Lung Ultrasound and High-Resolution Computed Tomography in Suspected COVID-19 Patients Admitted to the Emergency Department: A Comparison

Abstract

Objective: To analyze the diagnostic accuracy of lung ultrasonography (LUS) and high-resolution computed tomography (HRCT), to detect COVID-19.

Materials and methods: This study recruited all patients admitted to the emergency medicine unit, due to a suspected COVID-19 infection, during the first wave of the COVID-19 pandemic. These patients also who underwent a standardized LUS examination and a chest HRCT. The signs detected by both LUS and HRCT were reported, as well as the sensitivity, specificity, positive predictive value, and negative predictive value for LUS and HRCT.

Results: This cohort included 159 patients, 101 (63%) were diagnosed with COVID-19. COVID-19 patients showed more often confluent subpleural consolidations and parenchymal consolidations in lower lung regions of LUS. They also had "ground glass" opacities and "crazy paving" on HRCT, while pleural effusion and pulmonary consolidations were more common in non-COVID-19 patients. LUS had a sensitivity of 0.97 (95% CI 0.92-0.99) and a specificity of 0.24 (95% CI 0.07-0.5) for COVID-19 lung infections. HRCT abnormalities resulted in a 0.98 sensitivity (95% CI 0.92-0.99) and 0.1 specificity (95% CI 0.04-0.23) for COVID-19 lung infections.

Conclusion: In this cohort, LUS proved to be a noninvasive, diagnostic tool with high sensitivity for lung abnormalities that were likewise detected by HRCT. Furthermore, LUS, despite its lower specificity, has a high sensitivity for COVID-19, which could prove to be as effective as HRCT in excluding a COVID-19 lung infection.

Keywords: COVID-19; lung diseases; lung infiltrates; lung ultrasonography.

Figure 1.

A patient with moderate disease and lung involvement (mainly on the left side) and LUS performed about 5 hours after the HRCT. (A) Demonstrates the lung area designated as L1, showing A-lines with one isolated B-Line. Below the plane of the ribs, a hyper echogenic, regular pleural line is noted to be generating one horizontal reverberation artifact. It is parallel to the pleural line and the so-called A-lines, are visible in depth until the power of the ultrasound beam is exhausted. This is an area with preserved lung ventilation, also referred to “spared area” in a pathological setting. An isolated B-line is visible to the right of the image. (B) Demonstrates the lung areas designated as R3–R4. Non-confluent B-lines are seen extending from the pleural line into the depths obliterating the A-lines on the right of the field of view. (C) A HRCT image is provided, and GGOs are observed in the lateral areas of both lungs. This corresponds to L3–L4 and R3–R4 lung areas on LUS examination. (D and E) Demonstrating the lung area designated L5–L6. The pleural line is seen to be grossly irregular and thickened and appears to overlie several small subpleural thickenings. At the time of examination, these did not appear to converge but were inscribed in areas of high B-line condensation. (F) A HRCT image is provided and, in the posterior, lateral inferior area of the left lung, two small subpleural thickenings are noted which, corresponded to L6 lung are of the LUS examination. GGOs, ground glass opacities; HRCT, high-resolution computed tomogram; LUS, lung ultrasonogram.

Figure 2.

An example lung ultrasonogram that demonstrates non-confluent B-lines. There are faint vertical artifacts persisting up to the full depth of the field of view, but these do not completely erase the normal A-lines.

Figure 3.

(A) An example lung ultrasonogram that shows one intercostal space that is entirely occupied by multiple confluent B-lines, and this completely erases the normal A-lines. (B) This lung ultrasonogram also demonstrates a normal or “spared area” between two intercostal spaces with confluent B-line artifacts.

Figure 4.

An example lung ultrasonogram that shows an isolated subpleural consolidation. The one consolidation is less than 2 centimeters, with a posterior vertical artifact reaching the bottom of the image without fading. It was noted to dynamically move in synchronicity with lung sliding, in the context of an A-pattern artifact.

Figure 5.

An example lung ultrasonogram that demonstrated a confluent subpleural consolidations (CSpCs). These multiple consolidations, within the lung, were less than 2 centimeters in depth with a posterior vertical artifact, which completely obliterated the underlying lung parenchyma.

Figure 6.

An example lung ultrasonogram that demonstrated a subpleural consolidation of less than 2 centimeters.

Figure 7.

This patient case had to be excluded from the study because of a 48-hour interval time between the HRCT and the LUS. (A) Regardless of excluding this patient, the HRCT is provided. On the HRCT image, at the level of the bifurcation of the trachea, any parenchymal alterations are absent in the left posterior superior area. On this same image, GGOs and normal or “spared areas” are seen in the right posterior and lateral lung areas, as well as in the left lateral lung area. (B) A LUS was performed on this patient 48 hours after the HRCT. On the LUS image, A-lines are seen without other artifacts at L5 (corresponding to the posterior superior area), but multiple B-lines are noted and are otherwise aggregated at the L3 designation. (C) The same LUS image demonstrates a thickening greater than 2 centimeters, at the L4 designation, which corresponds to the lateral lung area on the HRCT. This could represent the evolution of the GGOs artifacts with that lung consolidation. GGOs, ground glass opacities; HRCT, high-resolution computed tomogram; LUS, lung ultrasonogram.