Chest CT images of COVID-19 patients showed some similarities with images of patients suffering from the same viral family (SARS & MERS)
-	 predominance in the peripheral regions
-	 associated with crazy-paving pattern
-	 “absence of pulmonary cavitations, pleural effusions, lymphadenopathy”



Chest CT images of COVID-19 patients that are different from other coronaviruses
-	 showed no definite lobar or craniocaudal distribution
-	 no discrete pulmonary emphysema or fibrosis


Chest CT Imaging Features of COVID-19 Patients from Study #1

- Chest CT Imaging Features of COVID-19 Patients from Study #1

University of California, Santa Cruz

Smith College

Sensitivity and specificity of CT scans varies with COVID-19, therefore cannot be the determinant factor in diagnosing infection. Produce a medical image using multiple X-ray measurements from different angles of the patient's chest.

CT Scans

Chung, M., Bernheim, A., Mei, X., Zhang, M., Huang, M., Zeng, X., Cui, Z., Xu, W., Yang, Y., Fayad, Z., Jacobi, A., Li, K., Li, S., Shan, H.  CT Imaging Features of 2019 Novel Coronavirus (2019-nCoV). Radiology. 2020. doi: 10.1148/radiol.2020200230

https://pubs.rsna.org/doi/10.1148/radiol.2020200230

CT Imaging Features of 2019 Novel Coronavirus (2019-nCoV)

CT Imaging Features of COVID-19 Patients

Unlike pathogenic laboratory testing which produces a significant amount of false-negatives while still being a time-consuming method, Machine Learning models  can display radio-logical and image patterns via CT chest scans. 3D CT volume can detect the virus's transmission by using a lung-region 3D CT scan which is segmented and fed into a 3D deep neural network to predict the infection rate.

Large-Scale Analysis of CT Scans: 

Coronavirus pneumonia tends to affect the entire lung as opposed to other viral pneumonia affecting small parts. As per respiratory clinicians and radiologists, the 3 features  of coronavirus pneumonia are ground-glass opacities, consolidation, and pleural effusion seen in the chest of the CT scans.

COVID-19 pneumonia vs. other viral pneumonia

CNN is a strong utility that is used for classifying images. Built on layers organized by three dimensions. The neurons of a specific layer do not attach to a full set of neurons to the later layer. The output is a diminished vector of probability scores to produce a model of COVID-19 classification.

Convolutional neural networks

- 103 patients with lung CTs were chosen for the study
- The CT scans from 107 healthy patients were taken (control group)
- Figures show the relationship between the volume of blood vessels and its cross-sectional area, where the area of a circle is measured by taking a slice of the blood vessel
- The volume of blood vessels in healthy patients was lower compared to COVID-19 patients
- However, there is a slight increase in blood vessel volume in healthy patients when the cross section area of the blood vessels is between ~1.25 and 3.5 mm^2
- The increase in blood vessel volume in COVID-19 patients could be due to inflammation in the blood vessels to increase oxygen uptake by the lungs


Learn Before

Related

Learn After