Automatic detection of microemboli during percutaneous coronary interventions

The objective of this study was to develop an analysis method for the automatic detection of intracoronary microemboli triggered high intensity signals (HITS) during percutaneous coronary interventions (PCI). The recorded ultrasonic Doppler velocity spectra from an intracoronary ultrasonic guide-wire were decomposed into 13 wavelet scales applying the continuous wavelet transform. From 7 wavelet scales which were most suitable for a differentiation between HITS and pulsatile flow, envelopes were calculated and combined to improve the HITS-to-background noise ratio. For different intensity thresholds the resulting number of HITS was automatically counted and compared with the number estimated by experienced observers. In a first validation trial HITS were detected within a simplified in vitro model with a sensitivity of 89.2% and a positive predictive value of 87.6%. In a following clinical study 211 HITS from 18 patients during PCI were counted manually by the observers. With the developed wavelet-based method 189 HITS were correctly detected (sensitivity of 89.6%, positive predictive value of 85.5%). The introduced new method automatically detects intracoronary HITS for the first time with a reliable accuracy. This may support further studies evaluating the incidence and consequences of coronary microembolization during coronary interventions.