Cardiovascular findings on chest computed tomography associated with COVID-19 adverse clinical outcomes

Figures

• 

  Fig. 1. Choice of chest CT per patient. Pathway of Chest CT selection is shown above. If multiple CT scans were performed, preference was given to contrast-enhanced CTs that were performed closest to the patient's time of death or escalation to the highest level of oxygen therapy. CT: computed tomography. O2: oxygen.

• 

  Fig. 2. Cardiovascular findings measured on Chest CT. PA diameter was measured just proximal to the bifurcation of the pulmonary trunk ascending aortic length and width were measured on the same axial CT slice at the level of the right pulmonary artery (red lines) (A). Presence or absence of coronary calcium and aortic calcification was measured in binary fashion (red arrow) (B). Presence or absence of IVC reflux was measured (red arrow) (C). RV and LV diameters were measured on a single axial CT slice as described in text (red lines) (D). Interventricular septum was evaluated in binary fashion for normal (D) versus abnormal (C) position. CT: computed tomography, IVC: inferior vena cava, LV: left ventricle, PA: pulmonary artery, RV: right ventricle.

• 

  Fig. 3. Pulmonary findings on chest CT. The pulmonary findings from the clinically recorded report included the typical COVID-19 findings of ground glass opacities, interlobular and intralobular septal thickening referred to as “crazy-paving” and peripheral distribution of consolidations, all of which are demonstrated above.

• 

  Fig. 4. Cardiac versus pulmonary multivariable models. As plotted on a receiver operation curve, the cardiac and pulmonary multivariable models were poorly predictive (AUC 0.62 [CI 0.54–0.70] and 0.61 [0.54–0.68], respectively), with no significant difference between the predictive ability between the two models, p = 0.95 AUC: area under the curve.