Sixty-Four-Section Cardiac Computed Tomography in Mechanical Prosthetic Heart Valve Dysfunction: Thrombus or Pannus.

Abstract

Distinguishing pannus and thrombus in patients with prosthetic valve dysfunction is essential for the selection of proper treatment. We have investigated the utility of 64-slice multidetector computed tomography (MDCT) in distinguishing between pannus and thrombus, the latter amenable to thrombolysis. Sixty-two (23 men, mean age 44±14 years) patients with suspected mechanical prosthetic valve dysfunction assessed by transesophageal echocardiography were included in this prospective observational trial. Subsequently, MDCT was performed before any treatment was started. Periprosthetic masses were detected by MDCT in 46 patients, and their attenuation values were measured as Hounsfield Units (HU). Patients underwent thrombolysis unless contraindicated, and those with a contraindication or failed thrombolysis underwent surgery. A mass which was completely lysed or surgically detected as a clot was classified as thrombus, whereas a mass which was surgically detected as tissue overgrowth was classified as pannus. A definitive diagnosis could be achieved in 37 patients with 39 MDCT masses (22 thrombus and 17 pannus). The mean attenuation value of 22 thrombotic masses was significantly lower than that in 17 pannus (87±59 versus 322±122; P<0.001). Area under the receiver operating characteristic curve was 0.96 (95% confidence interval: 0.91-0.99; P<0.001), and a cutoff point of HU≥145 provided high sensitivity (87.5%) and specificity (95.5%) in discriminating pannus from thrombus. Complete lysis was more common for masses with HU<90 compared with those with HU 90 to 145 (100% versus 42.1%; P=0.007). Sixty-four slice MDCT is helpful in identifying masses amenable to thrombolysis in patients with prosthetic valve dysfunction. A high (HU≥145) attenuation suggests pannus overgrowth, whereas a lower value is associated with thrombus formation. A higher attenuation (HU>90) is associated with reduced lysis rates.