Ultrasound contrast agent in intravascular echography: parametric mapping based on RF output

Detection of endoluminal contour on IntraVascular UltraSound (IVUS) images is a determining factor for early diagnosis of pathologies like atherosclerotic plaque and dissections. Nevertheless, the echogenicity of blood and that of tissues or plaque are close, making the detection of intraluminal contour a challenging task. The purpose of this study is to obtain automatic intraluminal edge detection using ultrasound contrast agent (USCA) perfusion. In a previous study, the contribution of USCA in intravascular echography has been shown using B-scan image processing. A parametric mapping is now proposed from RF signals. The signal processing method implemented is based on the computation of time variance parameters from a sliding window applied on each RF line. A statistical study has been performed to determinate the most relevant parameter. Since USCA circulates through the intra-lumen, the time variance gives a higher value inside the lumen than in the extra-arterial medium. Parametric mapping allows endoluminal contour enhancement by reducing the speckle and increasing the SNR of 7 dB. This method is a good pre-processing procedure to improve the automatic intraluminal edge detection technique based on a deformable model approach. The IVUS system consisted of a CVIS endovascular ultrasound scanner, an ultrasound catheter with a 30 MHz, central frequency. The RF signals were acquired by a Lecroy digital scope at a sampling frequency of 200 MHz, the 4 Mbytes memory of the oscilloscope allowing the capture of 7 images of 256 scan-lines each for an investigation depth of 7 mm. In vitro experiments were conducted with a cryogel phantom mimicking artery perfused by a sodium chloride solution at a flow rate of 88 ml/min. About 0.1 ml USCA (Sonovue^TM) bolus was injected into the system. With the benefit of USCA, we demonstrated the feasibility of the parametric mapping to enhance the contrast of luminal edges, and this enhancement improves the automatic intraluminal edge detection by a deformable model