The substantial increase in the number of confirmed and suspected COVID-19 cases has overwhelmed healthcare systems worldwide. Therefore, methods of accurate and rapid diagnosis and triaging are essential to avoid a crisis. Laboratory testing is the standard for diagnosing COVID-19 infection; however, delayed test results and insufficient supply of laboratory kits to cover the increasing number of suspected cases are major drawbacks because early quarantine and treatment are essential for control of the disease [4]. Xie et al. [8] also reported that 5 out of 167 patients (3%) had negative RT-PCR for COVID-19 at initial presentation despite chest CT findings typical of viral pneumonia.
Chest CT appears to be the most sensitive diagnostic tool for COVID-19 pneumonia. Various studies have reported that the morphology and extent of the lung lesions correlate with the severity of the inflammatory process [9, 10]. However, the extent of pulmonary inflammation caused by COVID-19 does not correlate with the severity of clinical manifestations presented at first evaluation in the emergency department [9]. Because CT cannot be performed for all suspected patients, another rapid, inexpensive, and widely available test is needed for the preliminary assessment of patients. Several studies have shown correlation between ABG values and the extent of pulmonary lesions on CT [5, 6]. Furthermore, COVID-19 can be complicated by acute respiratory distress syndrome (ARDS) [11]. Three degrees of ARDS based on the degree of hypoxemia are described: mild (PaO2/FiO2 ≤ 300 mmHg), moderate (PaO2/FiO2 ≤ 200 mmHg), and severe (PaO2/FiO2 ≤ 100 mmHg) [12]. Thus, we used PaO2/FiO2 values at admission and compared them with the CT total severity score. A moderate negative correlation was observed between the PaO2/FiO2 ratio and the CT total severity score, r = − 0.42, p < 0.001, indicating that pulmonary function decreased causing oxygenation impairment as the total severity score increased. This agrees with the results of Turcano et al. [6] who reported that a higher percentage of the lung being involved in the inflammatory process correlated with decreased oxygenation capacity which explained the difficulty of oxygenating the blood, even with high FiO2 values. Shang et al. [5] also recorded that the total CT score had a moderate correlation with arterial blood gas indices.
The most common presenting symptoms in this study were fever, cough, and malaise, and other less common symptoms were headache, dyspnea, sore throat, and diarrhea. Shi et al. [13] have also reported fever in 73% of patients and dry cough in 59%. However, 9.7% of the cases in this study were asymptomatic virus carriers while others showed only mild-onset symptoms with no fever and presented only as contacts of positive COVID-19 patients. Twenty-one of the asymptomatic cases (53.8%) showed positive chest CT findings suggesting that chest CT scans or laboratory tests should be done in asymptomatic cases with a history of contact with COVID-19-positive patients to allow early diagnosis and isolation to optimize control of the disease. The number of asymptomatic carriers may be higher than estimated and underdiagnosed due to insufficient laboratory and radiological testing. In the study by Nishiura et al. [14], asymptomatic carriers of COVID-19 have been estimated to comprise 17.9–33.3% of all infected cases. In this study, a normal chest CT scan was found in 84 patients (20.8%) despite positive PCR results. Chung et al. [7] have also reported that 14% of the patients (three out of 21) presented with a normal chest CT scan.
The positive chest CT findings in this study were classified into three main patterns: GGO only, consolidation only, and GGO with consolidation. GGO were the most common lesions at presentation and were observed in 40% of the patients; a slightly lower percentage of patients (34.4%) had associated consolidation. These results were consistent with the study by Li et al. [15] on 51 patients with confirmed COVID-19 infection by nucleic acid testing. Their results showed that only 3.9% of the patients did not have GGO or consolidation which indicated that these lesions are main signs of COVID-19 on CT images. Yoon et al. [16] have also reported that GGO lesions on chest CT without any consolidation were observed in 45% of their cases and in 45–67% of Chinese COVID-19 patients.
The shape of consolidation among our patients varied. Round consolidation was the most common shape observed in 11.9% of the patients, followed by arcade-like consolidation (10.4%), peripheral consolidation parallel to the pleura (7.3%), irregular patchy consolidation with architectural distortion (6.8%), peripheral wedge-shaped consolidation (6.2%), and lobar consolidation (5.4%). Arcade-like consolidation is a typical feature of perilobular fibrosis, which is frequently observed in cryptogenic organizing pneumonia. In 2004, Ujita et al. [17] observed the presence of perilobular fibrosis, with an “arch” pattern, in more than half of the patients with cryptogenic organizing pneumonia, which may be a sequalae of perilobular inflammation.
Associated secondary findings were interlobular septal thickening (152 patients, 47.5%), crazy paving (56 patients, 17.5%), vascular enlargement within the lesion (146 patients, 45.6%), and air bronchogram (48 patients, 15%). No lung cavitation was recorded. Bai et al. [18] described subsegmental vascular enlargement in 59% of the patients with COVID-19 pneumonia versus 22% in those with non-viral pneumonia. Reticular pattern with interlobular septal thickening was reported in several studies as one of the common manifestations of COVID-19 on chest CT, second to GGO and consolidation [13, 19, 20]. With an increase in the disease course, the prevalence of reticular pattern may increase in COVID-19 patients [13]. Posterior location of the opacities was observed in 58.1% of patients while combined peripheral and central distribution was observed in 57.7%. We explain this by late presentation of most of the patients (i.e., more than 7 days after the onset of symptoms). This was also evident in the pattern of affection because involvement of all the lung lobes was the most common pattern seen in 204 patients (50.5%).
Pleural thickening was observed in 79 patients (24.7%). In all cases, it was located adjacent to an area of consolidation or crazy paving indicating pleural involvement in the inflammatory process; however, no pleural effusion was recorded in any of the patients.
On chest CT, the pericardium is best visualized along the right ventricle, due to thin epicardial fat and the vicinity of pulmonary parenchyma along the lateral and posterior left ventricle wall making it difficult to visualize the pericardium on that side. Normal pericardial thickness ranges from 0.7 to 2.0 mm on CT images [21]. In this study, 16 out of 22 patients who presented with pleuritic centrothoracic chest pain (which improved when sitting forward and worsened with supine position in 11 patients) showed pericardial thickness of > 2 mm along the right ventricle suggesting pericarditis. This was confirmed by pericardial rub during physical examination and by the presence of widespread ST elevation on the ECG trace in some patients and T wave flattening in the inferior leads (II, III, and aVF) in others. Pericardial effusion was not observed in any of the patients. The affection of the cardiovascular system in COVID-19 infection was described in a study of 83 patients with severe and critical COVID-19 infection who underwent a CT scan. Chest pain was reported in 6% of the patients and pericardial effusion was observed in 4.8%, suggesting that acute pericarditis could be underdiagnosed [22]. Puntmann et al. [23] conducted a study that included 100 patients who had recovered from COVID-19 disease and underwent cardiac MRI. A total of 78% of the patients had abnormal findings on cardiac MRI, and 22% had pericardial enhancement.
The findings on chest CT may not be specific and can be similar to those in other viral pneumonia. However, together with clinical and laboratory findings, CT of the chest can make an essential contribution to the management of the disease. Among the disadvantages of using CT in the diagnosis of COVID-19 are radiation exposure to the patient, risk of COVID-19 transmission to uninfected healthcare workers and other patients, consumption of personal protective equipment, and need for cleaning and downtime of radiology rooms in resource-constrained environments [4]. In this study, no transmission of COVID-19 to any of the radiology staff was reported as a result of implementing the strict use of personal protective equipment (in those with direct contact with the patients). In addition, downtime caused by cleaning and disinfection was not a problem, because the CT machine was dedicated for COVID-19 patients only, which allowed sufficient time for examining new and follow-up cases.
This study had a few limitations. First, no follow-up imaging was done and thus the dynamic changes of COVID-19 pneumonia which occur over time were not assessed. Furthermore, the outcome of the disease and mortality were not evaluated which would be of great benefit in determining certain prognostic factors as well as their correlation with the PaO2/FiO2 ratio.