Digital subtraction for noise reduction in intravascular ultrasound data

Intravascular ultrasound data (IVUD) is susceptible to noise degradation from a variety of sources. In an in-vitro study, the potential of digital subtraction (DS) techniques to reduce image noise from IVUD is examined. A 20 MHz, rotating-transducer catheter was used to collect images of vascular phantoms immersed in canine blood, and images of blood only using the same instrumentation settings. Images were digitized and subtracted using a simple mask-mode algorithm. First-order gray-scale (GS) histogram statistics were computed on all image data. GS standard deviation and GS standard deviation normalized to mean GS were used as the primary measures of artificial noise. Statistical comparison of the image sets revealed significantly lower noise content in difference image data than in the initial image sets. This preliminary study indicates that DS may be an effective way to enhance interpretation of IVUD