Diagnostic performance of fluoroscopic video analysis for pulmonary embolism: a prospective observational study.

Abstract

Dynamic chest radiography using X-ray fluoroscopic video analysis has shown potential for the diagnosis of pulmonary embolism (PE), but its diagnostic performance remains uncertain. We aimed to evaluate the diagnostic performance of fluoroscopic video analysis for diagnosing PE. A prospective single-center observational study was conducted between October 2020 and January 2022. Fifty consecutive adult patients, comprising definitive PE, pulmonary hypertension (PH), or suspected PH, were enrolled. The study population was classified into 23 PE and 27 non-PE cases by contrast-enhanced computed tomography, lung scintigraphy, right heart catheterization, and pulmonary angiography. Cineradiographic images of 10-second breath-holds were obtained and analyzed using a fluoroscopic video analysis workstation to generate pulmonary circulation images. Two blinded cardiologists qualitatively assessed the presence or absence of perfusion defects on the pulmonary circulation images. The diagnosis obtained from the fluoroscopic analysis was compared with the definitive diagnosis. The primary outcomes included sensitivity, specificity, positive and negative predictive values, and overall accuracy for diagnosing PE. Perfusion defects were observed in 21 of 23 PE patients and 13 of 27 non-PE patients. The diagnostic performance of fluoroscopic video analysis for diagnosing PE showed a sensitivity of 91%, specificity of 52%, positive predictive value of 62%, negative predictive value of 88%, and overall accuracy of 70%. The high sensitivity of the fluoroscopic video analysis suggests its potential usefulness in ruling out PE without the need for contrast media or radionuclide; however, its specificity and overall accuracy remain limited.