SESSION TITLE: Pleural Disease
SESSION TYPE: Original Investigation Slide
PRESENTED ON: Wednesday, October 29, 2014 at 02:45 PM - 04:15 PM
PURPOSE: All patients hospitalized with pneumonia undergo radiographic imaging to assess for suspected complicated parapneumonic effusions, but there is limited evidence as to the best imaging modality to characterize pleural space infections. It is common practice to obtain chest computed tomography (CT). In recent years, chest ultrasound has been introduced as an alternative to assist in identifying the complexity of these effusions. We studied and compared the ability of chest CT versus chest ultrasound to determine the complexity of parapneumonic effusions.
METHODS: 47 patients with a parapneumonic pleural effusion who underwent thoracentesis were retrospectively reviewed. We compared the pre-procedure chest ultrasound determination of the complexity with the pre-procedure chest CT determination of the complexity. A dedicated thoracic radiologist reviewed all chest CT images. Pleural fluid analysis provided the gold standard as to whether the pleural space was complicated or not.
RESULTS: Chest CT had a sensitivity of 84%, specificity of 68%, positive predictive value (PPV) of 75% and negative predictive value (NPV) of 78%. Chest ultrasound had a sensitivity of 70%, specificity of 85%, PPV of 95% and NPV of 68%.
CONCLUSIONS: Chest CT can evaluate complex pleural abnormalities and confirm the diagnosis of complicated paraneumonic effusions, but exposes the patient to additional radiation, inflates hospital costs and requires more time. Chest ultrasound, which is fast, safe, portable and readily accessible, can correctly exclude the absence of complex pleural abnormalities and its greater PPV suggests higher accuracy in identifying them.
CLINICAL IMPLICATIONS: Chest ultrasound is comparable and can be an alternative to chest CT when evaluating complex parapneumonic effusions. In the absence of chest CT, it can determine the need for chest tube insertion.
DISCLOSURE: The following authors have nothing to disclose: Philip Svigals, John Huggins, Carlos Kummerfeldt, Jennings Nestor, James Ravenel, Peter Doelken
No Product/Research Disclosure Information