Diagnosis of Coronavirus Disease 2019 Pneumonia by Using Chest                    Radiography: Value of Artificial Intelligence

Zhang, Ran; Tie, Xin; Qi, Zhihua; Bevins, Nicholas; Zhang, Chengzhu; Griner, Dalton; Song, Thomas; Nadig, Jeffrey; Schiebler, Mark L.; Garrett, John W.; Li, Ke; Reeder, Scott B.; Chen, Guang-Hong

doi:10.1148/radiol.2020202944

Cited by 114 publications

(137 citation statements)

References 20 publications

Supporting

Mentioning

136

Contrasting

Order By: Relevance

“…An example of the former application is Truncated Inception Net that is being proposed as a screening tool for COVID-19 outbreak using chest x-rays taking advantage of the AI-driven tools active-learning based on cross-population train/test models that utilize multitudinal and multimodal data [25,26]. An example of the diagnostic application is CV19-Net that was able to diagnose COVID-19 pneumonia and differentiate it from non-COVID-19 pneumonia using CXR with high sensitivity and specificity [27]. An area of future research is to utilize this data set to design an AI based algorithm to predict clinical deterioration using radiographic images.…”

Section: Discussionmentioning

confidence: 99%

Correlation of chest radiography findings with the severity and progression of COVID-19 pneumonia

et al. 2021

View full text Add to dashboard Cite

Aim is to assess the temporal changes and prognostic value of chest radiograph (CXR) in COVID-19 patients. Material and methods: We performed a retrospective study of confirmed COVID-19 patients presented to the emergency between March 07-17, 2020. Clinical & radiological findings were reviewed. Clinical outcomes were classified into critical & non-critical based on severity. Two independent radiologists graded frontal view CXRs into COVID-19 pneumonia category 1 (CoV-P1) with <4 zones and CoV-P2 with ≥4 zones involvement. Interobserver agreement of CoV-P category for the CXR preceding the clinical outcome was assessed using Kendall's τ coefficient. Association between CXR findings and clinical deterioration was calculated along with temporal changes of CXR findings with disease progression. Results: Sixty-two patients were evaluated for clinical features. 56 of these (total: 325 CXRs) were evaluated for radiological findings. Common patterns were progression from lower to upper zones, peripheral to diffuse involvement, & from ground glass opacities to consolidation. Consolidations starting peripherally were noted in 76%, 93% and 48% with critical outcomes, respectively. The interobserver agreement of the CoV-P category of CXRs in the critical and non-critical outcome groups were good and excellent, respectively (τ coefficient = 0.6 & 1.0). Significant association was observed between CoV-P2 and clinical deterioration into a critical status (χ2 = 27.7, p = 0.0001) with high sensitivity (95%) and specificity (71%) within a median interval time of 2 days (range: 0-4 days). Conclusion: Involvement of predominantly 4 or more zones on frontal chest radiograph can be used as predictive prognostic indicator of poorer outcome in COVID-19 patients.

show abstract

Section: Discussionmentioning

confidence: 99%

Correlation of chest radiography findings with the severity and progression of COVID-19 pneumonia

et al. 2021

View full text Add to dashboard Cite

show abstract

“…There are several limitations to this work. First, we did not investigate all possible combinations of augmentation methods, and only relying upon combinations used in four scientific articles (6,(21)(22)(23) published recently. Second, a sample testing dataset of true positives provides an incomplete view of the deep learning model's performance.…”

Section: Limitations and Recommendations For Future Workmentioning

confidence: 99%

“…Our objective is to understand the impact of data augmentation and better understand whether one form of data augmentation is more useful than another. Four data augmentation methods proposed by Yoo et al ( 6 ), Nishio et al ( 21 ), Ahuja et al ( 22 ), and Zhang et al ( 23 ), implemented as follows:…”

Section: Data Augmentationmentioning

confidence: 99%

See 1 more Smart Citation

The Effectiveness of Image Augmentation in Deep Learning Networks for Detecting COVID-19: A Geometric Transformation Perspective

et al. 2021

View full text Add to dashboard Cite

Chest X-ray imaging technology used for the early detection and screening of COVID-19 pneumonia is both accessible worldwide and affordable compared to other non-invasive technologies. Additionally, deep learning methods have recently shown remarkable results in detecting COVID-19 on chest X-rays, making it a promising screening technology for COVID-19. Deep learning relies on a large amount of data to avoid overfitting. While overfitting can result in perfect modeling on the original training dataset, on a new testing dataset it can fail to achieve high accuracy. In the image processing field, an image augmentation step (i.e., adding more training data) is often used to reduce overfitting on the training dataset, and improve prediction accuracy on the testing dataset. In this paper, we examined the impact of geometric augmentations as implemented in several recent publications for detecting COVID-19. We compared the performance of 17 deep learning algorithms with and without different geometric augmentations. We empirically examined the influence of augmentation with respect to detection accuracy, dataset diversity, augmentation methodology, and network size. Contrary to expectation, our results show that the removal of recently used geometrical augmentation steps actually improved the Matthews correlation coefficient (MCC) of 17 models. The MCC without augmentation (MCC = 0.51) outperformed four recent geometrical augmentations (MCC = 0.47 for Data Augmentation 1, MCC = 0.44 for Data Augmentation 2, MCC = 0.48 for Data Augmentation 3, and MCC = 0.49 for Data Augmentation 4). When we retrained a recently published deep learning without augmentation on the same dataset, the detection accuracy significantly increased, with a χMcNemar′s statistic2=163.2 and a p-value of 2.23 × 10−37. This is an interesting finding that may improve current deep learning algorithms using geometrical augmentations for detecting COVID-19. We also provide clinical perspectives on geometric augmentation to consider regarding the development of a robust COVID-19 X-ray-based detector.

show abstract

“…A recent illustrative example of a binary diagnostic task is the study by Zhang et al ( 28 ). The authors studied chest radiographs from 2060 patients with COVID-19 pneumonia and 3148 patients with non-COVID-19 pneumonia.…”

mentioning

confidence: 99%

Artificial Intelligence of COVID-19 Imaging: A Hammer in Search of a Nail

Summers

2021

Radiology

View full text Add to dashboard Cite

Diagnosis of Coronavirus Disease 2019 Pneumonia by Using Chest Radiography: Value of Artificial Intelligence

Abstract: An artificial intelligence algorithm differentiated between COVID-19 pneumonia and non-COVID-19 pneumonia in chest x-ray radiographs with high sensitivity and specificity.

Cited by 114 publications

References 20 publications

Correlation of chest radiography findings with the severity and progression of COVID-19 pneumonia

Correlation of chest radiography findings with the severity and progression of COVID-19 pneumonia

The Effectiveness of Image Augmentation in Deep Learning Networks for Detecting COVID-19: A Geometric Transformation Perspective

Artificial Intelligence of COVID-19 Imaging: A Hammer in Search of a Nail

Contact Info

Product

Resources

About