DeepCOVID-XR: An Artificial Intelligence Algorithm to Detect COVID-19                     on Chest Radiographs Trained and Tested on a Large U.S. Clinical Data                     Set

Wehbe, Ramsey M.; Sheng, Jiayue; Dutta, Shinjan; Chai, Siyuan; Dravid, Amil; Barutcu, Semih; Wu, Yunan; Cantrell, Donald R.; Xiao, Nicholas; Allen, Bradley D.; MacNealy, Gregory A.; Savas, Hatice; Agrawal, Rishi; Parekh, Nishant; Katsaggelos, Aggelos K.

doi:10.1148/radiol.2020203511

Cited by 138 publications

(136 citation statements)

References 36 publications

Supporting

Mentioning

133

Contrasting

Order By: Relevance

“…This uncertainty notwithstanding, our core observation is that numerous well-cited studies build their datasets by gathering COVID-19-positive radiographs from various sources, as exemplified most thoroughly by the GitHub-COVID repository (in which the image sources and labelling method are clearly documented), and then combine these with COVID-19-negative radiographs originating from the NIH repository, so that we judge that our dataset I fairly represents the key aspects of the data used in these earlier works. Other publications 20,[26][27][28] generally use non-public data, precluding our ability to audit their models, and do not share this issue of strong correlation between data source labels and COVID-19 status. However, based on our review of the literature, we find this issue in an alarming proportion of the publications, including many of the most high-profile studies [4][5][6] .…”

Section: Resultsmentioning

confidence: 99%

AI for radiographic COVID-19 detection selects shortcuts over signal

2021

View full text Add to dashboard Cite

Artificial intelligence (AI) researchers and radiologists have recently reported AI systems that accurately detect COVID-19 in chest radiographs. However, the robustness of these systems remains unclear. Using state-of-the-art techniques in explainable AI, we demonstrate that recent deep learning systems to detect COVID-19 from chest radiographs rely on confounding factors rather than medical pathology, creating an alarming situation in which the systems appear accurate, but fail when tested in new hospitals. We observe that the approach to obtain training data for these AI systems introduces a nearly ideal scenario for AI to learn these spurious 'shortcuts'. Because this approach to data collection has also been used to obtain training data for the detection of COVID-19 in computed tomography scans and for medical imaging tasks related to other diseases, our study reveals a far-reaching problem in medical-imaging AI. In addition, we show that evaluation of a model on external data is insufficient to ensure AI systems rely on medically relevant pathology, because the undesired 'shortcuts' learned by AI systems may not impair performance in new hospitals. These findings demonstrate that explainable AI should be seen as a prerequisite to clinical deployment of machine-learning healthcare models.

show abstract

Section: Resultsmentioning

confidence: 99%

AI for radiographic COVID-19 detection selects shortcuts over signal

2021

View full text Add to dashboard Cite

show abstract

“…As remarked by van Ginneken [ 22 ], in this field, numerous specific dedicated architectures have shown exceptional diagnostic performance, such as the DeepCOVID-XR algorithm [ 23 ]; CAD4COVID-Xray [ 24 ]; and CV19-Net [ 25 ]. The use of three pre-trained convolutional neural networks [ 18 ], AlexNet [ 26 ], Goog-LeNet [ 27 ], and SqueezeNet [ 28 ], was shown to be successful by Pham [ 18 ].…”

Section: Introductionmentioning

confidence: 99%

Lessons from the COVID-19 Pandemic on the Use of Artificial Intelligence in Digital Radiology: The Submission of a Survey to Investigate the Opinion of Insiders

2021

View full text Add to dashboard Cite

The development of artificial intelligence (AI) during the COVID-19 pandemic is there for all to see, and has undoubtedly mainly concerned the activities of digital radiology. Nevertheless, the strong perception in the research and clinical application environment is that AI in radiology is like a hammer in search of a nail. Notable developments and opportunities do not seem to be combined, now, in the time of the COVID-19 pandemic, with a stable, effective, and concrete use in clinical routine; the use of AI often seems limited to use in research applications. This study considers the future perceived integration of AI with digital radiology after the COVID-19 pandemic and proposes a methodology that, by means of a wide interaction of the involved actors, allows a positioning exercise for acceptance evaluation using a general purpose electronic survey. The methodology was tested on a first category of professionals, the medical radiology technicians (MRT), and allowed to (i) collect their impressions on the issue in a structured way, and (ii) collect their suggestions and their comments in order to create a specific tool for this professional figure to be used in scientific societies. This study is useful for the stakeholders in the field, and yielded several noteworthy observations, among them (iii) the perception of great development in thoracic radiography and CT, but a loss of opportunity in integration with non-radiological technologies; (iv) the belief that it is appropriate to invest in training and infrastructure dedicated to AI; and (v) the widespread idea that AI can become a strong complementary tool to human activity. From a general point of view, the study is a clear invitation to face the last yard of AI in digital radiology, a last yard that depends a lot on the opinion and the ability to accept these technologies by the operators of digital radiology.

show abstract

“…Moreover, a study has compared the DeepCOVID-XR AI algorithm’s performance to five radiology experts on 300 images. The accuracy of DeepCOVID-XR’s accuracy was higher (82%) than the consensus (81%) of all five radiologists ( Wehbe et al, 2020 ).…”

Section: Introductionmentioning

confidence: 75%

Digital imaging, technologies and artificial intelligence applications during COVID-19 pandemic

Alhasan

Hasaneen²

2021

Computerized Medical Imaging and Graphics

View full text Add to dashboard Cite

The advancement of technology remained an immersive interest for humankind throughout the past decades. Tech enterprises offered a stream of innovation to address the universal healthcare concerns. The novel coronavirus holds a substantial foothold of planet earth which is combatted by digital interventions across afflicted geographical boundaries and territories. This study aims to explore the trends of modern healthcare technologies and Artificial Intelligence (AI) during COVID-19 crisis, define the concepts and clinical role of AI in the mitigation of COVID-19, investigate and correlate the efficacy of AI-enabled technology in medical imaging during COVID-19 and determine advantages, drawbacks, and challenges of artificial intelligence during COVID-19 pandemic. The paper applied systematic review approach using a deliberated research protocol and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart. Digital technologies can coordinate COVID-19 responses in a cascade fashion that extends from the clinical care facility to the exterior of the pending viral epicenter. With cases of healthcare robotics, aerial drones, and the internet of things as evidentiary examples. PCR tests and medical imaging are the frontier diagnostics of COVID-19. Computed tomography helped to correct the accuracy variation of PCR tests at a clinical sensitivity of 98%. Artificial intelligence can enable autonomous COVID-19 responses using techniques like machine learning. Technology could be an endless system of innovation and opportunities when sourced effectively. Scientists can utilize technology to resolve global concerns challenging the history of tangible possibility. Digital interventions have enhanced the responses to COVID-19, magnified the role of medical imaging amid the COVID-19 crisis and have exposed healthcare professionals to the opportunity of contactless care.

show abstract

DeepCOVID-XR: An Artificial Intelligence Algorithm to Detect COVID-19 on Chest Radiographs Trained and Tested on a Large U.S. Clinical Data Set

Cited by 138 publications

References 36 publications

AI for radiographic COVID-19 detection selects shortcuts over signal

AI for radiographic COVID-19 detection selects shortcuts over signal

Lessons from the COVID-19 Pandemic on the Use of Artificial Intelligence in Digital Radiology: The Submission of a Survey to Investigate the Opinion of Insiders

Digital imaging, technologies and artificial intelligence applications during COVID-19 pandemic

Contact Info

Product

Resources

About