Predicting Mortality Risk of Covid-19 Patients Using Chest X-Rays
Keywords:
Deep Learning, Convolutional Neural Networks(CNNs), Image Classification, X-Rays, COVID-19, MortalityAbstract
The outbreak of COVID-19 in late 2019 presents a challenging dimension exhibited by its fast and high rate of infection, even though its severity on infected patients is somewhat feeble, especially in people with strong immunity. Due to this rapid infection rate and the limited capacity of healthcare infrastructures, an optimal allocation of health facilities and resources becomes imperative. Consequently, forecasting an individual’s infection severity is crucial to efficiently determine whether the patient requires hospitalization or may be treated as an outpatient to free resources for those desperately deserving. Without such systems, health resources would be inefficiently utilized, resulting in needlessly lost lives. This study attempts to provide a model to determine the mortality of an infected patient on arrival to the health facilities to determine whether or not it is necessary to admit them to intensive care. A Convolutional Neural Networks (CNNs) model based on the ResNet-18 architecture was trained on chest X-rays of COVID-19 patients to estimate their mortality risk, with the best model achieving a training accuracy of 99.6 percent while the validation accuracy achieved 86.7% along with high sensitivity.
References
W. Zhu et al., “Establishing and Managing a Temporary Coronavirus Disease 2019 Specialty Hospital in Wuhan, China,” Anesthesiology, vol. 132, no. 6, pp. 1339–1345, 2020, doi: 10.1097/ALN.0000000000003299.
“Covid-19: Army opens hospitals to civilians, setting up temporary facilities | India News - Times of India.” https://timesofindia.indiatimes.com/india/covid-19-army-opens-hospitals-to-civilians-setting-up-temporary-facilities/articleshow/82318086.cms (accessed Jun. 18, 2022).
Centers for Disease Control and Prevention, “Coronavirus Disease 2019 (COVID-19) – Symptoms.” 2021.
Centers for Disease Control and Prevention, “COVID-19 and Your Health.” 2020.
R. Anand, “Patients turned away as ICU wards fill up in Malaysia.” 2021.
N. Islam et al., “Thoracic imaging tests for the diagnosis of COVID-19.,” Cochrane database Syst. Rev., vol. 11, p. CD013639, Nov. 2020, doi: 10.1002/14651858.CD013639.pub3.
S. Wang et al., “A fully automatic deep learning system for COVID-19 diagnostic and prognostic analysis,” Eur. Respir. J., vol. 56, no. 2, 2020, doi: 10.1183/13993003.00775-2020.
A. R. Kulkarni et al., “Deep learning model to predict the need for mechanical ventilation using chest X-ray images in hospitalised patients with COVID-19,” BMJ Innov., vol. 7, no. 2, pp. 261–270, 2021, doi: 10.1136/bmjinnov-2020-000593.
E. Luz et al., “Towards an effective and efficient deep learning model for COVID-19 patterns detection in X-ray images,” Res. Biomed. Eng., 2021, doi: 10.1007/s42600-021-00151-6.
B. G. S. Cruz, M. N. Bossa, J. Sölter, and A. D. Husch, “Public Covid-19 X-ray datasets and their impact on model bias - a systematic review of a significant problem,” medRxiv, 2021, doi: 10.1101/2021.02.15.21251775.
M. D. Zeiler et al., “On rectified linear units for speech processing,” in 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, 2013, pp. 3517–3521, doi: 10.1109/ICASSP.2013.6638312.
K. Clark et al., “The Cancer Imaging Archive (TCIA): Maintaining and Operating a Public Information Repository,” J. Digit. Imaging 2013 266, vol. 26, no. 6, pp. 1045–1057, Jul. 2013, doi: 10.1007/S10278-013-9622-7.
“Stony Brook University COVID-19 Positive Cases (COVID-19-NY-SBU) - The Cancer Imaging Archive (TCIA) Public Access - Cancer Imaging Archive Wiki.” https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=89096912 (accessed Jun. 18, 2022).
Y. Mohammed, R. S. Batha, A. Olowolayemo, and W. K. Shams, “Deep Learning Models for Prediction of Mortality Risk in Patients with Covid-19 Using Chest X-Rays,” Submitt. to Malaysian J. Comput. Sci., 2022.
