Distribution of alveolar‐interstitial syndrome in dogs and cats with respiratory distress as assessed by lung ultrasound versus thoracic radiographs

OBJECTIVE: To assess distribution of alveolar‐interstitial syndrome (AIS) detected by lung ultrasound (LUS) compared to thoracic radiographs (TXR). DESIGN: Prospective study. SETTING: University teaching hospital. ANIMALS: Seventy‐six dogs and 24 cats with acute respiratory distress or tachypnea. INTERVENTIONS: Patients underwent LUS and TXR within 6 hours. Lung ultrasound images were scored for presence and quantity of B‐lines in 4 lung quadrants (right cranial, right caudal, left cranial, left caudal). An individual LUS quadrant was scored positive if > 3 B‐lines were observed within a single intercostal space. Dorsoventral TXR were scored for presence of AIS in the same 4 quadrants. An individual TXR quadrant was scored positive if infiltrate was present in ≥ 25% of the quadrant. Medical records were evaluated for final diagnosis. MEASUREMENTS AND MAIN RESULTS: Quadrant‐by‐quadrant spatial agreement in assigning AIS using LUS versus TXR was fair (K = 0.24 – 0.56). Lung ultrasound scored a higher number of quadrants positive per patient (2.65 ± 1.59 vs. 2.13 ± 1.48; P = 0.012). Patterns of distribution of AIS differed significantly based on final diagnosis. Patients with left‐sided congestive heart failure were more likely to have diffuse AIS on LUS (P < 0.001) or bilateral caudal AIS on TXR (P = 0.04) while patients with noncardiac disease were more likely to have absence of AIS in all quadrants using either modality (P < 0.001). Differences in spatial distribution of AIS were also noted among disease subcategories. CONCLUSIONS: Lung ultrasound and TXR were both useful to detect and categorize distribution of alveolar or interstitial pulmonary pathology. Spatial agreement between modalities was only fair. Overall, LUS detected a higher incidence of AIS compared to TXR. Both modalities detected differences in distribution of AIS based on final diagnosis, suggesting that a regional pattern‐based approach to thoracic imaging may prove diagnostically useful.