Dynamic imaging of tumor perfusion using contrast enhanced ultrasound: In vivo results

Abstract

Despite almost 50 years of research on the use of microbubbles as ultrasound contrast agents (UCAs), the promise of high resolution dynamic perfusion imaging has not been fulfilled. This is due to the fact that the echogenicity enhancements from small clusters of bubbles in microvessels remain difficult to detect in the presence strong tissue echogenicity. A well-known pulse inversion (PI) method has been successful in exploiting the nonlinear behavior of UCAs and has led to enhanced myocardium and vascular imaging procedures. However, PI imaging has limited dynamic range due to noise amplification (fundamentally a subtraction method) We have recently proposed the use of the post-beamforming Volterra filter for separation of linear and non-linear echoes and maintaining very high dynamic range in addition to effective suppression of additive Gaussian noise. This method was shown to match or exceed the performance of PI imaging in static imaging of UCAs in flow phantoms. In this paper, we present in vivo results of UCA imaging using the cubic component of a third-order Volterra filter, together with a pixel-wise estimate of a temporal-perfusion index (TPI). We show that the TPI can be designed as a spatio-temporal estimator of the perfusion activity to provide separation between UCA activity and changes due to tissue motion due to breathing and/or pulsation. This is achieved using UCA injections with clinically-relevant concentrations and imaging at normal scanner setting, i.e. no sacrifice of bandwidth as may be necessary when using PI imaging. The results also demonstrate the high resolution nature of TPI imaging in both axial and lateral dimensions.