key: cord-307574-vmik4neu
authors: Ottaviani, S.; Franc, M.; Ebstein, E.; Demaria, L.; Lheure, C.; Debray, M. P.; Khalil, A.; Crestani, B.; Borie, R.; Dieudé, P.
title: Lung ultrasonography in patients with COVID-19: comparison with CT
date: 2020-08-12
journal: Clin Radiol
DOI: 10.1016/j.crad.2020.07.024
sha: 
doc_id: 307574
cord_uid: vmik4neu

Abstract Aim To determine whether findings from lung ultrasound and chest high-resolution computed tomography (HRCT) correlate when evaluating COVID-19 pulmonary involvement. Materials and methods The present prospective single-centre study included consecutive symptomatic patients with reverse transcription polymerase chain reaction (RT-PCR)-proven COVID-19 who were not in the intensive care unit. All patients were assessed using HRCT and ultrasound of the lungs by distinct operators blinded to each other’s findings. The number of areas (0–12) with B-lines and/or consolidations was evaluated using ultrasound and compared to the percentage and classification (absent or limited, <10%; moderate, 10–25%; extensive, 25–50%; severe, 50–75%; critical, >75%) of lung involvement on chest HRCT. Results Data were analysed for 21 patients with COVID-19 (median [range] age 65 [37–90] years, 76% male) and excellent correlation was found between the ultrasound score for B-lines and the classification (p<0.01) and percentage of lung involvement on chest HRCT (r=0.935, p<0.001). In addition, the ultrasound score correlated positively with supplemental oxygen therapy (r=0.45, p=0.041) and negatively with minimal oxygen saturation at ambient air (r=–0.652, p<0.01). Conclusion The present study suggests that among COVID-19 patients, lung ultrasound and HRCT findings agree in quantifying lung involvement and oxygen parameters. In the context of the COVID-19 pandemic, lung ultrasound could be a relevant alternative to chest HRCT.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (COVID- 19) is an emergent respiratory viral disease first reported in China that quickly spread worldwide. On 2 July 2020, the global COVID-19 pandemic had resulted in more than 10 billion cases and 512,842 deaths worldwide (1) . Similar to other coronavirus pneumonia diseases such as SARS caused by SARS-CoV and Middle East respiratory syndrome coronavirus, COVID-19 can also lead to acute respiratory distress syndrome. Recent studies have demonstrated that chest high-resolution computed tomography (HRCT) is important for screening, primary diagnosis, and evaluation of COVID-19 severity (2, 3) . Typical chest HRCT features are ground-J o u r n a l P r e -p r o o f glass opacities, interstitial changes, or multifocal patchy consolidation, with peripheral and posterior distribution (2) . The first imaging features are present in the early phase of the disease, whereas consolidations are seen later in the disease course (4) .

HRCT may have higher sensitivity for COVID-19 diagnosis than reverse transcription polymerase chain reaction (RT-PCR) (3) . In the context of a pandemic with a high fatality rate, there is an urgent need for a prompt identification of patients with COVID-19, thus limiting the access to chest HRCT. In addition, the need for transporting such contagious patients can be a limitation to wide use of the HRCT (5) . In this context of limiting the spread of infection, ultrasonography, with the possibility of sterilising equipment, seems useful. Moreover, ultrasound avoids radiation exposure, allows for repeated procedures, and can be performed at the bedside, which limits the need to transport patients (6-8).

Because pulmonary damage occurs mainly in the periphery of the lungs, these lesions could be accessible to ultrasonography. Lung ultrasound is based on the detection of pleural irregularities, consolidation, and B-line artefacts (9) . Among patients with normal aeration of the lung, the only ultrasound-detected structure is the pleura, seen as a hyperechoic line. This pleural line is suggested to represent an artefact secondary to a reflection phenomenon between alveolar air and the thoracic wall (10) . Some hyperechoic horizontal lines arising at regular intervals from the pleural line are non-pathological and are called A-lines (Fig. 1a) . When the alveolar air disappears due to the presence of any lung liquid or collagen, this acoustic artefact is decreased, and ultrasound can be reflected at deeper zones in the lung.

This modification of the propagation of ultrasound waves creates vertical artefact lines called B-lines (Fig. 1) . B-lines appear as a hyperechoic comet-tail vertical line extending from the pleural line to the bottom of the screen with fading. Ultrasound B-J o u r n a l P r e -p r o o f lines evoke lung interstitial syndrome and are different from Kerley B-lines, which are X-ray features of interlobular septum thickening. The number and distance between B-lines is also important for interpreting ultrasound of the lungs. The number increases when alveolar air decreases and the distance between B-lines <3 mm evokes ground-glass opacities (11) . Previous studies demonstrated that lung ultrasound is efficient for detecting interstitial lung disease, subpleural consolidations, or acute respiratory syndrome (9, (12) (13) (14) (15) . The detection of consolidation with dynamic air bronchograms may suggest a diagnosis of bacterial pneumonia (16). Although HRCT allows for a better detection of pulmonary tumours, ultrasound is able to detect chest-wall tumours by showing a mass with hypervascularisation on power Doppler (17) . Regarding COVID-19, four case studies involving 12 to 30 patients (18) (19) (20) (21) and two case reports (22, 23) are available. The proportion of patients with B-lines and consolidations ranged from 90% to 100% and 20% to 50%, respectively, in the case studies.

The aim of the present study was to determine the usefulness of lung ultrasound for COVID-19 diagnosis by investigating the correlation between ultrasound and chest HRCT findings.

This prospective single-centre study included consecutive symptomatic patients, age >18 years, with COVID-19 proven by positive RT-PCR for SARS-CoV2 who were not in the intensive care unit (ICU). All patients were admitted in the unit between 23

March and 14 April 2020. Exclusion criteria were hospitalisation in the ICU during the last month and previous thoracic surgery, acute heart failure, and bacterial J o u r n a l P r e -p r o o f pneumonia. All included patients underwent clinical evaluation (respiratory frequency, oxygen saturation measurement, quantification of oxygen supplementation), including disease history (presence of fever, diarrhoea, cough, dyspnoea) and Creactive protein (CRP) measurement. All patients underwent chest HRCT and lung ultrasound on the same day.

Informed consent was obtained for all patients. Because the study was observational without an invasive procedure, no ethics committee approval was required.

HRCT was performed by the local senior radiologist (>20 years of experience) who was blinded to clinical and ultrasound findings. The following characteristics were assessed (2): (1) presence of ground-glass opacities, (2) presence of consolidation, (15) and (2) presence of consolidation defined as a poorly delineated hypoechoic tissue structure. The presence of multiple (at least three B-lines) or confluent comet tails in each analysed area was graded B-score 1. At least one consolidation in the examined area was graded C-score 1. Thus, for each patient, the sum of the total B-score and total C-score was from 0 to 12.

To evaluate the interobserver variability of lung ultrasound, a second ultrasound were included in the study. The characteristics of patients are summarised in Table 1 .

The median (range) age was 65 (37-90) years; the delay between the first COVID-19 symptoms and ultrasound assessment was 7 (1-16) days. Symptoms related to COVID-19 were fever (n=15, 71%), dyspnoea (n=12, 57%), dry cough (n=12, 57%), and diarrhoea (n=3, 14%). Median CRP level was 71 (3-300) mg/l. The minimal median oxygen saturation at ambient air was 93% (75-100) and respiratory frequency 22 (14-28) cycles/min. For patients requiring oxygen supplementation, the median oxygen requirement was 2 (0-15) l/min. Lung imaging assessment (Table 2) For chest HRCT, the classification of lung involvement for COVID-19 was as follows: absent or limited (n=2, 9.5%), moderate (n=10, 48%), extensive (n=5, 24%), severe (n=4, 19%) and critical (none). The median percentage of lung involvement was 20% (0-70).

For lung ultrasound assessment, 19 (90.5%) and 13 (61.9%) patients had at least one area with B-lines and consolidations, respectively. The median B-score and Cscore was 6 (0-11) and 1 (0-4), respectively. 

Lung ultrasound and correlation with chest HRCT and clinical variables (Fig. 2) On comparing ultrasound and chest HRCT findings, excellent correlation was found between the ultrasound score for B-lines and classification (p<0.01) and percentage lung involvement on chest HRCT (r=0.935, p<0.001; Fig. 2 

A recent study suggested that agreement between lung ultrasound and chest HRCT findings might vary depending on the severity of lung involvement, with diagnostic accuracy from 77% (mild involvement) to 93% (severe) (19) . The prevalence of Blines (90%) in the present COVID-19 patients agrees with that in previous case studies (18) (19) (20) (21) . In those studies, ultrasound analysis of 12 to 22 COVID-19 patients revealed a prevalence of B-lines ranging from 90% to 100% and the proportion of consolidations ranged from 20% to 25%. A larger proportion (>60%) with consolidations was observed in the present study. This discrepancy could be explained by distinct populations or by the fact that the number of consolidations observed by ultrasound is low (median=1), thus leading to increased difficulty to find them. The ability of ultrasound to detect COVID-19-related lung damage could be explained by a specific distribution of mainly peripheral and posterior lung lesions (18) .

Several arguments strengthen the relevance of lung ultrasound for COVID-19 evaluation. First, chest ultrasound can be easily performed at the bedside (26), requiring a mean time for global assessment of 15 minutes (data not shown).

Second, lung ultrasound does not require mobilising radiologist specialist staff and is less time-consuming than chest HRCT. Third, in the context of the pandemic and highly contagious patients, which limits wide access to chest HRCT (5), lung ultrasound could be an acceptable alternative for detecting COVID-19-related lung involvement. Fourth, ultrasound assessment of the lung could be a good alternative in countries with limited access to chest HRCT.

Regarding ultrasound detection of consolidations, the reproducibility was moderate, and ultrasound consolidations were not correlated with any oxygen parameters. Only the percentage of lung involvement on HRCT was slightly correlated with the J o u r n a l P r e -p r o o f consolidation score. Moreover, the interobserver reliability for ultrasound consolidation was moderate. This low reproducibility could be explained by a low number of consolidations observed by ultrasound, thus leading to increased difficulty to find them. These data suggest that using ultrasound to search for consolidations has limited use in clinical practice.

The lung ultrasound B-lines score correlated with oxygen parameters (Fig. 2) . The present prospective study suggests that ultrasound could help estimate the severity of COVID-19. Thus, ultrasound could help clinicians in emergency units determine the prognostic factors for patients at risk of rapid decrease in respiratory function; however, this impact on prognosis should be investigated in future studies.

The present study has some limitations. First, a relatively small number of patients were analysed, and only half were assessed by the two ultrasound operators.

Regarding the low number of patients, other studies analysed the same range of patients. Despite the relatively low number of patients, the results were clearly significant, which suggests the robustness of the present findings. Concerning ultrasound assessment by both operators, the second examination by the trainee operator was performed only to evaluate interobserver reliability. When comparing lung ultrasound assessment between experienced and trainee ultrasound operators, the interobserver reproducibility was substantial and was slightly higher than that observed a previous study (kappa = 0.53) (19) . Another limitation was the use of a non-validated ultrasound score for lung assessment of COVID-19 patients. A simplified score for B-lines was used to maximise the feasibility in the context of the urgent need for alternative screening tools for COVID-19 pulmonary disease. A proposal for standardising the use of lung ultrasound was published after the the present study commenced (25) . Using the present score, good agreement between J o u r n a l P r e -p r o o f both imaging procedures emphasises the relevance of using lung ultrasound as an alternative rapid screening tool for COVID-19. Nonetheless, HRCT should be considered the reference standard for both the diagnosis and extent of SARS-CoV2 infection allowing for elimination of the differential diagnoses, such as pulmonary embolism, considering that thrombotic events are rapidly emerging as a key threat in COVID-19 (28) . Finally, ultrasound of the lung was easy to perform in non-ICU patients. Indeed, the patients did not have severe tachypnoea, pulmonary drains, or lobar collapse, which could limit the interpretation on ultrasound.

In conclusion, the present findings suggest that lung ultrasound and chest HRCT findings are well correlated in patients with COVID-19, and ultrasound could be helpful for screening patients with suspected COVID-19. 

World Health Association. WHO coronavirus disease dashboard Update of

CT imaging features of 2019 novel coronavirus (2019-nCoV)

Correlation of chest CT and RT-PCR testing in coronavirus disease 2019 (COVID-19) in China: a report of 1014 cases

Time course of lung changes on chest CT during recovery from 2019 novel coronavirus (COVID-19) pneumonia

Point-of-care lung ultrasound in patients with COVID-19-a narrative review

Application of lung ultrasound during the COVID-19 pandemic: a narrative review

A brief review of lung ultrasonography in COVID-19: is it useful?

Thoracic ultrasound and SARS-COVID-19: a pictorial essay

Thoracic ultrasonography: a narrative review

How I do it: lung ultrasound

Lung B-line artefacts and their use

Ultrasonography in lung pathologies: new perspectives

Determination of a potential quantitative measure of the state of the lung using lung ultrasound spectroscopy

Lung ultrasound for critically ill patients

Bedside lung ultrasound in the assessment of alveolar-interstitial syndrome

CT in detecting chest wall invasion by tumor: a prospective study

COVID-19 pneumonia manifestations at the admission on chest ultrasound, radiographs, and CT: single-center study and comprehensive radiologic literature review

A clinical study of noninvasive assessment of lung lesions in patients with coronavirus disease-19 (COVID-19) by bedside ultrasound

Can lung ultrasound help critical care clinicians in the early diagnosis of novel coronavirus (COVID-19) pneumonia?

Lung ultrasound findings in patients with COVID-19 pneumonia

Point-of-care lung ultrasound findings in novel coronavirus disease-19 pnemoniae: a case report and potential applications during COVID-19 outbreak

Lung ultrasound findings in a 64-year-old

COVID-19 patients and the radiology department -advice from the European Society of Radiology (ESR) and the

European Society of Thoracic Imaging (ESTI)

Proposal for international standardization of the use of lung ultrasound for patients with COVID-19: a simple, quantitative, reproducible method

Bedside ultrasound assessment of positive end-expiratory pressure-induced lung recruitment

Radiographic and CT features of viral pneumonia

Incidence of thrombotic complications in critically ill ICU patients with COVID-19

Ultrasonography of the lung in axial view (linear probe) showing the normal artefact A-lines (white dashed lines) and consolidation (white arrow). (c,d) Ultrasound of the lungs with convex probe representing multiple (c) and coalescent (d) B-lines

Author contributions

Guarantor of integrity of the entire study: Ottaviani 2. Study concepts and design: Ottaviani, Franc, Dieudé 3. Literature research: Ottaviani, Debray, Khalil, Crestani, Borie 4. Clinical studies: Ottaviani, Franc, Borie, Crestani, Ebstein, Demaria, Lheure 5. Experimental studies / data analysis: Ottaviani 6. Statistical analysis: Ottaviani 7