Dataset Information

COVID-19: Histopathological correlates of imaging patterns on chest computed tomography.

ABSTRACT:

Background and objective

Patients with coronavirus disease 2019 (COVID-19) pneumonia present with typical findings on chest computed tomography (CT), but the underlying histopathological patterns are unknown. Through direct regional correlation of imaging findings to histopathological patterns, this study aimed to explain typical COVID-19 CT patterns at tissue level.

Methods

Eight autopsy cases were prospectively selected of patients with PCR-proven COVID-19 pneumonia with varying clinical manifestations and causes of death. All had been subjected to chest CT imaging 24-72 h prior to death. Twenty-seven lung areas with typical COVID-19 patterns and two radiologically unaffected pulmonary areas were correlated to histopathological findings in the same lung regions.

Results

Two dominant radiological patterns were observed: ground-glass opacity (GGO) (n = 11) and consolidation (n = 16). In seven of 11 sampled areas of GGO, diffuse alveolar damage (DAD) was observed. In four areas of GGO, the histological pattern was vascular damage and thrombosis, with (n = 2) or without DAD (n = 2). DAD was also observed in five of 16 samples derived from areas of radiological consolidation. Seven areas of consolidation were based on a combination of DAD, vascular damage and thrombosis. In four areas of consolidation, bronchopneumonia was found. Unexpectedly, in samples from radiologically unaffected lung parenchyma, evidence was found of vascular damage and thrombosis.

Conclusion

In COVID-19, radiological findings of GGO and consolidation are mostly explained by DAD or a combination of DAD and vascular damage plus thrombosis. However, the different typical CT patterns in COVID-19 are not related to specific histopathological patterns. Microvascular damage and thrombosis are even encountered in the radiologically normal lung.

SUBMITTER: Kianzad A

PROVIDER: S-EPMC8447040 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:ObjectiveChest computed tomography (CT) imaging is being increasingly used for potential lung donor assessment. However, the efficacy of CT imaging in this setting remains unknown. We hypothesize that chest CT imaging independently affects the decision-making process in donor lung utilization.MethodsWe conducted a retrospective cohort study of all adult donation after brain death donors managed through our local organ procurement organization from June 2011 to November 2016. An experienced thoracic radiologist independently reviewed donor chest CT and chest x-ray images in a blinded, standardized manner to determine the presence of structural lung disease (eg, emphysema, interstitial lung disease [ILD]) and acute abnormalities (eg, traumatic lung injury [TLI]). Distinct models of lung utilization were fit to groups with initial partial pressure of oxygen (iPaO2) ≤300 mm Hg (suboptimal) and iPaO2 >300 mm Hg (optimal).ResultsThe organ procurement organization managed 753 donors during the study period, with a lung utilization rate ([lung donors/all organ donors] × 100) of 36.5% (275 of 753). Four hundred forty-five (59.1%) donors received chest CT imaging, revealing emphysema (13.7%), ILD (2.5%), and TLI (7.2%). In univariate analysis, findings of TLI (odds ratio [OR], 2.23; 95% confidence interval [CI], 1.08-4.61) were positively associated with lung utilization, whereas findings of emphysema (OR, 0.18; CI, 0.08-0.40) were negatively associated with utilization. In multivariate analysis, CT findings of emphysema (OR, 0.21; CI 0.08-0.54) remained negatively associated with utilization. No potential donors with CT findings of ILD became lung donors. After controlling for chest x-ray findings, chest CT imaging findings of structural lung disease remained negatively associated with utilization (P = .0001). Lung utilization rate in the suboptimal and optimal iPaO2 populations was 35.1% and 41.4%, respectively, and CT findings of emphysema had a significant association with nonutilization in both groups.ConclusionsIn the evaluation of potential lung donors, chest CT imaging findings of structural lung disease, such as emphysema and ILD, have a significant negative association with lung utilization. Our findings suggest that chest CT imaging might be an important adjunct to conventional lung donor assessment criteria.

Project description:ObjectiveThe aim of our study was to assess the effect of a short-term treatment with low-moderate corticosteroid (CS) doses by both a quantitative and qualitative assessment of chest HRCT of COVID-19 pneumonia.MethodsCORTICOVID is a single-center, cross-sectional, retrospective study involving severe/critical COVID-19 patients with mild/moderate ARDS. Lung total severity score was obtained according to Chung and colleagues. Moreover, the relative percentages of lung total severity score by ground glass opacities, consolidations, crazy paving, and linear bands were computed. Chest HRCT scores, P/F ratio, and laboratory parameters were evaluated before (pre-CS) and 7-10 days after (post-CS) methylprednisolone of 0.5-0.8 mg/kg/day.FindingsA total of 34 severe/critical COVID-19 patients were included in the study, of which 17 received Standard of Care (SoC) and 17 CS therapy in add-on. CS treatment disclosed a significant decrease in HRCT total severity score [median = 6 (IQR: 5-7.5) versus 10 (IQR: 9-13) in SoC, p < 0.001], as well in single consolidations [median = 0.33 (IQR: 0-0.92) versus 6.73 (IQR: 2.49-8.03) in SoC, p < 0.001] and crazy paving scores [mean = 0.19 (SD = 0.53) versus 1.79 (SD = 2.71) in SoC, p = 0.010], along with a significant increase in linear bands [mean = 2.56 (SD = 1.65) versus 0.97 (SD = 1.30) in SoC, p = 0.006]. GGO score instead did not significantly differ at the end of treatment between the two groups. Most post-CS GGO, however, derived from previous consolidations and crazy paving [median = 1.5 (0.35-3.81) versus 2 (1.25-3.8) pre-CS; p = 0.579], while pre-CS GGO significantly decreased after methylprednisolone therapy [median = 0.66 (0.05-1.33) versus 1.5 (0.35-3.81) pre-CS; p = 0.004]. CS therapy further determined a significant improvement in P/F levels [median P/F = 310 (IQR: 235.5-370) versus 136 (IQR: 98.5-211.75) in SoC; p < 0.001], and a significant increase in white blood cells, lymphocytes, and neutrophils absolute values.ConclusionThe improvement of all chest HRCT findings further supports the role of CS adjunctive therapy in severe/critical COVID-19 pneumonia.

Project description:PurposeThe novel coronavirus COVID-19, which spread globally in late December 2019, is a global health crisis. Chest computed tomography (CT) has played a pivotal role in providing useful information for clinicians to detect COVID-19. However, segmenting COVID-19-infected regions from chest CT results is challenging. Therefore, it is desirable to develop an efficient tool for automated segmentation of COVID-19 lesions using chest CT. Hence, we aimed to propose 2D deep-learning algorithms to automatically segment COVID-19-infected regions from chest CT slices and evaluate their performance.Material and methodsHerein, 3 known deep learning networks: U-Net, U-Net++, and Res-Unet, were trained from scratch for automated segmenting of COVID-19 lesions using chest CT images. The dataset consists of 20 labelled COVID-19 chest CT volumes. A total of 2112 images were used. The dataset was split into 80% for training and validation and 20% for testing the proposed models. Segmentation performance was assessed using Dice similarity coefficient, average symmetric surface distance (ASSD), mean absolute error (MAE), sensitivity, specificity, and precision.ResultsAll proposed models achieved good performance for COVID-19 lesion segmentation. Compared with Res-Unet, the U-Net and U-Net++ models provided better results, with a mean Dice value of 85.0%. Compared with all models, U-Net gained the highest segmentation performance, with 86.0% sensitivity and 2.22 mm ASSD. The U-Net model obtained 1%, 2%, and 0.66 mm improvement over the Res-Unet model in the Dice, sensitivity, and ASSD, respectively. Compared with Res-Unet, U-Net++ achieved 1%, 2%, 0.1 mm, and 0.23 mm improvement in the Dice, sensitivity, ASSD, and MAE, respectively.ConclusionsOur data indicated that the proposed models achieve an average Dice value greater than 84.0%. Two-dimensional deep learning models were able to accurately segment COVID-19 lesions from chest CT images, assisting the radiologists in faster screening and quantification of the lesion regions for further treatment. Nevertheless, further studies will be required to evaluate the clinical performance and robustness of the proposed models for COVID-19 semantic segmentation.