Temporal averaging for quantification of lumen dimensions in intravascular ultrasound images

Abstract

Quantitative analysis of arterial dimensions from high frequency intravascular ultrasound images (30 MHz) may be hampered by strong blood scattering. Replacement of blood by saline is one method to provide a clear view of the arterial lumen; another method is that of temporal averaging of successive ultrasound images. The accuracy of this latter method was tested by comparing the lumen area measurements on the temporal-averaged image, with the data of the same cross-section obtained from the single-frame and saline-filled images. The mean lumen area measured on the temporal-averaged images was similar to that measured on the single-frame images (mean difference: −0.02 ± 1.16 mm 2; p = ns). The mean lumen area of the saline-filled images was 8% larger than the values obtained from the temporal-averaged and single-frame images (mean difference: −1.14 ± 0.85 mm 2, p < 0.05), probably due to the difference in sound velocity between saline and blood. Intraobserver variation in the averaging method were 2.4 times smaller than the measurements of the single-frame images and close to the data obtained by saline injection (variation coefficient: single-frame: 8.8%; temporal-averaged: 3.6%; saline-filled: 2.9%). It is concluded that analysis from temporal-averaged images is more efficient, enabling accurate and reproducible measurement of the luminal dimensions from images containing blood scattering echoes. This technique is suitable to replace the laborious saline injection method and facilitates off-line quantitative analysis.