key: cord-0993013-dpxfcryw authors: Sadiq, Zuhair; Rana, Shehroz; Mahfoud, Ziyad; Raoof, Ameed title: Systematic review and meta-analysis of chest radiograph (CXR) findings in COVID-19 date: 2021-07-27 journal: Clin Imaging DOI: 10.1016/j.clinimag.2021.06.039 sha: 20c7bcf7bcc082e4372e09d24ec99e858e3bb34a doc_id: 993013 cord_uid: dpxfcryw Chest radiography (CXR) is most likely to be the utilized modality for diagnosing COVID-19 and following up on any lung-associated abnormalities. This review provides a meta-analysis of the current literature on CXR imaging findings with to determine the most common appearances of lung abnormalities in COVID-19 patients to equip medical researchers and healthcare professionals in their efforts to combat this pandemic. Twelve studies met the inclusion criteria and were analyzed. The inclusion criteria consisted of: (1) published in English literature; (2) original research study; (3) sample size of at least 5 patients; (4) reporting clinical characteristics of COVID-19 patients as well as CXR imaging features; and (5) noting the number of patients with each corresponding imaging feature. A total of 1948 patients were included in this study. To perform the meta-analysis, a random-effects model calculated the pooled prevalence and 95% confidence intervals of abnormal CXR imaging findings. Seventy-four percent (74%) (95% CI: 51–92%) of patients with COVID-19 had an abnormal CXR at the initial time of diagnosis or sometime during the disease course. While there was no single feature on CXR that was diagnostic of COVID-19 viral pneumonia, a characteristic set of findings were obvious. The most common abnormalities were consolidation (28%, 95% CI: 8–54%) and ground-glass opacities (29%, 95% CI: 10–53%). The distribution was most frequently bilateral (43%, 95% CI: 27–60%), peripheral (51%, 95% CI: 36–66%), and basal zone (56%, 95% CI: 37–74%) predominant. Contrary to parenchymal abnormalities, pneumothorax (1%, 95% CI: 0–3%) and pleural effusions (6%, 95% CI: 1–16%) were rare. The first case of coronavirus disease 2019 (COVID- 19) was first reported in Wuhan, China in December 2019 and it was declared a pandemic by the World Health Organization (WHO) in March 2020 (1) . Healthcare workers need diagnostic tools to study cases of potential COVID-19 that are both sensitive and specific. The gold standard for diagnosing COVID-19 patients is a viral nucleic acid test conducted by reverse transcription-polymerase chain reaction (RT-PCR) with a sensitivity of 79% and specificity of 100% (2) . Chest CT has been found to have comparable diagnostic performance with a sensitivity of 77% and specificity of 96% (3). The pulmonary system is primarily affected by COVID-19, as such, it is common practice to request a chest radiograph (CXR) in suspicious cases as the initial imaging exam. The diagnostic An aggregate of 401 non-duplicate citations was identified by using the aforementioned search strategies from which 37 potentially applicable papers were evaluated. Ultimately, 12 studies were eligible with reported cases ranging from 5 to 636 were included in the meta-analysis (Table 1, 2 & 3) . In total, these studies contained 1,948 patients, 643 cases with normal CXR and 1,305 cases with abnormal CXR findings ( Figure 1 ). Table 1 contains the basic demographic of the participant population in each study as well as the reported CXR findings. In the following sections, we present an analysis and review of CXR findings in COVID-19 patients as gathered from the eligible studies. The ensuing info was collected from each published paper for the analysis of CXR findings: total number of patients in the study; gender of participants and average age; lung lesion distribution (diffuse, peripheral or perihilar); lung involvement (right, left, or bilateral); typical abnormalities such as consolidation and groundglass opacities; frequency of zone predominance (superior or basal); and less common findings such as pleural effusion, pulmonary edema, atelectasis and pneumothorax. Stata version 15.0 was utilized to perform the statistical analysis. Table 1 and Table 2 summarize the collected data. The figures displaying the CXR findings were obtained with permission via email from Radiopedia.org. A meta-analysis was conducted using a variance components model to determine the pooled proportion values and the analogous 95% confidence intervals (CI) of distinctive imaging features linked to patients with COVID-19. The pooled-effects meta-analysis model involved an assumption that the effect size (ES) being approximated in the different studies are not alike; rather, they adhere to some distribution ( Figure 1 ). The center of the symmetric distribution characterizes the average of the effects, whilst the width describes the degree of heterogeneity (10) . Statistical heterogeneity was calculated using the I² test (p<0.05 represented significant statistical heterogeneity). Data was entered into Stata version 15.0 to perform statistical analyses. The pooled-effects model was calculated for each CXR finding. All the graphs for the metaanalyses and forest plots can be found in the supplementary document. Statistically significant heterogeneity was found amongst all the studies in regards to the data extracted from their findings (p<0.05). The results from the graphs were summarized in Table 3 . J o u r n a l P r e -p r o o f Results CXR was utilized in 12 studies diagnosing COVID-19. Both CXR and CT were used in 4 studies, with chest CT proving to be superior to CXR in the assessment and detection of each type of lung abnormality, demonstrating the limitations of CXR (11) (12) (13) (14) . There is some heterogeneity between studies in regards to the average age of the patient cohort. In addition, the diverse array of findings on CXR sometimes made it hard to make sense of because of vague and nonstandard terminology such as "patchy opacities," "nodules," "infiltrates," "airspace disease," and "airspace opacities", which may contribute to the limitations in CXR analysis. Of the 12 studies, 9 (75.0%) found normal CXR findings in COVID-19 patients. The proportion of patients with normal CXR findings varied from 5% at the low end of the spectrum to 58% at the higher end; this range was found to be dependent on how COVID-19 presented in patients, with high normal CXR percentages being reported in asymptomatic patients or those with mild symptoms. The overall pooled prevalence was 26% (95% CI: 8-49%). Furthermore, ≥25% of normal CXR findings were reported in 70.0% (7/10) studies. A total of 11 studies (91.7%) noted information in regards to bilateral, right, or left lung involvement (Table 1) ; from which 9 studies noted a greater proportion of bilateral lung involvement (43%, 95% CI: 27-60%) when compared to right lung involvement (19%, 95% CI: 7-33%) or left lung involvement (1%, 95% CI: 0-4%) ( Tables 3 and 4 ). All 11 studies noted that bilateral lung involvement was more commonly seen in urgent or severe cases of COVID-19 and that bilateral lung involvement was much greater than the involvement of the right or left lung in COVID-19 patients. Six studies (50.0%) noted the proportions on the distribution of lesions or abnormalities in the perihilar or peripheral lung regions or diffuse lung distribution on CXR, with peripheral distribution (51%, 95% CI: 36-66%) showing significant dominance compared to perihilar J o u r n a l P r e -p r o o f distribution (13%, 95% CI: 4-24%) and diffuse distribution (13%, 95% CI: 1-35%). Eight studies (66.7%) reported the details of zone predominance of abnormalities or lesions in the basal or superior or neither lung regions, with basal predominance (56%, 95% CI: 37-74%) being a more common finding compared to superior predominance (8%, 95% CI: 2-16%) or no predominance (17%, 95% CI: 0-46%). Ippolito et al. noted information only on the involvement of peripheral lung distribution and basal predominance (15) . Consolidation and ground-glass opacities (GGOs) were reported in 83.3% (10/12) of the studies; they were the most frequent, atypical abnormalities ( Table 1 ). The pooled prevalence of both consolidation (28%, 95% CI: 8-54%) and GGO (29%, 95% CI: 10-53%) were similar ( Table 4) . The other two studies did not report any data on these features (16, 17) . Besides for consolidation and GGOs, other lung abnormalities were reported in 75.0% (8/12) of these studies, including less common findings like pleural effusion (6%, 95% CI: 1-16%; Figure 5a ) and pneumothorax (1%, 95% CI: 0-3%; Figure 5b ). One of the most common abnormalities seen on CXR in COVID-19 patients is consolidation. Consolidation refers to an occupation of the air space that causes filling of the alveolar spaces by pathological products such as water, pus and blood in the lungs. It appears as a homogeneous increase in pulmonary parenchymal attenuation (increased density) that obscures the margins of the vessels and the walls of the airways (18) . It may present with an air bronchogram, referring to the visualization of the air bronchial lumens within a parenchymal opacity of the lung and, therefore, implies airway patency (18) . On CXR, it appears to have indistinct margins and presents as fluffy opacities that become confluent over time, so the air that normally surrounds J o u r n a l P r e -p r o o f the bronchi becomes opacified and appears white, while the bronchi remain air filled and appear like black tubular structures within the area of consolidation. In Figure 2a , linear black tubular structures that represent air bronchograms are visible in the left lower lobe and this is one of the key features of consolidation. Ground-glass opacities (GGOs) describes the lung parenchymal opacification that produces a minor increase in attenuation for the consolidation ( Figure 3 ). As such, despite the increase in density, the walls of the bronchi and the pulmonary vessels remain differentiated from the affected parenchyma (18) . GGOs represent a partial occupation of the airspace; they are less opaque than consolidations and, as an important consequence, the CT is more sensitive in its detection than CXR. GGOs are common in earlier presentations and can potentially precede the appearance of consolidation (19) . The term "air opacities" is often used in conjunction with "nodules" to denote the filling of air spaces with the products of disease (18) . Compared to community-acquired bacterial pneumoniae, which typically involve one lobe and tend to be unilateral, COVID-19 similar to other viral pneumoniae usually presents with air opacities greater than a single lobe (20) . Recognizing multifocal air opacities on CXR is a pertinent imaging feature of COVID-19 pneumonia. Researchers have noted that the air opacities typically tend to have a bilateral, peripheral and predominantly basal distribution (4) (Figure 4 ). Reticular opacities with regions of ground-glass attenuation are usually better visualized on CXR than chest CT (19) . Pleural effusion is considered to be characteristically rare in COVID-19 and its presence may J o u r n a l P r e -p r o o f indicate co-existing bacterial pneumonia (21) . Pleural effusion presents as blunting of the costophrenic angle on CXR and often appears as a white area at the lung base. In one study that compared COVID-19 to other viral pneumoniae, pleural effusion was more common in non-COVID-19 viral pneumoniae ( Figure 5 ) (22) . It may also be a poor prognostic indicator in COVID-19 patients (23) . Pneumothorax is also a rare finding in COVID-19 patients and presents as a thin, sharp white line at the visceral pleural edge ( Figure 5 ). The authors of one case report of a COVID-19 patient who had a left-sided tension pneumothorax could not be certain whether this pneumothorax was secondary to COVID-19 or the complication was purely coincidental (24) . Atelectasis, cardiomegaly, and pneumomediastinum have also been found in COVID-19 patients; yet, it to be determined if these are findings are unusual manifestations of COVID-19, a sign of dual pathology, or purely incidental (25) . Chest radiographs may be found to be "normal" in early or mild disease, however, COVID-19 patients can later develop radiological or clinical signs of viral pneumonia. Seventy-four percent (74%) (95% CI: 51-92%) of patients with COVID-19 had an abnormal CXR at the initial time of diagnosis or sometime during the disease course. Although the number of participants covered by individual COVID-19 CXR studies is low when compared to similarly designed CT studies, a characteristic set of findings are clear. There is no single feature on CXR that is diagnostic of COVID-19 viral pneumonia. Our findings are consistent with meta-analyses of pertinent COVID-19 features on chest CT (18, 21) , which suggest that the disease presents as atypical or organizing pneumonia on imaging. The most common abnormalities are consolidation (28%, 95% CI: 8-54%) and ground-glass opacities (29%, 95% CI: 10-53%). The distribution is most J o u r n a l P r e -p r o o f frequently bilateral (43%, 95% CI: 27-60%), peripheral (51%, 95% CI: 36-66%), and basal zone (56%, 95% CI: 37-74%) predominant. In contrast to parenchymal abnormalities, pneumothorax (1%, 95% CI: 0-3%) and pleural effusions (6%, 95% CI: 1-16%) are rare. Our meta-analysis has several strengths. First, the studies, which met the inclusion criteria, occurred in different countries and health centers, providing more generalizable results. Second, the total number of cases was somewhat large for 6 months of early publications, generating an accumulation of evidence to evaluate the diagnosis of COVID-19 by CXR. Third, a variety of different imaging features was gathered, including the distribution patterns in the lungs as well as specific imaging features. However, there are also several limitations. First, the reported imaging findings may not be consistent across studies because of different interpreting radiologists. Second, most of the included studies did not differentiate between COVID-19 patients presenting as clinically mild, moderate, or severe cases. Third, some COVID-19 patients may have had chronic diseases and comorbidities like hypertension and diabetes that could have affected CXR findings. CXR imaging can play a role in triaging patients due to long PCR turnaround times. Moreover, CXR has several advantages over CT such as cost and practical considerations, which limits CT's utility. The American College of Radiology has noted that CT decontamination, which is required after scanning COVID-19 patients, can disrupt radiology service availability and has suggested that portable CXR be used to reduce the chances of cross-infection (5) . Common CT findings of lower zone predominance, peripheral distribution and bilateral involvement can also be appreciated on CXR (6) . As such, CXR is most likely to be the utilized modality for identification of COVID-19 and follow-up of any associated lung abnormalities. Possible biases may be due to the inclusion criteria only using studies published in English, J o u r n a l P r e -p r o o f which may exclude data that show different CXR finding trends and patient demographics, which may skew the results of this study. In addition, the CXR pathology findings that were analyzed in this study were chosen based on what we considered to be most pertinent based on the various studies that were included, which may leave out possibly common CXR findings. Another possible bias could be the patient demographics, as the paper studied reported and mostly inpatient cases of COVID-19, this possibly narrows the data to more severe cases of COVD-19. Our results confirm that CXR will continue to be a useful tool in the evaluation and management of patients diagnosed with COVID-19. Although less sensitive than chest CT, CXR can prove useful in its prognostic predictions such as triaging patients and answering questions in regards to whether a patient should stay home, if they will need the ICU in a few days, if they will respond to a specific treatment such as intubation or some drug, whether they should be taken off a ventilator and their overall chances for survival. With an ever-growing number of suspected cases in this pandemic, CXR should be seen as a quick and easy-to-use modality to assess lung abnormalities. depicting the the 95% CIs, and each box in the center describes the proportion for that study comparison. The effect size of each box to the hollow blue diamond represents the strength or (b) Bilateral pneumothorax and pneumomediastinum with diffuse air space in both lungs [33] ) ) ) ) ) ) COVID-19) events as they happen Diagnostic performance between CT and initial real-time RT-PCR for clinically suspected 2019 coronavirus disease (COVID-19) patients outside Wuhan, China An update on COVID-19 for the radiologist -A British society of Thoracic Imaging statement Frequency and Distribution of Chest Radiographic Findings in COVID-19 Positive Patients ACR-Position-Statements/Recommendations-for-Chest-Radiography-and-CT-for-Suspected-COVID19-Infection A systematic review of chest imaging findings in COVID-19 Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72 314 Cases From the Chinese Center for Disease Control and Prevention Essentials for Radiologists on COVID-19: An Update-Radiology Scientific Expert Panel Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement A re-evaluation of random-effects metaanalysis Chest Radiographic and CT Findings of the 2019 Novel Coronavirus Disease (COVID-19): Analysis of Nine Patients Treated in Korea Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study Clinical Characteristics of Coronavirus Disease 2019 in China Imaging Profile of the COVID-19 Infection: Radiologic Findings and Literature Review Chest X-ray features of SARS-CoV-2 in the emergency department: a multicenter experience from northern Italian hospitals Clinical and Chest Radiography Features Determine Patient Outcomes In Young and Middle Age Adults with COVID-19 Clinical Course and Outcomes of Patients with Severe Acute Respiratory Syndrome Coronavirus 2 Infection: a Preliminary Report of the First 28 Patients from the Korean Cohort Study on COVID-19 Fleischner Society: Glossary of Terms for Thoracic Imaging Portable chest X-ray in coronavirus disease-19 (COVID-19): A pictorial review Radiology of bacterial pneumonia Chest X-Ray Findings Among Urgent Care Patients with COVID-19 Are not Affected by Patient Age or Gender: A Retrospective Cohort Study of 636 Ambulatory Patients Evaluation and Treatment Coronavirus (COVID-19) Review of Chest Radiograph Findings of COVID-19 Pneumonia and Suggested Reporting Language Tension pneumothorax in a patient with COVID-19 The role of chest radiography in confirming covid-19 pneumonia COVID-19) Outbreak: What the Department of Radiology Should Know COVID-19 pneumonia | Radiology Case | Radiopaedia.org Post-intubation pneumomediastinum and pneumothorax -background COVID-19 pneumonia | Radiology Case | Radiopaedia.org COVID 19 Pneumonia | Radiology Case | Radiopaedia.org COVID-19 pneumonia | Radiology Case | Radiopaedia.org COVID-19 pneumonia | Radiology Case | Radiopaedia.org COVID-19 pneumonia -evolution over a week | Radiology Case | Radiopaedia.org COVID-19 | Radiology Reference Article | Radiopaedia.org