Theodore J Dubinsky1, Hardik Shah2, Rachelle Sonneborn2, Daniel S Hippe2. 1. Department of Radiology, University of Washington, Seattle, WA. Electronic address: tdubinsky@comcast.net. 2. Department of Radiology, University of Washington, Seattle, WA.
Abstract
BACKGROUND: We prospectively identified B-lines in patients undergoing ultrasonographic (US) examinations following liver transplantation who also had chest radiography (CXR) or chest CT imaging, or both, on the same day to determine if an association between the presence of B-lines from the thorax on US images correlates with the presence of lung abnormalities on CXR. METHODS: Following institutional review board (IRB) approval, patients who received liver transplants and underwent routine US examinations and chest radiography or CT imaging, or both, on the same day between January 1, 2015 through July 1, 2016 were prospectively identified. Two readers who were blinded to chest films and CT images and reports independently reviewed the US interreader agreement for the presence or absence of B-lines and performed an evaluation for the presence or absence of diffuse parenchymal lung disease (DPLD) on chest films and CT images as well as from clinical evaluation. Receiver operating characteristic (ROC) curves were constructed. RESULTS: There was good agreement between the two readers on the presence of absence of B-lines (kappa = 0.94). The area under the ROC curve for discriminating between positive DPLD and negative DPLD for both readers was 0.79 (95% CI, 0.71-0.87). CONCLUSIONS: There is an association between the presence of extensive B-lines to the point of confluence and "dirty shadowing" on US examinations of the chest and associated findings on chest radiographs and CT scans of DPLD. Conversely, isolated B-lines do not always correlate with abnormalities on chest films and in fact sometimes appear to be a normal variant.
BACKGROUND: We prospectively identified B-lines in patients undergoing ultrasonographic (US) examinations following liver transplantation who also had chest radiography (CXR) or chest CT imaging, or both, on the same day to determine if an association between the presence of B-lines from the thorax on US images correlates with the presence of lung abnormalities on CXR. METHODS: Following institutional review board (IRB) approval, patients who received liver transplants and underwent routine US examinations and chest radiography or CT imaging, or both, on the same day between January 1, 2015 through July 1, 2016 were prospectively identified. Two readers who were blinded to chest films and CT images and reports independently reviewed the US interreader agreement for the presence or absence of B-lines and performed an evaluation for the presence or absence of diffuse parenchymal lung disease (DPLD) on chest films and CT images as well as from clinical evaluation. Receiver operating characteristic (ROC) curves were constructed. RESULTS: There was good agreement between the two readers on the presence of absence of B-lines (kappa = 0.94). The area under the ROC curve for discriminating between positive DPLD and negative DPLD for both readers was 0.79 (95% CI, 0.71-0.87). CONCLUSIONS: There is an association between the presence of extensive B-lines to the point of confluence and "dirty shadowing" on US examinations of the chest and associated findings on chest radiographs and CT scans of DPLD. Conversely, isolated B-lines do not always correlate with abnormalities on chest films and in fact sometimes appear to be a normal variant.