ABSTRACT
Thoracic radiography (TR), the most common screening test for pulmonary metastases in dogs, can fail to detect small lesions <3 mm. Lung ultrasonography (LUS) is a widely available imaging modality capable of detecting peripheral nodules but is underutilized for this purpose. Thoracic computed tomography (CT) is the criterion standard for diagnosis of lung metastases and nodular disease but is less practical for a variety of reasons. We hypothesized that LUS would be more sensitive but less specific at detecting nodules consistent with metastatic pulmonary disease in dogs compared to TR, using CT as the criterion standard. This was a masked, single-center prospective study of 62 client-owned dogs evaluated for respiratory signs or pulmonary metastatic neoplasia screening using TR, LUS and CT. Dogs were included if metastatic pulmonary disease was a differential. All imaging modalities were scored as having nodules (yes/no) and other types of pathologic lesions were recorded. Sensitivity (Se), specificity (Sp) and positive (LR+) and negative likelihood ratios (LR-) were determined for TR and LUS. For TR, Se and Sp were 64% and 73%, and LR+ and LR- were 2.37 and 0.49, respectively. For LUS, Se and Sp were 60% and 65% and LR+ and LR- were 1.71 and 0.62, respectively. The results of the study indicate that LUS had a similar Se to TR, with both modalities missing nodules when used for screening. The low Sp and LR- suggests caution should be used when assuming TR and LUS rule out the presence of nodules.
