In-vivo perfusion quantification by contrast ultrasound: Validation of the use of linearized video data vs. raw RF data

Abstract

Perfusion quantification software commonly analyzes log-compressed video images provided by ultrasound systems. Compared to analysis of raw RF signals, errors in perfusion parameter estimates may be introduced, even when proper linearization is applied to reverse the log-compression. In this work, the effects of log-compression, demodulation, scan conversion and 8-bit quantization have been investigated, in terms of the errors introduced on perfusion parameter estimates. An in-vivo contrast bolus sequence was processed using both raw RF data and linearized video data. The outcome is that linearizing log-compressed video sequences can provide estimates as accurate as processing RF data under two conditions: a) system gain should be such as to prevent both excessive truncation of pixel values to minimum level and saturation to maximum level; b) the dynamic range of log-compression should be 45 dB or higher.