UTILIZING ULTRASOUND TO PREDICT PLEURAL EFFUSION ETIOLOGY

Abstract

SESSION TITLE: Imaging Posters SESSION TYPE: Original Investigation Posters PRESENTED ON: October 18-21, 2020 PURPOSE: Our primary objectives were to associate specific features on ultrasound imaging of pleural effusions with fluid etiology and to determine the diagnostic accuracy of these sonographic features in predicting these etiologies. Specifically, we aimed to associate the following findings on pleural fluid ultrasound imaging with their potential etiologies: anechoic fluid, fibrin strands, septation, loculations, and plankton sign (i.e., swirling echoic particles that move with cardiac and pulmonary motion). METHODS: We conducted a retrospective analysis, from July 2015 to November 2019, of all ultrasound studies of pleural effusions captured in the medical intensive care unit of a tertiary care center. Each study was cross-referenced with its respective patient’s medical record to determine whether a thoracentesis procedure and cytological fluid analysis was conducted. Pleural fluid analyses were adjudicated as transudative or exudative, on the basis of Light’s criteria. Light’s criteria outcomes were then correlated with specific sonographic features, including anechoic fluid, fibrin strands, septation, loculations, and plankton sign. Diagnostic test characteristics (sensitivity, specificity, positive predictive value [PPV], and negative predictive value [NPV]) were calculated for each sonographic feature along with their respective 95% confidence intervals. RESULTS: Of 832 patients who underwent ultrasound examinations of pleural effusions during our study period, 184 patients had a thoracentesis procedure performed upon which fluid analysis was conducted and effusion type was determined. Effusions were exudative in 115 (62.5%) and transudative in 69 (37.5%) patients. In identifying exudative effusions, fibrin strands, plankton sign, septation, and loculations all had a 97% specificity and 86% PPV or greater. The sensitivity and NPV of anechoic fluid for identifying transudative effusions were 94.1% and 92.1%, respectively. The specificity and NPV of plankton sign for identifying cytologypositive malignancy was 97.3% and 88.4%, respectively. Combining ultrasound features did not lead to an increase in diagnostic test characteristics for determining effusion type. CONCLUSIONS: Sonographic features such as fibrin strands, plankton sign, septation, and loculations are highly specific for identifying exudative pleural effusions. Plankton sign is highly specific for identifying cytology-positive malignancy. CLINICAL IMPLICATIONS: Identification of specified sonographic features can guide medical and procedural management of pleural effusions by providing important diagnostic information in a cost-effective and non-invasive manner. DISCLOSURES: No relevant relationships by Verdah Ahmad, source=Web Response No relevant relationships by Sahar Ahmad, source=Web Response No relevant relationships by Jason Chua, source=Web Response No relevant relationships by Anthony Lisa, source=Web Response No relevant relationships by Kinner Patel, source=Web Response