A Comparative Study of Ventilation Perfusion Mismatch Derived from Non-Contrast CT Scans in Hospitalized COVID-19 Patients
American Journal of Respiratory and Critical Care Medicine
Rationale: Severe coronavirus disease 2019 (COVID-19) is associated with important variations within the immune system and the coagulation cascade. We have developed a robust method for computing ventilation (CT-V) and perfusion (CT-P) from dynamic non-contrast CT scan, which can detect ventilation-perfusion (VQ) mismatch at a voxel level. We hypothesize that COVID-19 patients with mild disease will still have a higher VQ mismatch compared to patients with no respiratory symptoms. Methods: We included 12 random patients with mild symptoms from a prospective study characterizing quantitative lung function in patients with COVID-19 (NCT04320511) and compared their VQ scores to 12 patients with no respiratory symptoms in the NORM dataset (NCT00848406) matched to age, gender and BMI. The CT-P and CT-V methods apply image processing and physical modeling to an Inhale/Exhale CT image pair to generate a quantitative CT-P and CT-V images. We calculated VQ mismatch as the percent of lung voxels with residual fit errors 4 standard deviations apart from a least median of squares quadratic regression model describing CT-P as a function of CT-V.Results: All included COVID cases were hospitalized to regular floors and were breathing at room air, except for 3 patients on supplemental oxygen < 3L/min. The mean CT-V scores were significantly lower in COVID-19 cases compared to controls (0.46 vs. 1.39, 95% CI difference 0.51, 1.33, p 0.001). Likewise, the mean CT-P scores were significantly lower in COVID-19 cases vs. controls (0.14 vs. 0.18, 95% CI difference 0.02, 0.08, p 0.004). However, the median VQ mismatch scores were significantly higher in COVID-19 cases [0.300 (IQR 0.287,0.320) vs. 0.270 (IQR 0.233,0.293), p 0.04], see figure. Conclusion: Patients with COVID-19 have significant derangements in pulmonary physiology, VQ mismatch despite having minimal to no-oxygen requirements. Progression of VQ mismatch from an early stage could be studied to identify patients at risk for mechanical ventilation and mortality. Figure: (Left) Box plots of the VQ scores for each COVID cases vs. controls. (Right) Representative CT-V and CT-P in a patient with COVID-19 pneumonia with corresponding CT scan. Higher function areas appear red and low function areas appear bluer. Top row depicts “Dead space ventilation” with normal appearing CT scan but diminished areas of perfusion in the right lung and preserved ventilation. Bottom row demonstrates area of pneumonia in the left upper lung zone with “Shunt physiology”. CT-P shows increased activity with reddish hue, whereas corresponding ventilation is diminished in the same area.
Nair GB, Al-Katib S, Turner-Lawrence D, Ionescu F, Galban CJ, Van den Berge M, et al. A Comparative Study of Ventilation Perfusion Mismatch Derived from Non-Contrast CT Scans in Hospitalized COVID-19 Patients. In American Thoracic Society; 2021. p. A4467. (American Thoracic Society International Conference Abstracts). Available from: https://www.atsjournals.org/doi/pdf/10.1164/ajrccm-conference.2021.203.1_MeetingAbstracts.A4467