Impact of ultrasound settings on lung vertical artefacts: an observational study in mechanically ventilated patients

• Published in: ERJ open research 2025-01, Vol.11 (1), p.483
• Main Authors: Leote, João, Gonçalves, Andreia, Fonseca, Júlia, Loução, Ricardo, Dias, Hermínia, Inês Ribeiro, Maria, Meireles, Ricardo, Varudo, Rita, Bacariza, Jacobo, Gonzalez, Filipe
• Format: Article
• Language: English
• Subjects: Original s
• ISSN: 2312-0541; 2312-0541
• DOI: 10.1183/23120541.00483-2024

The number of vertical artefacts (VAs) in lung ultrasound (LUS) impacts patients' clinical management. This study aimed to demonstrate the influence of ultrasound settings on the number of VAs in patients under invasive mechanical ventilation (IMV). Patients under IMV were recruited for LUS, including three breathing cycles with a motionless curvilinear probe on the thoracic region with the most VAs. Three experts in LUS were asked about the number of VAs at random, and blinded after altering the settings for a total of 20 test recordings per patient. The correlation between expert classifications was tested after grading the classifications. The number of VAs across clinicians was compared between baseline recordings and test condition recordings to determine statistical differences. 29 patients were enrolled with a median Sequential Organ Failure Assessment score of 6 (interquartile range (IQR) 3). IMV was mainly due to stroke (n=10) and pneumonia (n=6). LUS was made between days 1 and 6 (IQR). Baseline recordings showed a median of 2±2 VAs in inspiration and a median of 1±2 in expiration from a total of 3636 expert classifications, with a strong agreement within patients. A probe frequency of 8 MHz, artefact filtering, speckle reduction and frame average reduced the median VA number by one. A power of -20 dB and dynamic range of 32 dB abolished the VAs. A gain above 90% increased the median number of VAs by one. In this study, the LUS settings influenced the VA number in IMV patients, after controlling for physiological and operator confounders.