International Journal on Advanced Science, Engineering and Information Technology, Vol. 12 (2022) No. 3, pages: 1215-1221, DOI:10.18517/ijaseit.12.3.12579

CT-based Analysis of Vascular Tree Abnormalities in Different Phenotypes of COVID-19 Pneumonia

Gihad Ibrahim, Ahmed Hassan

Abstract

In this paper, Computed Tomography (CT) images of six confirmed COVID-19 patients were analyzed in order to investigate the physiological abnormalities in the vascular tree in response to the disease-induced hypoxia. The CT images were classified into an L-type and H-type groups based on the cumulative voxel distribution of the CT scan. A 3-Dimentional model of the vascular tree was reconstructed out of each CT image following a computational framework. Then, the Cross-Sectional Area (CSA) of the vessels belonging to each vascular tree was computed. The acquired results were compared against averaged measurements of three healthy subjects that were computed following the same approach. The results showed that as the severity of COVID-19 lean towards H-type phenotype, signs of vasoconstriction in small blood vessels with a CSA less than 10 mm2 tend to decreases, whereas signs of vasodilation in medium to large blood vessels increases. The intensity of dilated blood vessels proximal to consolidated areas of the lung was found to increase significantly as the disease progresses in the lungs. Furthermore, signs of vasoconstriction and vasodilation in the vascular tree were observed in all lobes of the lung in both phenotypes including seemingly healthy lobes. The results in this paper are suggestive of intrapulmonary blood flow shunting towards unaerated areas of the lungs which may lead to a ventilation/perfusion mismatch even at minor cases of COVID-19. The results also suggest that subject-specific regulated use of vasodilating medication may reduce the number of cases that require mechanical ventilation.

Keywords:

COVID-19 pneumonia; SARS-CoV-2; COVID-19 phenotypes; vascular tree analysis.

Viewed: 1091 times (since abstract online)

cite this paper     download