Detection of construction errors in ex vivo coronary artery anastomoses by 13-MHz epicardial ultrasonography

Abstract

Intraoperative detection of suboptimal coronary anastomoses allows revision before chest closure. We evaluated an epicardial 13-MHz ultrasound minitransducer as a means to detect three different coronary anastomosis construction errors. In total, 120 internal thoracic artery-to-coronary artery anastomoses were constructed correctly (n = 60) or incorrectly (n = 60) with one technical error: suture crossover, purse-string or deep toe stitch (n = 20 each). Anastomoses were performed on ex vivo pressure-perfused porcine (96 anastomoses) and human hearts (24 anastomoses). Two blinded observers scanned and scored the anastomoses with epicardial ultrasonography. In 24 human and 24 porcine anastomoses, angiograms were made of 24 correct and 24 incorrect anastomoses and scored by two other blinded observers. Angioscopy and cast injection served as a reference. Overall, 119 of 120 anastomoses were accurately scored as correct or incorrect within a median of 67 seconds (8-381 seconds) by both observers (sensitivity 0.98, specificity 1.00, kappa 1.00 (1.00, 1.00, and 1.00 in angiography subset, respectively). One deep toe stitch that induced outflow corner stenosis was spotted by both observers but regarded as insignificant and thus inaccurately scored as correct. In 5 anastomoses, unintended irregularities were detected. By angiography, anastomoses were accurately scored with a sensitivity of 0.75 and a specificity of 0.81 ( P < .001 vs ultrasonography) and kappa of 0.54. Angioscopy and cast confirmed ultrasonographic findings and did not reveal irregularities other than detected by ultrasonography. Ex vivo epicardial 13-MHz ultrasonography allowed rapid and accurate evaluation of coronary anastomoses and detected technical construction errors with higher sensitivity and specificity than angiography.