholarly journals Detection of COVID-19 Using Transfer Learning and Grad-CAM Visualization on Indigenously Collected X-ray Dataset

Sensors ◽  
2021 ◽  
Vol 21 (17) ◽  
pp. 5813
Author(s):  
Muhammad Umair ◽  
Muhammad Shahbaz Khan ◽  
Fawad Ahmed ◽  
Fatmah Baothman ◽  
Fehaid Alqahtani ◽  
...  
Keyword(s):  
Transfer Learning ◽  
Input Image ◽  
Fine Tuning ◽  
Feature Maps ◽  
Virus Family ◽  
X Ray ◽  
Normal Chest ◽  
Chest X Ray ◽  
Confusion Matrices

The COVID-19 outbreak began in December 2019 and has dreadfully affected our lives since then. More than three million lives have been engulfed by this newest member of the corona virus family. With the emergence of continuously mutating variants of this virus, it is still indispensable to successfully diagnose the virus at early stages. Although the primary technique for the diagnosis is the PCR test, the non-contact methods utilizing the chest radiographs and CT scans are always preferred. Artificial intelligence, in this regard, plays an essential role in the early and accurate detection of COVID-19 using pulmonary images. In this research, a transfer learning technique with fine tuning was utilized for the detection and classification of COVID-19. Four pre-trained models i.e., VGG16, DenseNet-121, ResNet-50, and MobileNet were used. The aforementioned deep neural networks were trained using the dataset (available on Kaggle) of 7232 (COVID-19 and normal) chest X-ray images. An indigenous dataset of 450 chest X-ray images of Pakistani patients was collected and used for testing and prediction purposes. Various important parameters, e.g., recall, specificity, F1-score, precision, loss graphs, and confusion matrices were calculated to validate the accuracy of the models. The achieved accuracies of VGG16, ResNet-50, DenseNet-121, and MobileNet are 83.27%, 92.48%, 96.49%, and 96.48%, respectively. In order to display feature maps that depict the decomposition process of an input image into various filters, a visualization of the intermediate activations is performed. Finally, the Grad-CAM technique was applied to create class-specific heatmap images in order to highlight the features extracted in the X-ray images. Various optimizers were used for error minimization purposes. DenseNet-121 outperformed the other three models in terms of both accuracy and prediction.

scholarly journals Weakly Supervised Contrastive Learning for Chest X-Ray Report Generation

2021 ◽  
Author(s):  
An Yan ◽  
Zexue He ◽  
Xing Lu ◽  
Jiang Du ◽  
Eric Chang ◽  
...  
Keyword(s):  
Report Generation ◽  
X Ray ◽  
Chest X Ray ◽  
Weakly Supervised

scholarly journals Comparison between the molecular diagnostic test and chest X-ray combined with multi-slice spiral CT in the diagnosis of lobar pneumonia

2021 ◽  
Vol 67 (3) ◽  
pp. 129-132
Author(s):  
Huiping Zhu ◽  
Jianjun Dong ◽  
Xufeng Xie ◽  
Lei Wang
Keyword(s):  
Streptococcus Pneumoniae ◽  
Klebsiella Pneumoniae ◽  
Diagnostic Test ◽  
Spiral Ct ◽  
Molecular Diagnostic ◽  
X Ray ◽  
Lobar Pneumonia ◽  
Chest X Ray ◽  
Molecular Diagnostic Test ◽  
Better Than

Lobar pneumonia is an inflammatory condition of the lung that mainly affects the lobes of the lungs and the alveoli, and it is usually caused by a bacterial infection. There are many ways to diagnosis this disease. But an early and accurate method for lobar pneumonia diagnosis has an important role in its treatment. Therefore, in this study, a comparison between the molecular diagnostic test and chest x-ray combined with multi-slice spiral CT was done to find out better diagnosis of lobar pneumonia. For this purpose, 122 individuals suspected of lobar pneumonia were studied by clinical examination, chest X-ray, and multi-slice spiral CT. For the molecular diagnosis test, the multiplex PCR was used for two main causes of the disease, Streptococcus pneumoniae and Klebsiella pneumoniae. Results showed that the specificity for Chest X-ray + Multi-slice Spiral CT had the highest amount (82.8%), but high sensitivity (100%) belonged to a molecular diagnostic test for both bacteria. On the other hand, the sensitivity and specificity of Streptococcus pneumoniae were better than Klebsiella pneumoniae and the possibility of error in Streptococcus pneumoniae was lower than Klebsiella pneumoniae. In general, although the Chest X-ray + Multi-slice Spiral CT method was better than the molecular diagnosis test, it could not identify the causative agent and did not show a difference between pathogens for better antibiotic treatment, and also the possibility of diagnosis is low at the beginning of the disease. Therefore, according to the results of the current study, the best way to diagnose lobar pneumonia is to use both methods, simultaneously.

scholarly journals Artificial Intelligence Framework for Efficient Detection and Classification of Pneumonia Using Chest Radiography Images

2020 ◽  
Author(s):  
Ali Mohammad Alqudah ◽  
Shoroq Qazan ◽  
Ihssan S. Masad
Keyword(s):  
Artificial Intelligence ◽  
Input Image ◽  
Image Size ◽  
X Ray ◽  
Softmax Classifier ◽  
Non Invasive ◽  
Efficient Detection ◽  
X Ray Imaging ◽  
Chest X Ray

Abstract BackgroundChest diseases are serious health problems that threaten the lives of people. The early and accurate diagnosis of such diseases is very crucial in the success of their treatment and cure. Pneumonia is one of the most widely occurred chest diseases responsible for a high percentage of deaths especially among children. So, detection and classification of pneumonia using the non-invasive chest x-ray imaging would have a great advantage of reducing the mortality rates.ResultsThe results showed that the best input image size in this framework was 64 64 based on comparison between different sizes. Using CNN as a deep features extractor and utilizing the 10-fold methodology the propose artificial intelligence framework achieved an accuracy of 94% for SVM and 93.9% for KNN, a sensitivity of 93.33% for SVM and 93.19% for KNN and a specificity of 96.68% for SVM and 96.60% for KNN.ConclusionsIn this study, an artificial intelligence framework has been proposed for the detection and classification of pneumonia based on chest x-ray imaging with different sizes of input images. The proposed methodology used CNN for features extraction that were fed to two different types of classifiers, namely, SVM and KNN; in addition to the SoftMax classifier which is the default CNN classifier. The proposed CNN has been trained, validated, and tested using a large dataset of chest x-ray images contains in total 5852 images.

scholarly journals Contrastive Attention for Automatic Chest X-ray Report Generation

2021 ◽  
Author(s):  
Fenglin Liu ◽  
Changchang Yin ◽  
Xian Wu ◽  
Shen Ge ◽  
Ping Zhang ◽  
...  
Keyword(s):  
Report Generation ◽  
X Ray ◽  
Chest X Ray

scholarly journals Ultrasound to Detect Central Venous Catheter Placement Associated Complications

2020 ◽  
Vol 132 (4) ◽  
pp. 781-794 ◽  
Author(s):  
Jasper M. Smit ◽  
Mark E. Haaksma ◽  
Endry H. T. Lim ◽  
Thei S. Steenvoorden ◽  
Michiel J. Blans ◽  
...  
Keyword(s):  
Central Venous Catheter ◽  
Diagnostic Accuracy ◽  
Central Venous Catheterization ◽  
Venous Catheter ◽  
Reference Standard ◽  
Central Venous ◽  
Central Venous Catheter Placement ◽  
X Ray ◽  
Venous Catheterization ◽  
Chest X Ray

Abstract Background Mechanical complications arising after central venous catheter placement are mostly malposition or pneumothorax. To date, to confirm correct position and detect pneumothorax, chest x-ray film has been the reference standard, while ultrasound might be an accurate alternative. The aim of this study was to evaluate diagnostic accuracy of ultrasound to detect central venous catheter malposition and pneumothorax. Methods This was a prospective, multicenter, diagnostic accuracy study conducted at the intensive care unit and postanesthesia care unit. Adult patients who underwent central venous catheterization of the internal jugular vein or subclavian vein were included. Index test consisted of venous, cardiac, and lung ultrasound. Standard reference test was chest x-ray film. Primary outcome was diagnostic accuracy of ultrasound to detect malposition and pneumothorax; for malposition, sensitivity, specificity, and other accuracy parameters were estimated. For pneumothorax, because chest x-ray film is an inaccurate reference standard to diagnose it, agreement and Cohen’s κ-coefficient were determined. Secondary outcomes were accuracy of ultrasound to detect clinically relevant complications and feasibility of ultrasound. Results In total, 758 central venous catheterizations were included. Malposition occurred in 23 (3.3%) out of 688 cases included in the analysis. Ultrasound sensitivity was 0.70 (95% CI, 0.49 to 0.86) and specificity 0.99 (95% CI, 0.98 to 1.00). Pneumothorax occurred in 5 (0.7%) to 11 (1.5%) out of 756 cases according to chest x-ray film and ultrasound, respectively. In 748 out of 756 cases (98.9%), there was agreement between ultrasound and chest x-ray film with a Cohen’s κ-coefficient of 0.50 (95% CI, 0.19 to 0.80). Conclusions This multicenter study shows that the complication rate of central venous catheterization is low and that ultrasound produces a moderate sensitivity and high specificity to detect malposition. There is moderate agreement with chest x-ray film for pneumothorax. In conclusion, ultrasound is an accurate diagnostic modality to detect malposition and pneumothorax. Editor’s Perspective What We Already Know about This Topic What This Article Tells Us That Is New

DIAGNOSTIC ACCURACY OF CHEST X-RAY IN THE DETECTION OF PNEUMOTHORAX IN BLUNT CHEST TRAUMA PATIENTS WITH CT AS THE GOLD STANDARD

2021 ◽  
Vol 5 (7) ◽  
Author(s):  
Danquale Vance Kynshikhar ◽  
Chaman Lal Kaushal ◽  
Ashwani Tomar ◽  
Neeti Aggarwal
Keyword(s):  
Diagnostic Accuracy ◽  
Predictive Value ◽  
Gold Standard ◽  
Blunt Chest Trauma ◽  
Chest Trauma ◽  
Trauma Patients ◽  
Care Level ◽  
X Ray ◽  
Primary Health ◽  
Chest X Ray

Background: To study the diagnostic accuracy of chest X-ray in the detection of pneumothorax in blunt chest trauma patients with CT as the Gold Standard Methods: The present study was conducted from 31th July 2018 to 30th July 2019. A total of 36 patients were enrolled in the study. Results: On Chest X-Ray Supine AP view, pneumothorax was detected in 11 of 24 patients. The sensitivity of Chest X-Ray Supine AP view was 45.83%, specificity was 100%, positive predictive value (PPV) was 100%, negative predictive value (NPV) was 48% and accuracy was 63.89% for the diagnosis of pneumothorax. Conclusion: A Chest radiograph is the most preferred and relevant primary investigation in the diagnosis of pneumothorax even with the various advanced techniques that are available. X-ray being relatively cheaper and is easily available even at the peripheral centers at the primary health care level. Keywords: X-ray, CT, Pneumothorax

Sign in / Sign up

Export Citation Format

Share Document