Atilio Atencio, MS4
The Article: Benchoufi, et al. “Lung injury in patients with or suspected COVID-19 : a comparison between lung ultrasound and chest CT-scanner severity assessments, an observational study.” https://doi.org/10.1101/2020.04.24.20069633
The Idea: SARS-CoV2 is a virus that has caused the most recent pandemic commonly referred to as COVID-19. COVID-19 can vary in presentation and severity from patient to patient. Respiratory symptoms may quickly progress to acute respiratory distress syndrome (ARDS) making the efficient evaluation of lung status important for patient management and prognosis. CT is the gold standard for COVID-19 pneumonia testing, but the test is limited by scanner availability and radiation exposure. Using lung ultrasonography as an alternative to CT scans to assess and screen suspected COVID-19 patients would relieve some of the heavy patient flow bottlenecking at the CT scanner while being a convenient and efficient exam that can be interpreted real-time and at a fraction of the cost. This study compares the use of ultrasound versus CT scan to quickly assess lung damage severity in patients with suspected or diagnosed COVID-19 and to compare the performance of newly trained operator and an expert operator in terms of ultrasound assessment of pulmonary lesions.
The Study: This study was a multicenter, observational non-randomized study conducted in the emergency units of 3 hospitals of Assistance Publique – Hôpitaux de Paris (APHP) from 3/19/2020 – 4/1/2020.
Inclusion Criteria: patients age >18 years with suspected or diagnosed COVID-19 who underwent CT.
Exclusion Criteria: patients for whom the LU exploration could not be performed (morbid obesity, extensive thoracic subcutaneous emphysema, absorbent subcutaneous infiltrations) or with any comorbidity that justified priority immediate intensive care.
There were 107 patients included in the study, each of which underwent a lung ultrasound (LU), a clinical assessment by an emergency physician, and a CT. The emergency physician and ultrasound operator were blinded to each other’s study. 90 patients (84.1%) had their CT quantified by an ordinal scale. The remaining patients had information missing in the radiology report and were not included.
48 patients underwent lung ultrasound by both and expert and a newly trained physician. The physicians newly trained in LU underwent a 30-min training protocol before being able to explore normal lungs and to recognize pathologies on ultrasound images form an image bank.
The ultrasound score for assessing lung condition was derived from the standard LU score which was simplified to involve 8 zones instead of 12. The scoring consisted of assigning a numerical value to the lesions from 0 to 3 that would determine the severity global score (GS) ranging from 0 to 24. The 8 zones were: right antero-superior, left antero-superior, right antero-inferior, left antero-inferior, right postero-superior, left postero-superior, right postero-inferior, and left postero-superior. The study also defined 4 grades of severity on LU from 0 – 3: 0, up to a maximum of 3 observed B-lines; 1, 4 to 8 B-lines through intercostal space at one of the pulmonary bases; 2, B-lines in “curtain sign” (> 8 B-lines) and/or diffusion of more than 4 B-lines in two-thirds of the pulmonary field; 3, consolidation foci.
CT results were extracted form the radiologists’ report by determining if the lung injuries were typical or not of SARS-CoV2 infection and defining the severity of lung injury on a scale from minimal to critical. The criteria for determining lung injury: up to 10% involved of pulmonary parenchyma corresponds to minimal, 10%-25% to moderate, 25%-50% to extended, 50%-75% to severe, and >75% to critical.
Results: The mean severity score for all pulmonary quadrants (maximum score of 24) was 9.6 ± 6.0 with the most severely affected lung quadrant being the right posterior-inferior. Chest CT examinations were pathological for 101 (94.4%) patients and considered typical of COVID-19 for 86 (85.1%).
The study found moderate agreement between LU when considered as a 4-category ordinal scale and the CT scale (weighted kappa of 0.52). It also showed that GS had good performance to predict evaluation of the disease by CT classified normal versus pathologic (AUC 0.93 and Brier score of 0.04).
The study found good agreement between GS evaluated by a new trainee and expert (weighted kappa of 0.85). Including when considering each quadrant individually (weighted kappa 0.62 - 0.81).
The Takeaway: The study supports two findings. One, that LU can be used to assess the lung injuries of patients with suspected or diagnosed COVID-19 as it is consistent with CT findings. And two, that the findings of a newly trained physicians on ultrasound and an expert on ultrasound were not considerably different, suggesting you do not need to be an ultrasound expert to use LU to assess COVID-19 lung injury. While CT will continue to be the gold standard for COVID-19 due to its superior specificity, LU offers a bedside, low-cost, real-time, and non-invasive exam with similar sensitivity. LU is therefore a useful tool when assessing and triaging COVID-19 patients, especially in the setting of limited CT availability.