Lung Ultrasound Prediction Model for Acute Respiratory Distress Syndrome: A Multicenter Prospective Observational Study

Abstract

Rationale: Lung ultrasound (LUS) is a promising tool for diagnosis of acute respiratory distress syndrome (ARDS), but adequately sized studies with external validation are lacking. Objectives: To develop and validate a data-driven LUS score for diagnosis of ARDS and compare its performance with that of chest radiography (CXR). Methods: This multicenter prospective observational study included invasively ventilated ICU patients who were divided into a derivation cohort and a validation cohort. Three raters scored ARDS according to the Berlin criteria, resulting in a classification of "certain no ARDS," or "certain ARDS" when experts agreed or "uncertain ARDS" when evaluations conflicted. Uncertain cases were classified in a consensus meeting. Results of a 12-region LUS exam were used in a logistic regression model to develop the LUS-ARDS score. Measurements and Main Results: Three hundred twenty-four (16% certain ARDS) and 129 (34% certain ARDS) patients were included in the derivation cohort and the validation cohort, respectively. With an ARDS diagnosis by the expert panel as the reference test, the LUS-ARDS score, including the left and right LUS aeration scores and anterolateral pleural line abnormalities, had an area under the receiver operating characteristic (ROC) curve of 0.90 (95% confidence interval [CI], 0.85-0.95) in certain patients of the derivation cohort and 0.80 (95% CI, 0.72-0.87) in all patients of the validation cohort. Within patients who had imaging-gold standard chest computed tomography available, diagnostic accuracy of eight independent CXR readers followed the ROC curve of the LUS-ARDS score. Conclusions: The LUS-ARDS score can be used to accurately diagnose ARDS also after external validation. The LUS-ARDS score may be a useful adjunct to a diagnosis of ARDS after further validation, as it showed performance comparable with that of the current practice with experienced CXR readers but more objectifiable diagnostic accuracy at each cutoff.

Keywords: ARDS; LUS; diagnosis; intensive care.

Figure 1.

Study profile. A total of 519 patients were included in the Diagnosis of Acute Respiratory disTress Syndrome (DARTS) study. Sixty-six included patients who could not be analyzed in the present study because they did not receive a lung ultrasound (LUS) exam or because more than 4 of 12 lung regions were missing in the LUS exam. Patients included in the Amsterdam University Medical Centers, location Academic Medical Center, served as the derivation cohort, and patients included in the Maastricht University Medical Center+ were assigned to the validation cohort. ARDS = acute respiratory distress syndrome; MV = mechanical ventilation.

Figure 2.

Distribution of lung ultrasound (LUS) patterns for different lung regions in patients from the derivation cohort (n = 324) and the validation cohort (n = 129). Patients underwent a LUS exam with six regions scanned per hemithorax. Patterns found in these regions for both lungs combined are presented in this figure and are stratified for ARDS category. ARDS = acute respiratory distress syndrome; NA = not available; UTS = unable to score.

Figure 3.

Diagnostic accuracy of 1) the lung ultrasound-acute respiratory distress syndrome (LUS-ARDS) score, 2), the current practice with single-rater assessment of the Berlin definition using chest radiography (CXR) and computed tomography (CT), and 3) two previously published LUS methods for diagnosis of ARDS. Diagnostic accuracy was assessed in patients from both the derivation and validation cohorts who had a CT scan available, as CT is considered the most reliable imaging modality (n = 229). ARDS diagnosis by the expert panel was used as the reference standard. The area under the receiver operating characteristic curve (AUROCC) of the LUS-ARDS score in this group was 0.84; 95% confidence interval (CI), 0.79–0.89 (red). The three raters of the expert panel scored the Berlin definition using CT scans in all 229 patients with a CT (green). For the purpose of this study, and using CXR as the imaging modality, 1) three raters of the expert panel scored the presence of ARDS in all 229 patients with a CT scan, 2) seven independent ICU physicians scored the presence of ARDS in the same subset of 50 patients with a CT scan, and 3) one chest radiologist scored the presence of ARDS in a subset of 121 patients with a CT scan from the derivation cohort. These assessments resulted in a total of 11 CXR raters (blue). The expert panel and independent raters had clinical and ventilation data available next to the images. We also assessed the diagnostic accuracy of two previous definitions of ARDS that included LUS criteria: the Kigali definition by Riviello and colleagues (10) (yellow) and the Kigali definition with stricter criteria by Vercesi and colleagues (11) (orange). The LUS criteria of these methods were compared with ARDS diagnosis by the expert panel in all 229 patients who had a CT scan available. The Kigali definition requires at least one region with a B- or C-pattern bilaterally. The Kigali definition with stricter criteria requires at least two B2- or C-patterns or three B1-patterns bilaterally. Lower CIs are presented for each rater and LUS method to assess potential overlap with the CI of the LUS-ARDS score. *P < 0.05, indicating significant differences between the AUROCC of a rater/method and the AUROCC of the LUS-ARDS score.