Differential diagnosis of COVID-19 from community acquired pneumonia by CT scan and follow-up

Objective: Coronavirus disease-19(COVID-19) is the most serious infectious disease in the world at present. Accurate diagnosis of this disease in clinic is very important. this study aims to improve the differential ability of COVID-19 and other community acquired pneumonia(CAP) by CT,and to evaluate the short-term prognosis of patients. Methods: The clinical and imaging data of 165 COVID-19 and 118 CAP patients diagnosed in seven hospitals in Anhui Province,China from January 21,2020 to February 28, 2020 were retrospectively analyzed. Results (cid:0) The sensitivity and specicity of age, white blood cell count, and ground glass shadow in the diagnosis of COVID-19 were 92.7% and 66.1%,respectively. Pulmonary consolidation,ber cord,and bronchial wall thickening were used as indicators to exclude COVID-19.The sensitivity and specicity were 78.0% and 63.6%, respectively.The follow-up results showed that 67.8% (112/165) of the cases of COVID-19 had abnormal changes in lung residual levels,and the pulmonary sequelae of patients over 60 years of age became more severe with age. Conclusion (cid:0) CT combined with clinical conditions, lung lesion density (cid:0) morphological characteristics and other associated signs, has a high accuracy for the early diagnosis of COVID-19 and the differential diagnosis of CAP.patients with COVID-19 infection over 60 years have poor prognosis.


Introduction
Coronavirus disease-19  is an acute respiratory infectious disease caused by new coronavirus infection [1]. As of May 23, 2020. a total of 5299580 cases were diagnosed worldwide. Common symptoms after coronavirus infection are respiratory symptoms such as fever, cough, shortness of breath, dyspnea and so on [2][3][4]. In more serious cases, infection can lead to pneumonia, severe acute respiratory syndrome, renal failure, and even death. The study of antibodies against new coronavirus is still under study [5][6], and there is no speci c and effective treatment [7].
The gold standard for COVID-19 diagnosis is positive for new coronavirus nucleic acid test [8][9], but due to the hysteresis of the test results or "false negatives" in the report, some patients cannot be diagnosed and treated in time.COVID-19 has certain imaging characteristics in lungs .chest CT examination is an important method for clinical diagnosis of COVID-19. During the current outbreak of COVID-19, early detection of lesions is very important for early isolation and treatment, and to prevent the spread of disease. Previous studies have mostly focused on the imaging manifestations and changes of COVID-19 [10][11][12]. The disease needs to be distinguished from many diseases in imaging, especially the community-acquired pneumonia (CAP). However, a large number of comparative studies on imaging in this area are still lacking. Therefore, the purpose of this study is to compare the clinical features and CT imaging manifestations of COVID-19 with community-acquired pneumonia, to explore the value of CT in the diagnosis and differential diagnosis of COVID-19, and to evaluate the short-term prognosis of the two types of pneumonia. The inclusion and exclusion process of patients is shown in Figure 1, and the results of CAP-RPP are shown in Figure 2.

2.3CT image evaluation
CT images of patients in the COVID-19 and CAP groups were performed by 3 experienced radiologists (working in radiography for> 5 years). Use double-blind method to read independently. Carefully observe and record the location, shape, number, size, burr and extent of the lesion. If the diagnosis is inconsistent, the consensus reached by the 3 experts shall prevail.

Statistical analysis
The measurement data is expressed by sequelae`x±s, the comparison between groups is expressed by independent sample t test; the count data is expressed by percentage(%), and the comparison between groups is expressed by χ 2 test. Correlation analysis was used to examine the relationship between age and the size of the lung lesions, the number of involved lobes, and the CT ndings of patients with COVID-19. SPSS 20.0 software package was used for the above statistical analysis.

Study Population Characteristics
The baseline data and clinical manifestations of the two groups are shown in Table 1. The differences in age white blood cell (WBC) Lymphocyte proportion and the presence or absence of underlying disease in the lungs of the two groups of patients were signi cant.χ 2 = 8.03, P <0.001;χ 2 = 5.58, P <0.001; t = 4.62, P = 0.003; χ 2 = 21.72, P <0.001.

3.2CT imaging
CT ndings of chest in patients with COVID-19 and CAP group onset 3-7 days were showed in Table 2. The ground glass opacities GGO on chest CT images were found in140 cases (84.8%) of in COVID-19 group.The typical lesions early in the COVID-19 group were shown as single or multiple small rounds. In this group, 53 cases (32.1%) were found, 87cases (52.7%) of ground-glass-like density shadows with small spot large piece or fusion ( Figure 1A).In the CAP group, there were 47 cases (39.8%) of ground glass shadow in the lungs, 71 cases (60.2%) showed consolidation, and the density difference between the two groups was signi cant (χ 2 = 48.75, P < 0.001).In the COVID-19 group, 54 cases (32.8%) of "crazy paying" pattern were also observed, which was mainly characterized by interlobular septal thickening( Figure 3B) were different from the subpleural reticular or honeycomb changes of 34 cases 28.8% in the CAP group. Eight (4.8%) patients in the COVID-19 group showed "walking" lung lesions ( Figure 3C-D).
In addition to some fresh exudative in ammatory lesions mentioned above, chest CT in the CAP group is often accompanied by chronic obstructive manifestations such as emphysema, bullae, and "mosaic" signs or small airway obstructive changes. In the CAP group, 68 cases (57.6%) with brous cords or pulmonary texture aggregates were signi cantly higher than those in the COVID-19 group (5cases, 3.1%) (x 2 =107.14, P< 0.001); the bronchial wall thickening was also in the CAP group (37 cases, 31.4%) and 18 case (10.1%) was seen in the COVID-19 group, the difference between the two groups was signi cant (χ2 = 18.4 P <0.001). In CAP group, there were 12 cases (10.2%) with small cavities and 8 cases with pulmonary bullae, which were signi cantly higher than those in COVID-19 group.

The relationship between clinical characteristics and imaging manifestations
The total number of lung lobes involved in the COVID-19 and CAP groups was 384 and 257, respectively.
The single lesion in the COVID-19 was signi cantly higher than that in the CAP group (χ 2 = 8.38, P = 0.03), and the multi-leaf multi-foci of the CAP were signi cantly higher than the COVID-19 group (χ 2 = 6.68, P = 0.01). The lesions in both groups were mainly in the middle and outer of lobes. There were 148 cases (89.7%) of COVID-19 group located in the outer zone, and 62 cases (52.5%) of CAP. There was a signi cant difference between the two groups (χ 2 = 49.6, P < 0.001). Statistics showed that there was a positive correlation between the age of onset and the number of involved leaf segments (r = 0.62, P < 0.001).
In the above clinical and imaging manifestations, age factors, the total number of leukocytes, ground glass shadow have a higher signi cance in the diagnosis of covid-19 (Table 3). The characteristics of CAP group were consolidation of lung lesions, accumulation of brous cords or textures and thickening of bronchial wall (Table 4).

CT follow-up of COVID-19
A total of 135 cases in COVID-19 group were reexamined 3-7 days after the rst CT examination, 32 cases ( 23.70%) showed anlargement of lesion size and increase of the number, which were classi ed asprogressing, 25 cases (18.5%) were stable, and 78 cases (57.8%) were improvement.
All patients with COVID-19 underwent CT Re-examination one month after treatment, , and 15 cases (15.6%) showed no abnormalities in both lungs at the time of reexamination Figure 3E-F ; the rest 87 cases (52.7%) showed localized strip shadow, 43 cases (26.1%) of extensive brosis, 10 cases (6.1%) of limited aky shadow, 6 (3.6%) cases of wide strip shadow 4 cases (2.4%) of diffuse grid shadow ( Figure  3G-H). The relationship between age and prognosis shows that in patients with COVID-19 <60 years of age, the late CT changes have no signi cant relationship with age; in patients over 60 years of age, the CT sequelae changes with increasing age (Figure 4).

Discussion
At present, retrospective studies on various aspects of COVID-19 are being further developed. Due to the strong infectivity of COVID-19, the source of the pathogen is unknown, no speci c treatment method, and the mortality rate is high [14], so the early diagnosis of COVID-19 a challenge for us. Correct and timely diagnosis is great signi cance for the prognosis and prevention of dissemination. Due to the delayed laboratory etiology results and limited sensitivity, there is still controversy for early clinical diagnosis [15].
Thoracic imaging plays an important role in early diagnosis. This study believes that in the case of an epidemic area, the key to early imaging diagnosis is to distinguish it from other CAP.
The data were similar to most clinical observations [16][17][18]. The age of COVID-19 group was mostly young adults, 46 cases (71.9%) were <60 years old, and only 7 cases (15.2%) in the CAP group, It does not comply with the general law of spread of infectious diseases, that is, children and the elderly are often susceptible groups, Young adults had the strongest resistance to infectious diseases and the incidence rate should be low. 72314 cases of Chinese Center for Disease Control and Prevention showed that the incidence rate of China under 10 years old was less than 1% [19]. The high incidence rate of young adults can't exclude the possibility of more virus contact and infection opportunities due to more social interactions. Of course, further studies on epidemiology and pathogenesis are needed. Fever was the prominent clinical manifestation in both groups, but the degree of temperature rise was more obvious in COVID-19 group than that in CAP group. In the early stage of the disease (within 7 days), the total number of leukocytes in the observed cases of COVID-19 was normal or low (< 4×10 9 /L), or showed a decreasing trend, similar to that reported in the literature [20], which may be caused by the decrease of lymphocytes in the rst few days after the disease. The symptoms of muscle soreness and fatigue in COVID-19 group were signi cantly higher than those in CAP group, while the symptoms of expectoration in the latter group were signi cantly higher than those in the former group,which may be related to the fact that CAP is prone to produce more purulent secretion due to bacterial infection [21]. The results showed that the incidence of heart disease, hypertension, chronic lung disease and diabetes in CAP group was signi cantly higher than that in COVID-19.
The chest CT changes of COVID-19 usually showed ground glass opacities shadow of different size 3-4 days after clinical onset.The size and shape of COVID-19 ground glass opacities shadow are various, among which the most typical is round or quasi round ground glass shadow, accounting for 32.1% (53 cases) in this group, most of which are the early changes of young people. The statistics of this group showed that the single lesion of COVID-19 was signi cantly higher than that of CAP group, which may be related to the early stage of COVID-19. With the progress of the disease, the area of ground glass opacities shadow in some patients has expanded, and gradually developed into ground glass shadow with multiple patches, large fusion or large and small pieces coexisting, but the density change is relatively small, which is in contrast with the rapid consolidation in the progress of CAP pneumonia, while the area expansion is relatively not obvious. Of course, there are also some light patients with lung lesions area no longer expanded until absorption. The pathological basis of ground glass changes is mainly related to the local severe pulmonary edema, hyaline membrane formation and uid exudation in the alveolus cavity in the early lesions. Another 32.7% (54 cases) of COVID-19 showed ne reticular shadow overlapped in ground glass shadow, which may be related to the alveolar edema in the lung lesion area of COVID-19 and the slight thickening of alveolar septum with in ltration of monocyte, lymph and plasma cells [22]. Because some CAP patients also showed ground glass opacities shadow (47 cases, 39.8%), which overlapped with the change of COVID-19 lung disease, the imaging manifestations must be combined with clinical and other lung changes.
In CAP group, the new exudative in ammatory lesions were often accompanied by obvious brous components or texture aggregation and cord adhesion in other parts of the lung.which were manifested as bronchitis or bronchitis thickening and blurring of the outer edge, suggesting that these patients had experienced the process of lung in ammation in the past, and left some chronic in ammation or later changes, This is quite different from the acute onset, rapid progress and multiple onset of lung in ammation of COVID-19. Therefore, this study suggests that the lung with brous cord and bronchial wall thickening can be used as a counter indication to exclude COVID-19. In addition, a certain proportion of emphysema, pulmonary bullae, reticular or honeycomb changes under the pleura and "mosaic" sign and/or small cavity shadow in the consolidation area were observed in CAP group. In conclusion, the diversity of lung lesions and the coexistence of new and old lesions in CAP group are helpful to distinguish from COVID-19.
In this group, 6 patients with COVID-19 underwent chest CT Reexamination about one week after the onset of the disease, the lesions showed "wandering" characteristics, which may indicate the heterogeneity of pathological changes in different lung areas, that is, the early changes and the changes in the organic phase of diffuse alveolar injury can appear in different segments at the same time, which is more common in young patients. Due to the small number of cases in this group, whether it has a good diagnostic speci city remains to be further con rmed.
The results of this study showed that there was a positive correlation between the age and the size of lesion and the total number of segments involved in the lesion in the COVID-19 group. The correlation coe cient between the age and the size of lesion was 0.522, P < 0.001, and the correlation coe cient between the age and the total number of segments involved was 0.531, P < 0.001. In this group, the range of lesions in the rst visit of the elderly patients was signi cantly larger than that in the young patients, and the number of involved segments was also signi cantly higher than that in the latter. This may be due to the rapid development of the disease in the elderly infected patients. The rst CT scan has involved multiple segments, which is also one of the reasons affecting the prognosis of the patients.
The follow-up of COVID-19 group in the later period showed that the patients aged more than 60 years old had a lot of changes left in the lung in the later period, and the most serious manifestation was extensive brous cord shadow in the lung. In this group, 32 patients 19.4%) were all over 60 years old, suggesting that the prognosis of the patients over 60 years old was poor. The follow-up of 2-3 months showed that the remaining changes in the lung could still be absorbed, but the absorption was slow.

Conclusion
At present, the most widely used clinical diagnosis of COVID-19 is the positive detection of new coronavirus nucleic acid, but the rapid spread and strong infectivity of COVID-19, the detection reagent can not meet the clinical application in time. imaging examination, especially CT examination plays an important role in the early diagnosis and screening of COVID-19, At present, most of the early diagnosis of COVID-19 is based on imaging. Therefore, for the diagnosis of COVID-19, the accuracy of differential diagnosis between COVID-19 and CAP should be improved by closely combining the epidemiological history, the clinical situation of patients and the changes of lung image.

Declarations
Ethics approval and consent to participate Acquisition of all clinical images was granted by subject verbal consent. The images in this paper are obtained from 7 research centers. These are repositories of anonymised images accessible locally for educational purposes and no identi able information is stored or available. Ethical approval for their use in publication was therefore deemed unnecessary.    days later, showing that the original consolidation of both lungs was diminished, and another ground glass-like density shadow appeared in the middle lobe of the right lung, and the lesions in the lungs showed "walking" E-F. A 32-year-old female patient reexamined after 10 days of treatment. CT showed that multiple lung lesions disappeared G-H. a 68-year-old female patient, was reexamined after 15 days of treatments. CT showed that the diffuse lesions of both lungs were absorbed into extensive brosis.

Figure 4
The relationship between the age and the prognosis of COVID-19 patients in this group.