AIDCOV: An Interpretable Artificial Intelligence Model for Detection of COVID-19 from Chest Radiography Images
Authors:
Open Access
- 22 October 2021
- journal article
- research article
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Management Information Systems
- Vol. 12 (4) , 1-20
Abstract
As the Coronavirus Disease 2019 (COVID-19) pandemic continues to grow globally, testing to detect COVID-19 and isolating individuals who test positive remains the primary strategy for preventing community spread of the disease. Therefore, automatic and accurate detection of COVID-19 using medical imaging modalities, which are more widely available and accessible, can be beneficial as an alternative diagnostic tool. In this study, an Artificial Intelligence model for Detection of COVID-19 (AIDCOV) is developed to classify chest radiography images as belonging to a person with either COVID-19, other infections, or no pneumonia (i.e., normal). The hierarchical structure in AIDCOV captures the dependencies among features and improves model performance while an attention mechanism makes the model interpretable and transparent. We used several publicly available datasets of both computed tomography (CT) and X-ray modalities. The main public dataset for chest X-ray images contains 475 COVID-19 samples, 3949 samples from other viral/bacterial infections, and 1583 normal samples. Our model achieves a mean cross-validation accuracy of 98.4%. AIDCOV has a sensitivity of 99.8%, a specificity of 100%, and an F1-score of 99.8% in detecting COVID-19 from X-ray images on that dataset. Using a large dataset of CT images, our model obtained mean cross-validation accuracy and sensitivity of 98.8% and 99.4%, respectively. Additionally, our interpretable model can distinguish subtle signs of infection within each radiography image. Assuming these results hold up in larger datasets obtained from a variety of patients over the world, AIDCOV can be used in conjunction with or instead of RT-PCR testing (where RT-PCR testing is unavailable) to detect and isolate individuals with COVID-19, prevent onward transmission to the general population and healthcare workers, and highlight the areas in the lungs that show signs of COVID-related damage.Keywords
All Related Versions
This publication has 25 references indexed in Scilit:
- Sensitivity of Chest CT for COVID-19: Comparison to RT-PCRRadiology, 2020
- Relation Between Chest CT Findings and Clinical Conditions of Coronavirus Disease (COVID-19) Pneumonia: A Multicenter StudyAmerican Journal of Roentgenology, 2020
- Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in ChinaJAMA, 2020
- Evolution of CT Manifestations in a Patient Recovered from 2019 Novel Coronavirus (2019-nCoV) Pneumonia in Wuhan, ChinaRadiology, 2020
- Epidemiological and clinical features of COVID-19 patients with and without pneumonia in Beijing, ChinamedRxiv, 2020
- Initial Public Health Response and Interim Clinical Guidance for the 2019 Novel Coronavirus Outbreak — United States, December 31, 2019–February 4, 2020MMWR. Morbidity and Mortality Weekly Report, 2020
- Imaging Profile of the COVID-19 Infection: Radiologic Findings and Literature ReviewRadiology: Cardiothoracic Imaging, 2020
- Monitoring of fatigue in radiologists during prolonged image interpretation using fNIRSJapanese Journal of Radiology, 2019
- A survey on deep learning in medical image analysisMedical Image Analysis, 2017
- ImageNet: A large-scale hierarchical image database2009