Diagnostic accuracy of real-time ultrasound for detecting pulmonary edema by lung region

Lung ultrasound is a useful tool for rapidly diagnosing pulmonary edema at the bedside, with data suggesting improved accuracy compared with chest radiographs [ ]. However, there have been variations in techniques used, ranging from a simple two-zone technique to assessment of individual rib spaces across the entire lung field. While previous ultrasound studies for pulmonary edema have focused on different techniques [ ], there are limited data on which regions are most accurate. This study aims to evaluate the diagnostic accuracy of individual lung zones included in the 12-zone lung ultrasound technique.

This was a prospective, observational study evaluating the diagnostic accuracy of specific regions on lung ultrasound for identifying pulmonary edema using real-time assessment. Adult (age ≥ 18 years) Emergency Department patients with suspected cardiogenic pulmonary edema were identified by research assistants using a convenience sample of when an ultrasound fellow was present. Patients were eligible if they were willing to participate in the study, spoke English as a primary language, and there was a clinical concern for pulmonary edema. Patients were excluded if they had symptomatic COVID-19 infection, interstitial lung disease, the treating clinician was not concerned for pulmonary edema, they declined to participate, were clinically unstable, or had been previously enrolled in the study. All exams and real-time interpretations were performed by a single ultrasound fellow who had completed an emergency medicine residency, followed by a one-hour study-specific training session and 25 proctored lung ultrasound examinations [ ].

Using the C1–5 transducer (GE Venue; GE Healthcare) in the lung preset at a depth of 18 cm in the sagittal plane, a single ultrasound fellow performed thoracic ultrasound with the 12-zone technique. This included the following views: Left-Anterior-Inferior (LAI), Left-Anterior-Superior (LAS), Left-Lateral-Inferior (LLI), Left-Lateral-Superior (LLS), Left-Posterior-Inferior (LPI), Left-Posterior-Superior (LPS), Right-Anterior-Inferior (RAI), Right-Anterior-Superior (RAS), Right-Lateral-Inferior (RLI), Right-Lateral-Superior (RLS), Right-Posterior-Inferior (RPI), and Right-Posterior-Superior (RPS). The fellow recorded a six-second video clip in each region and reported a real-time assessment of whether the lung field was positive (≥3 B lines or a large dense B line) or negative (<2 B lines and the absence of a large, dense B line) for pulmonary edema. Lung regions were excluded if they were unable to be adequately visualized (e.g., cardiac interference). The clips were then reviewed independently by two expert sonographers (ultrasound leaders with >10,000 prior ultrasound image reviews) who were blinded to the initial determination by the fellow. The experts reviewed all discordant values and reached consensus on whether the field was positive or negative using the same criteria as defined above. We calculated the overall accuracy, sensitivity, specificity, positive likelihood ratio (LR+), and negative likelihood ratio (LR-) with 95% confidence intervals (CI) for all lung zones with at least 10 acceptable images.

Seventy-one patients ( n = 752 lung fields) were included in the study. We evaluated all lung zones except the LAI (only 3 sufficient images available). Overall, lung ultrasound was 96.7% (95% CI 93.8% to 98.5%) sensitive and 79.0% (95% CI 75.1% to 82.6%) specific. The most accurate regions were RAS, RPS, and LPS, while the least accurate regions were LPI and LAS ( Table 1 ).

Table 1

Diagnostic accuracy by lung region.

Zone	Accuracy	Sensitivity	Specificity	LR+	LR-
LAS	80.6% (68.6% to 89.6%)	100.0% (79.4% to 100%)	73.9% (58.9% to 85.7)	3.83 (2.36 to 6.23)	0
LLI	84.1% (73.3% to 91.8%)	94.3% (80.8% to 99.3%)	73.5% (55.6% to 87.1%)	3.56 (2.02 to 6.27)	0.08 (0.02 to 0.30)
LLS	84.5% (74.0% to 92.0%)	96.3% (81.0% to 99.9%)	77.3% (622% to 88.5%)	4.24 (2.44 to 7.34)	0.05 (0.01 to 0.33)
LPI	78.9% (67.6% to 87.7%)	96.2% (80.4% to 99.9%)	68.9% (53.4% to 81.8%)	3.09 (1.99 to 4.81)	0.06 (0.01 to 0.39)
LPS	88.6% (78.7% to 94.9%)	94.1% (71.3% to 99.9%)	86.8% (74.7% to 94.5%)	7.13 (3.54 to 14.35)	0.07 (0.01 to 0.46)
RAI	84.6% (73.5% to 92.4%)	94.7% (74.0% to 99.9%)	80.4% (66.1% to 90.6%)	4.84 (2.67 to 8.78)	0.07 (0.01 to 0.44)
RAS	91.2% (81.8% to 96.7%)	100.0% (85.2% to 100%)	86.7% (73.2% to 94.9%)	7.50 (3.56 to 15.80)	0
RLI	86.6% (76.0% to 93.7%)	100.0% (88.1% to 100%)	76.3% (59.8% to 88.6%)	4.22 (2.39 to 7.47)	0
RLS	87.0% (76.7% to 93.9%)	100.0% (82.4% to 100%)	82.0% (68.6% to 91.4%)	5.56 (3.07 to 10.04)	0
RPI	85.3% (74.6% to 92.7%)	93.9% (79.8% to 99.3%)	77.1% (59.9% to 89.6%)	4.11 (2.22 to 7.60)	0.08 (0.02 to 0.30)
RPS	89.9% (80.2% to 95.8%)	96.2% (80.4% to 99.9%)	86.0% (72.1% to 94.7%)	6.89 (3.27 to 14.53)	0.04 (0.01 to 0.31)
OVERALL	85.4% (82.6% to 87.8%)	96.7% (93.8% to 98.5%)	79.0% (75.1% to 82.6%)	4.61 (3.87 to 5.49)	0.04 (0.02 to 0.08)