key: cord-0888109-rutpcax0
authors: Henkel, Maurice; Weikert, Thomas; Marston, Katharina; Schwab, Nathalie; Sommer, Gregor; Haslbauer, Jasmin; Franzeck, Fabian; Anastasopoulos, Constantinos; Stieltjes, Bram; Michel, Anne; Bremerich, Jens; Menter, Thomas; Mertz, Kirsten D.; Tzankov, Alexandar; Sauter, Alexander W.
title: Lethal COVID-19: Radiological-Pathological Correlation of the Lungs
date: 2020-11-19
journal: Radiol Cardiothorac Imaging
DOI: 10.1148/ryct.2020200406
sha: 3f441fa4483dd9e288ce3bddc985aaeec55736d0
doc_id: 888109
cord_uid: rutpcax0

BACKGROUND: CT has emerged as an important diagnostic tool in COVID-19, but the underlying pathological changes behind CT findings are not yet fully elucidated. PURPOSE: The purpose of this retrospective study was to correlate CT patterns of fatal cases of COVID-19 with post-mortem pathology observations. MATERIAL AND METHODS: The study included 70 lung lobes of 14 patients who died from RT-PCR confirmed COVID-19. All patients underwent ante-mortem CT and autopsy between March 9 and April 30, 2020. Board-certified radiologists and pathologists performed lobe-wise correlations of pulmonary observations. In a consensus reading, 267 radiological and 257 histopathological observations of the lungs were recorded and systematically graded according to severity. These observations were matched and evaluated. RESULTS: Predominant CT observations were ground glass opacities (GGO; 59 of 70 lobes examined) and areas of consolidation (33/70). The histopathological observations were consistent with diffuse alveolar damage (70/70), capillary dilatation and congestion (70/70), often accompanied by microthrombi (27/70), superimposed acute bronchopneumonia (17/70) and leukocytoclastic vasculitis (7/70). Four patients had pulmonary emboli. Bronchial wall thickening on CT histologically corresponded with acute bronchopneumonia. GGOs and consolidations corresponded with mixed histopathological observations including capillary dilation and congestion, interstitial edema, diffuse alveolar damage and microthrombosis. Vascular alterations were prominent observations in both CT and histopathology. CONCLUSION: A significant proportion of GGO correlated with the pathologic processes of diffuse alveolar damage, capillary dilatation and congestion and microthrombosis. Our results confirm the presence and underline the importance of vascular alterations as a key pathophysiological driver in lethal COVID-19.

Coronavirus disease 2019 (COVID-19) caused a pandemic with more than 34'000'000 cases worldwide (1) . It is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) and has an estimated mortality rate of up to 5-10% (2) . Given the high reproduction number of the virus, diagnostic tools for rapid diagnosis and evaluation are important to track and mitigate transmission. Computed tomography (CT) of the chest is sensitive for COVID-19 in high prevalence areas and may have a role in identifying other causes of respiratory failure (3) . Frequently found observations include ground-glass opacities (GGO), consolidations, crazy-paving pattern, reticulations, thickened interlobular septa and air bronchogram. A better characterization of the main pathological drivers for each radiological pattern, which is currently lacking (4, 5) , could provide a strong, rational foundation for treatment strategies. The first post-mortem examination of a COVID-19 patient (6) showed pulmonary edema and hyaline membrane formation in both lungs, which is the assumed histopathological correlate of GGO (7) . Based on knowledge previously acquired from SARS cases, diffuse alveolar damage is suspected to represent the primary histological response that accompany acute lung injury (8) . Furthermore, there is new evidence that associate thromboembolic complications in COVID-19 with coagulation activation, endothelial dysfunction, capillary congestion, and acute (exudative) diffuse alveolar damage (DAD) (9) (10) (11) . Microthrombosis of lung capillary networks and thromboembolism are emerging as key pathophysiological drivers behind hypoxemia leading to mechanical ventilation (12, 13) . Radiological imaging might reflect these vascular abnormalities by the vascular thickening sign (14) and perfusion abnormalities (15) .

The purpose of this retrospective study is to gain knowledge on the pathologies underlying CT patterns in COVID-19 by means of radiological-pathological correlation analysis of a local COVID-19 autopsy series of patients who underwent ante-mortem chest CT.

This retrospective, dual-center radiology and pathology study was approved by the local ethics committee, written informed consent was obtained from all subjects (ID *blinded for review*)

We retrospectively identified all lethal COVID-19 cases confirmed by reverse transcription polymerase chain reaction (RT-PCR) with both an antemortem CT and autopsy performed between March 9 and April 30, 2020 (n=14). In cases with multiple CTs, the last ante-mortem scan was selected for analysis. Complete autopsy was performed in 13 and partial autopsy of the upper respiratory tract, lungs and heart in one case. Autopsies were executed at (*blinded for review*) (n=10) and (*blinded for review*) (n=4 

Our institute developed a COVID-19 optimized autopsy protocol in line with recently published recommendations (18) . Thoracic organs, lungs, trachea and larynx were completely exenterated and perfused via the trachea with 4% refrigerated (+4°C) phosphate-buffered formalin (pH 7.4). The trachea was then closed with a clamp and specimens were left in formalin at room temperature for 72 hours before dissection.

The lungs were subsequently cut into 0.5-1 cm parasagittal slices. At least two sections of the first 3 cm of subpleural parenchymal of each lobe as well as the trachea I n p r e s s were histologically analyzed. The observations were graded in a four-step scale according to severity (none, mild, moderate, severe; Table 1 ).

Tissue samples were processed using standard laboratory equipment and stained with standard histochemical methods (haematoxylin and eosin staining). The histopathological specimens were reviewed in consensus by two pathologists with 13 (*blinded for review*) and 20 (*blinded for review*) years of experience. The observations were graded in a four-step scale according to severity (none, mild, moderate, severe; Table 2 ). 

To ensure the highest possible granularity despite retrospective study design and temporal and local variability of inflammatory processes, the rad-path correlation was performed at a lobe-wise level instead of a patient-wise level. The observations of each lobe were recorded separately and graded according to severity (either binary, positive/negative, or a four-step scale, using the terms none, mild, moderate, and severe) in close coordination between radiologists and pathologists. The correlation analysis examined whether radiological observations were associated with pathological patterns. In cases that displayed multiple radiological features within a lobe, features were independently compared between the corresponding gross specimen and the histopathological findings.

Continuous variables were analysed by means and ranges (demographics), means and interquartile ranges (IQR; laboratory values) and frequencies (number of lobes with a given feature). To test for differences in observations between patients with a short and long time interval between onset of symptoms and CT (imaging features), and between onset of symptoms and autopsy (histopathology features), patients were assigned to short interval and long interval cohorts according to the respective median.

Patients having a shorter time interval than the median were assigned to the short interval group and patients having a longer time interval than the median were assigned to the long interval group. Chi-squared tests were conducted to assess differences between long / short interval groups regarding age (T-test) and sex (Chisquared test). Analysis was conducted in R (19) and figures were produced using ggplot2 (20) . P-values ≤ 0.05 were considered to be statistically significant.

Clinical features including comorbidities and symptoms are listed in Table 3 . The mean interval from death to autopsy was 38.5 hours (range: 11.0-97.0 hours) and the mean interval between chest CT and death was 3.7 days (range: 0.0-17.0 days). The average hospitalization time before death was 5.5 days. The mean age in our collective was 76 years (range: 58-96 years); 29% of patients were female. There were no statistically significant differences in age and sex between patients with a short and long time interval between onset of symptoms and CT (age: p=0.19; sex: p=0.48) and death (age: p=0.27; sex: p=0.73), respectively.

The most prevalent clinical symptoms were cough (n=10), followed by fever (n=7) and dyspnea (n=3). All patients suffered from at least three comorbidities (three comorbidities: n=7, four comorbidities: n=6; more than four comorbidities: n=1; Table   3 ). Patients in our cohort were treated with hydroxychloroquine (n=11), iopinavir/ritonavir (n=7), tocilizumab (n=5), antibiotics (n=4), remdesivir (n=1), or did not receive COVID-19 specific mediation (n=3). 

The overall gross findings of all lobes in all patients are shown in Table 4 

While most recent studies also described the radiological pattern of organization of pneumonia in patients with COVID-19 (4, 21) , no histopathological features for fibrosis were found in our study. Interestingly, in a recent study using post-mortem transbronchial lung cryobiopsy from six patients with a median illness duration of 32 days, three showed late/fibrotic phase diffuse alveolar damage, one of them with honeycombing (22) . In another study, for five patients, who died around 20 days after the beginning of symptoms, the histologic pattern was an acute fibrinous and organizing pneumonia (AFOP) (23) . The various radiological patterns showed only minor differences in the frequency of underlying histopathological changes. The most important components for GGO were capillary dilatation and congestion (26.9%), interstitial edema (25.0%) and acute (exudative) DAD (21.7%). Consolidation showed similar histopathological patterns with slightly more microthrombosis (12.6% vs. 10 .4%) and leukocytoclastic vasculitis (3.6% vs. 2.4%). Importantly, the observed CT patterns were not linked to a specific histopathological finding.

Furthermore, an increase in the frequencies of bronchial wall thickening, pulmonary arterial enlargement and pulmonary embolism in patients with longer symptom onset to CT was found. The increase in bronchial wall thickening and consolidation might be explained by superimposed bacterial superinfections and subsequent acute bronchopneumonia, which was evident in our collective. Figure 5a provides an example. Of note, while there was no evident temporal dynamics of microthrombosis, more pulmonary emboli were detected in patients with a longer interval between onset of symptoms and CTPA in our cohort. Surprisingly, the temporal dynamic of radiographic changes could not be clearly correlated with corresponding histopathological results. Further evaluation of the histopathological observations as I n p r e s s a function of disease duration showed an increase of acute (exudative) and organizing (proliferative) phase DAD, whereas no fibrotic changes were observed.

The observation of enlarged pulmonary arteries in our series might be related to an increase of parenchymal and predominantly intravascular pressure (24) , due to the severe COVID-19 pulmonary microangiopathy affecting the alveolar capillary network (25) . In the context of recent reports of pathologically altered coagulation, these results may also indicate an increase in vascular incidents (10, 26) . 30 of 388 patients in the study by Lodigiani et al. underwent CTPA confirming pulmonary embolism in 10 cases (9) . Although higher, this positive rate of 33% seems in line with the observed rate in our collective of 20%; importantly, all our patients received anticoagulation. While most of the currently available literature relies on non-contrast CT (27) , the need to assess vascular abnormalities is being recognized as an increasingly important factor (28, 29) , both to distinguish COVID-19 pneumonia from other viral infections, and to exclude pulmonary embolism. Interestingly, the high incidence of microthrombosis and the low number of pulmonary emboli detected with CTPA suggests a possible underestimation of the vascular alterations associated with COVID-19 using imaging, especially in un-enhanced scans. Newly described signs such as "vascular thickening", "pulmonary arterial enlargement", or "vascular congestion" could reflect these alterations. Severe influenza pneumonia has previously been described to cause a hyper-inflammatory response with virally associated platelet activation, which can lead to pulmonary thrombosis with passive congestion. Similar to influenza pneumopathy, COVID-19 is able to limit compensatory ventilation responses by means of vascular leakage and alveolar edema, thus contributing to widespread haemorrhage (10, 30) . According to a comparative autopsy study, alveolar capillary I n p r e s s microthrombi were 9 times as prevalent in patients with COVID-19 as in patients with influenza and extent of angiogenesis was 2.7 times higher than in influenza (25) . Therefore, described vascular alterations seem to be specific for COVID-19.

Our study has several limitations. First, its retrospective design and the relatively small number of cases. For this reason, no complex statistical analyses have been performed. Second, the time intervals between symptoms, CT and autopsy varied. This is a factor that cannot be controlled in an observational study. Third, as all patients died from COVID-19, there is a bias towards more severe and rapidly progressive courses of disease. Fourth, the correlation analysis was performed on a per-lobe level.

A finer granularity was not possible due to the specific COVID-19 autopsy protocol.

For future studies, image-guided tissue sampling is desirable to further increase the resolution of analysis. However, as stated in the results, multiple pathologic processes were found also in lobes with pure GGO / consolidation patterns in CT. Finally, CT parameters and vendors differed due to the dual-center nature of this study. However, all scans were acquired with a high diagnostic quality clearly suitable for this correlation analysis as shown in supplementary materials.

The results of this study deepen our understanding of COVID-19 pathophysiology confirming the importance of vascular alterations. Our observations imply that both severe acute lung injury and vascular complications contribute to fatal outcomes. 

Corona virus resource center

Case-Fatality Risk Estimates for COVID-19 Calculated by Using a Lag Time for Fatality. Emerging infectious diseases

ACR recommendations for the use of chest radiography and computed tomography (CT) for suspected COVID-19 infection. ACR website Advocacy-and Economics/ACR-Position-Statements/Recommendations-for-Chest-Radiography-and-CTfor-Suspected-COVID19-Infection Updated

The many faces of COVID-19: spectrum of imaging manifestations

Essentials for radiologists on COVID-19: an update-radiology scientific expert panel

Pathological findings of COVID-19 associated with acute respiratory distress syndrome. The Lancet Respiratory Medicine

Chest CT manifestations of new coronavirus disease 2019 (COVID-19): a pictorial review

Chest CT Findings in Patients With Coronavirus Disease 2019 and Its Relationship With Clinical Features

Venous and arterial thromboembolic complications in COVID-19 patients admitted to an academic hospital in

Autopsy Findings and Venous Thromboembolism in Patients With COVID

Epub 2020/05/07

Postmortem examination of COVID-19 patients reveals diffuse alveolar damage with severe capillary congestion and variegated findings in lungs and other organs suggesting vascular dysfunction

Prevalence of venous thromboembolism in patients with severe novel coronavirus pneumonia

Severe COVID-19 infection associated with endothelial activation

Vascular Abnormalities as Part of Chest CT Findings in COVID-19

Hypoxaemia related to COVID-19: vascular and perfusion abnormalities on dual-energy CT

SARS-CoV-2 From the Trenches: A Perspective From New York City

Fleischner Society: glossary of terms for thoracic imaging

Autopsy in suspected COVID-19 cases

R: A language and environment for statistical computing

ggplot2: elegant graphics for data analysis

Longitudinal CT Findings in COVID-19 Pneumonia: Case Presenting Organizing Pneumonia Pattern

Fibrotic progression and radiologic correlation in matched lung samples from COVID-19 post-mortems

COVID-19 pneumonia: different respiratory treatments for different phenotypes?

Significance of main pulmonary artery dilation on imaging studies

Confirmation of the high cumulative incidence of thrombotic complications in critically ill ICU patients with COVID-19: An updated analysis

The role of chest imaging in patient management during the COVID-19 pandemic: a multinational consensus statement from the Fleischner Society

Acute Pulmonary Embolism Associated with COVID-19 Pneumonia Detected by Pulmonary CT Angiography

Acute Pulmonary Embolism in COVID-19 Patients on CT Angiography and Relationship to D-Dimer Levels

Aberrant coagulation causes a hyper-inflammatory response in severe influenza pneumonia

We thank all patients and their families for their willingness to dedicate the bodies to (*blinded for review*) for autopsy. (*blinded for review*) received funding from (*blinded for review*). We thank Rita Achermann (statistician) for statistical advice. I n p r e s s

p r e s s I n p r e s s 5a Radiological-pathological correlation in patient 1, a 67-year-old female. Gross pathology findings revealed interlobular septal edema, congestion, thickened bronchial walls (arrow) and consolidation (arrowhead). The corresponding CT in sagittal plane discloses bronchial wall thickening (arrow) and dystelectasis in the posterior parts of the lower lobes (arrowhead). Microscopy disclosed bronchopneumonia in the left lower lobe (upper right), thickened bronchial walls (lower left), and microthrombosis (lower right, immunohistochemistry for fibrin¸ polyclonal antibody (Dako, Glostrup, Denmark, A0080)).5b Radiological-pathological correlation in patient 2, a 66-year-old man. Gross findings document interlobular septal edema and segmental hemorrhage in the anterobasal segment of the left lower lobe (arrow), while CT shows peripherally pronounced GGO in the left lower lobe (arrow). Inserted miniP reveals some focal bronchial dilatation in association to ground glass opacities. Microscopy reveals hyaline membranes as remnants of acute exudative (arrowhead) and intraalveolar fibroblastic proliferations as signs of proliferative DAD (arrow). Furthermore, capillaries show extensive congestion (asterisk). 6c Radiological-pathological correlation in patient 9, a 58-year-old male. Gross findings show congestion, interlobular septal edema and multiple thromboembolisms (insert at lower left). CT reveals subtotal consolidation of the left lower lobe (arrow) and bilateral pulmonary embolisms (not shown). Microscopy shows hyaline membranes as correlates of acute exudative (arrow) and fibroblastic proliferations as signs of proliferative (arrowhead) DAD as well as alveolar hemorrhage (far right).

Anonymized, zoomable high-resolution scans of the histology sections of each patient as well as the corresponding transverse chest CT series in lung kernel reconstruction can be found following the links provided below.