Comparison of Chest Computed Tomography (CT) Imaging Patterns and Severity among COVID-19 Patients during the First and Fourth Waves in Addis Ababa, Ethiopia

Background Studies done globally had shown that chest imaging patterns of Corona virus disease 2019 (COVID-19) infection varied depending on the strains of the virus and the waves of the pandemic. There is no published literature done in Ethiopia to examine whether there is any difference in chest computed tomography (CT) findings of COVID-19 patients during the first and fourth waves. Thus, this study tries to fill the gap of knowledge in that regard. Objective To describe and compare chest CT scan imaging pattern and assess the predictors of chest CT severity of the first and fourth wave of COVID-19 infection. Methods An institution-based cross-sectional study was conducted on 200 polymerase chain reaction test confirmed COVID-19 patients who underwent chest CT scan imaging in two diagnostic centers in Addis Ababa city. Pioneer and Wudassie diagnostic centers were selected due to the high case load and availability of well-experienced cardiothoracic radiologists. Data were collected from July 1 to August 3, 2022, using a structured Google form sheet questionnaire. Binary logistic regression was performed, and statistical significance was assessed at a level of significance α = 0.05. Results Comparatively higher proportion of patients from the first wave had positive chest CT finding than fourth wave (99% vs. 69%). Bilateral lung involvement and lower lobe predilection were seen for both waves of COVID-19. Ground glass opacity and consolidation were the most common CT features for both waves. Delayed chest CT features such as traction bronchiectasis were primarily seen among first-wave patients. Mean global CT severity score was higher for the first-wave patients (13.18 vs. 8.31), and the mean difference is statistically significant (p < 0.001). Duration of symptoms was a statistically significant predictor of CT severity during the first wave of COVID-19, and patients that presented later than 14 days had 4.12 times higher odds of being in the severe CT score category than those that presented less than 7 days (AOR = 4.12, p = 0.011). There was no statistically significant predictor of CT severity for the fourth wave in this study. Conclusion Chest CT positivity was comparatively higher for first wave patients. Common features included bilateral involvement, lower lobe involvement, ground glass opacity, and consolidation. Mean chest CT severity was comparatively higher for the first wave than the fourth wave, and the duration of symptoms was a statistically significant predictor of the CT severity for first wave.


Introduction
Novel coronavirus disease 2019 (COVID- 19) is a viral infectious disease that was frst detected in Wuhan, Hubei Province of China in 2019 [1].It is a multisystem syndrome with a spectrum of presentation ranging from asymptomatic disease to life-threatening complications requiring invasive management like extracorporeal membrane oxygenation [2].More than 489 million confrmed cases with 6 million deaths were reported globally as of December 2022 [3].
Te frst case in Ethiopia was identifed on March 13, 2020 [4].Tere were four waves of COVID-19 based on the epidemiologic curve.Te frst wave of COVID-19 lasted from June to November 2020, while the fourth wave lasted from December to February 2022.Although genomic sequencing was not available, the frst wave was attributed to the alpha variant, and the fourth wave was attributed to the omicron variant [5].
Reverse transcriptase polymerase chain reaction (RT-PCR) performed on nasopharyngeal and oropharyngeal swab is the gold standard for diagnosing COVID-19 [6].However, chest CT has become an important diagnostic tool for COVID-19 when used in close combination with clinical presentation and epidemiological evidence for the disease.Furthermore, chest CT has also been shown to be useful for the management of COVID-19 patients as well as diferentiating various diseases that COVID-19 could mimic [2].A study done by Fang et al. observed that the sensitivity of CT in diagnosing COVID-19 was very higher than that of RT-PCR [7].
Over the past years, chest imaging fndings of the COVID-19 infection have been described in details.On chest imaging, the lungs could appear normal.However, studies conducted globally had shown that some of the commonest radiological fndings included ground glass opacities on plain radiographs and ground glass attenuation on CT scans, focal or difuse areas of consolidation, and septal thickening most commonly noted in the lower lobes.Bilateral lung changes were also more common than unilateral lung involvement.Atypical fndings of COVID-19 infections described in literature include pleural and pericardial efusions, calcifcations, pneumothorax, cavitation, lung collapse, mediastinal enlargement, and cardiomegaly [8][9][10][11].Based on previous studies, chest CT features were shown to vary based on age of the patients, duration of symptom at presentation, COVID-19 vaccination status, and the strain of the virus [12][13][14][15].A study done in Ghana had demonstrated that more patients in the second wave had evidence of COVID-19 disease on chest imaging than those patients in the frst wave [8].Furthermore, a study done Italy in 2021 had shown that, although there was no diference in the typical chest fnding among the frst four waves of COVID-19, there was a signifcant diference in the occurrence of atypical CT features including pleural efusion [16].Recognizing the diferences in chest CT imaging between various waves COVID-19 is vital to diagnosing and treating patients.
Tis study, to the best of the investigators' knowledge, is the frst study done in Ethiopia with the objective of describing and comparing chest CT imaging fnding as well as assesses the predictors of chest CTseverity among confrmed COVID-19 patients during the frst and fourth wave.Terefore, the study would provide useful information and fll the gap of knowledge on CT features.

Study Setting and Study Design.
Te study was conducted in Addis Ababa city, the capital city of Ethiopia, using a data collected from Pioneer and Wudassie diagnostic centers.Among health facilities with a capacity to perform CT scan imaging, Wudassie diagnostic center and Pioneer diagnostic center were purposively selected due to the high caseload, quality of CT scans (128 slices), and the availability of highly experienced radiologists.Furthermore, in both institutions, chest CT reading was conducted by experienced cardiothoracic radiology subspecialists.Both centers provided imaging and other diagnostics services to dedicated COVID-19 treatment centers and other health facilities.
Noncontrast chest CT scan in standard dose was acquired in volumetric mode, scanning extending from thoracic inlet to caudally include upper abdomen.Te acquired CT images are reconstructed into soft tissue lung and mediastinal window and in 1.2-1.5 mm section thickness for interpretations.Contrast enhanced CT angiography was done for those suspected of having pulmonary thromboembolism.A comparative cross-sectional retrospective study design was conducted from July 1, 2022 to August 3, 2022.

Study Population.
All patients with confrmed COVID-19 who underwent chest CT scan imaging in the selected health institutions during the frst and fourth waves in Addis Ababa city.

Inclusion Criteria.
All patients with PCR confrmed COVID-19 infection that underwent chest CT scan evaluation during the frst and fourth waves of COVID-19.

Exclusion Criteria.
Patients whose chest CT imaging was not reviewed due to poor quality image to interpret and those with incomplete data were excluded from this study.

Sampling Technique and Sample Size.
All patients who underwent a chest CT scan with a confrmed COVID-19 pneumonia were retrieved from the archives of the two institutions, and 240 patients with confrmed COVID-19 cases were found.Among them, 130 patients were from the frst wave, and 110 patients were from the fourth wave.After applying the exclusion criteria, however, a total of 210 were eligible candidates (108 from the frst wave and 102 from the fourth wave).Subsequently, a 1 : 1 proportion of 100 patients from each wave was selected for comparability.

Data Collection Instruments and Techniques.
A structured questionnaire was developed after reviewing various pieces of literature.Te questionnaire was then converted in to a Google form sheet. Te selected patient's electronic records and chest CT images were retrieved from both diagnostic centers.Te data from patient's records and chest CT images were flled into the preprepared Google form using a mobile phone.
2.6.Data Processing and Analysis.Te collected data were assessed for completeness and exported to the statistical package for social science (SPSS) v.20 and after data cleaning, analysis was done.
Categorical variables were summarized using frequency and percentage, while continuous variables were summarized using mean and standard deviation after testing for 2 Radiology Research and Practice normality of data using the Kolmogorov-Smirnov test.Subgroup analysis was performed to assess diference in CT imaging character among various groups.Association between variables was assessed using chisquare (χ 2 ) and Independent sample t-test was performed to assess diferences in mean CT score.Simple binary logistic regression was performed to assess predictors of chest CT severity and those variables with a p value of <0.25 were considered as candidate for multivariable logistic regression to determine predictors of chest CT severity, using α � 0.05 as the signifcance level.Te association was measured using odds ratio with the corresponding 95% confdence interval.
Te chest CT severity score was assessed using a semiquantitative CT severity scoring adopted from Gurumurthy et al. [18].It was calculated based on the involvement of the 5 lobes considering the anatomic extent and the involvement of each lobe.It is scored as follows: 0, no involvement; 1 � <5% involvement; 2 � 5-25% involvements; 3 � 26-50% involvements; 4 � 51-75% involvement; and 5 � >75% involvements.Te global CT score was the sum of each lobar score graded from 0 to 25.A score of <7 is classifed to be mild, between 8 and 17 moderate, and >18 as severe.For analysis purposes, the score was dichotomized in to severe and nonsevere chest CT score.Tose with scores <18 were classifed as nonsevere and those with score ≥18 were classifed as severe CT score.

Ethical Statement.
Ethical approval was obtained from research and ethics committee of the department of radiology.Permission was requested and granted from the respective institutions to access the medical records of the patients and to retrieve review the patients imaging study for the research purpose only.A written informed consent was waived due to secondary nature of the data.However, no identifers were included in the data and patients were not contacted for further follow-up.In addition, all the data were kept confdential and only available to the investigator.Te data would be discarded appropriately after the objective of the study is achieved.

Sociodemographic Characteristics of the Participants.
Among the 200 COVID-19 confrmed patients that participated in this study, majority of them were greater than ffty years of age (43% vs. 62% for frst wave and fourth wave, respectively).Furthermore, more than half of the study subjects were male patients for both groups (53% vs. 69% for the frst wave and fourth wave, respectively).

Comorbidity Status.
Based on comorbidity status, less than half of the study participants (44%) that had undergone CT imaging during the frst wave had comorbid illness, while around two-third (63%) of those participants from the fourth wave had associated comorbidity.Moreover, higher proportions of patients with two or more comorbidity were seen during the fourth wave compared to the frst wave (8% vs. 36% during the frst and fourth wave, respectively).Hypertension (23% vs. 30%) and diabetes mellitus (22% vs. 29%) were the most common comorbidities in both waves of COVID-19.However, there were comparatively higher proportion patients with comorbid underlying lung disease (2% vs. 11%), renal disease (1% vs. 7%), cardiac illness (0% vs. 16%), and malignancy (0% vs. 9%) during the fourth wave than the frst wave (Table 1).

Clinical Symptom of the Participants.
More than twothird (74%) of the participants during the fourth wave of COVID-19 presented within the frst one week of symptom onset, while around half (59%) of patients from the frst wave presented in the frst one week of symptom onset.Patient presentation later than 14 days was more common among patients from the frst wave compared to fourth wave (27% vs 16%, respectively).However, the diference was not statistically signifcant (p � 0.075).
Cough and shortness of breath were the most common presenting symptoms, and there was no statistically signifcant diference between the two waves.Te study had also shown that there were signifcantly lower proportions of patients that presented with symptoms of arthralgia/myalgia (3% vs. 33%) and sore throat (2% vs. 49%) during the frst wave compared to the fourth wave (p < 0.001) (Table 2).

Chest CT Protocols and Findings of COVID-19 Patients.
Te majority of patients had undergone chest CT angiography on both waves with a slightly higher proportion during the frst wave (61% vs. 50% for frst wave and fourth wave, respectively).A comparatively higher proportion of COVID-19 patients had positive chest CT fnding in the frst wave than the fourth wave with 99% CT positivity in the frst wave compared to a 69% in the fourth wave of COVID-19 and it was statistically signifcant (p value <0.001) (Table 3).
Among patients that had positive chest CT features, a signifcant majority (80%) had typical features alone, followed by mixed features (both typical and atypical features), while there were no patients with atypical features alone.However, mixed features followed by typical features were the most common CT feature among patients with positive fndings during the fourth wave (Figure 1).
Among patients that presented with positive CT fnding, a higher proportion of patients (87.9%) during the frst wave presented with two or more chest CT fndings simultaneously compared to those during the fourth wave (78.3%),Radiology Research and Practice but it was not statistically signifcant (p � 0.095).Regarding the symmetry of lung involvement, bilateral CT fnding was the most prevalent feature on patients from both COVID-19 waves.Furthermore, a higher proportion of bilateral lung involvement was seen among patient in the frst wave (p � 0.02).
In relation to the zones of distribution, there was a statistically signifcant diference in distribution (p � 0.004) with peripheral distribution as the most common presentation during the frst wave (58; 58.6% vs. 25; 36.2%), while difuse lung involvement was the most common for the fourth wave (39; 39.4% vs. 37; 53.6%).Central distribution was the least common feature for both waves.
With regard to ground glass opacity, comparably higher percentage of patients presented with GGO in the frst wave than fourth wave but it was not statistically signifcant (p � 0.359).In addition, a higher proportion of patients presented with peripheral GGO in the frst wave (56.63% vs. 42.59%)while there were higher proportion of central and peripheral GGO in the fourth wave (39.76% vs. 55.56%).Meanwhile, central GGO alone was the least common feature in both waves of COVID-19 among patients that had GGO on chest CT (3.61% and 1.85%).Among patients during the frst wave that presented with consolidation on chest CT, a signifcant majority had peripheral consolidation (74.71%; p < 0.001) and hazy margin of consolidation (93.1%; p < 0.001).Meanwhile, difuse consolidation (53.84%) and sharp margin of consolidation (58.97%) were predominantly seen in fourth wave of COVID-19 patients.Central consolidation was the least common feature in both waves (Table 4).

Atypical Chest CT Features among COVID-19 Patients with Positive
Findings.In this study, mediastinal lymphadenopathy was more commonly seen among frst-wave patients (12.12% vs. 1%) and it was statistically signifcant (p � 0.024).Meanwhile, pleural efusion (p < 0.001) and nodular opacity (p < 0.001) were predominantly observed among patients from the fourth wave.In addition, there were no patients that presented with tree in bud opacity among frst-wave patients while 5 (7.24%) of the study participants from the fourth wave had the feature.Tere were no patients that had either pneumothorax or cavitation in both waves of COVID-19 (Table 5).

Chest CT Severity Score of COVID-19 Patients during the
First and Fourth Waves.Te study has revealed that the mean CT severity score for patients during the frst wave of COVID-19 was higher than those participants from the fourth wave (13.18 vs. 8.31) and the mean diference was statistically signifcant (p < 0.001) (Table 6).
Regarding CT severity score, there were more than two times higher proportion of patients with severe CT score category among subjects from the frst wave than the fourth wave (28% vs. 13%).
Based on age category, the lowest proportion of patients with severe chest CTscore was seen among young age groups who are less than the age of 35 years.Te study had also revealed females had a higher proportion of patients with severe chest CT score compared to males in both waves.However, there was no statistically signifcant association between CT severity categories with age and sex.With regards to the duration of symptoms, those patients presenting after two weeks of symptoms had the highest proportion of patients with severe CT score in both waves while those presenting with in the frst week had the smallest proportion.Furthermore, the association between duration of symptoms with chest CT severity category was statistically signifcant (p value �0.037) for patients from the frst wave of COVID-19.
In this study, comparatively higher proportions of patients with severe CT score category were seen among patients who had comorbidities compared to their counterparts.Meanwhile, there were a comparable proportion of patients among participants who have Radiology Research and Practice hypertension and diabetes mellitus in comparison to those who don't in both waves of COVID-19.However, there was no statistically signifcant association during both waves between CT severity category with age, sex, comorbidity, hypertension, and diabetes mellitus, respectively, in this study (Table 7).

Factors Associated with Chest CT Severity Category among COVID-19
Patients.On bivariable binary logistic regression, only duration of symptom was found to be a statistically signifcant predictor of CT severity (p value �0.038).Tose variables with p value of <0.25 on bivariable association were included in the multivariable model to adjust for possible confounding efects.
After adjusting for age category and sex, duration of symptom was found to be a statistically signifcant predictor of chest CT severity during the frst wave and those patients that presented after 14 days of symptom onset had 4.12 times higher odd of having severe chest CT score compared to those who presented in the frst 7 days of symptom onset (AOR � 4.12; 95% CI: 1.38, 12.27) (Table 8).Binary logistic regression to determine predictor of severity among fourth-wave patients was not done since there was no observed association between severity category and other variables on bivariate association (Table 7).

Discussion
In this study, CT positivity was higher among patients from the frst wave than fourth wave.Tis could be related to the hypothetical COVID-19 variant during those waves.Our study showed that CT angiography was done during the frst wave of COVID more than the fourth wave.Tis could be due to the more severe feature of the frst wave mimicking pulmonary thromboembolism.
Te study found that GGO and consolidations were the most common chest CT features among COVID-19 patients during both waves.Furthermore, bilateral lung involvement with lower lobe involvement was predominantly seen in both waves which were in comparative agreement with previous knowledge [9,19,20].Traction bronchiectasis and mediastinal lymphadenopathy, delayed CT features of COVID-19, were seen predominantly among the frst-wave patients [9].Tis could correlate with the higher proportion  6 Radiology Research and Practice of delayed presentation observed during the frst wave compared to the fourth wave which allows enough time for the development of fbrosis and bronchiectasis.However, comparatively higher patients with atypical and mixed features were seen during the fourth wave.Previous fndings suggested that pleural efusion is an uncommon CT feature among COVID-19 patients [16,21], but our study found that around a third of patients during the fourth wave had pleural efusion.It was found that there were statistically signifcant higher proportions of patients with severe chest CT features in the frst wave of COVID-19 compared the fourth Radiology Research and Practice wave.It was also shown that there was a statistically signifcant mean diference between the two waves with the frst wave having a higher mean CT severity score.Tis could be due to diferences in the hypothetical strain of COVID-19 and its virulence.A study done in South Korea had revealed patients with Delta variant had more severe global CT scores compared to those who contracted the omicron variant [22,23].Te frst wave of COVID-19 in Ethiopia was suspected to be related to the alpha variant while the fourth wave was attributed due to omicron [4].Similar to previous studies, this study had revealed that duration of symptoms was a statistically signifcant predictor of severity during the frst wave of COVID-19 after adjusting for other variables [11,24].In comparative agreement with previous studies done, we did not fnd any association between chest CT severity score and comorbidity status in our study [25,26].
Chest CT imaging played a crucial role in the diagnosis and management of COVID-19 in both waves of COVID-19 infections by determining severity score and enabling to decide on the mode of management.It was also useful in diagnosing other mimickers like pulmonary thromboembolism.Despite its value, the routine use of chest CTand CTangiographic studies during COVID-19, especially in resource limited settings, added additional burden to the limited the imaging resources and afected imaging of other diseases, particularly in oncologic patients who do have regular imaging workups with CT.

Strength and Limitation.
Tis study is the frst study that tried to explore diferences in CT imaging between diferent waves of COVID-19 infection.However, this study was not without limitations.Firstly, this study utilized secondary  8 Radiology Research and Practice data and lacks important variables including smoking history and vaccination status, re-infection status.Secondly, the study included subjects from purposively selected two diagnostic centers which would limit the generalizability of the result.Finally, features of pulmonary thromboembolism and COVID-19 on chest CT without contrast can be similar.Tis could have led to misclassifcation and overestimation of the magnitude of chest CT positivity.

Conclusion and Recommendation
Te study has revealed that chest CT positivity was comparatively higher for frst-wave patients.Common chest features included bilateral involvement, lower lobe involvement, and consolidation.Te mean chest CT severity was comparatively higher for frst wave than fourth wave, and the duration of symptoms was a statistically signifcant predictor of CT severity for frst wave.We recommend further with variables such as smoking history and immunization history included as well as a study with a larger sample size involving multiple diagnostic center.

Figure 1 :
Figure 1: Chest computed tomography (CT) features comparison between the frst and fourth waves of COVID-19.

Figure 2 :
Figure 2: A 45 years old male patient from frst wave with bilateral difuse confuent consolidation and ground glass opacity.

Figure 3 :
Figure 3: A 60 years old male patient from the fourth wave with peripheral patch ground glass opacity and minimal bilateral pleural efusion.

Table 1 :
Sociodemographic and comorbidity profle of patients during the 1 st and 4 th wave COVID-19 in Addis Ababa city.

Table 3 :
Comparison of chest computed tomography (CT) fnding of COVID-19 patients during the 1 st and 4 th waves COVID in Addis Ababa city.

Table 6 :
Independent sample t-test to compare mean CT severity scores between 1 st and 4 th waves of COVID-19.

Table 8 :
Predictors of chest CT severity category among COVID-19 patients during the frst wave of COVID-19.