A method for estimating the microbubble concentration in contrast-enhanced ultrasound imaging

Abstract

The estimation of the contrast agent concentration can provide useful information in medical diagnostics. Because the intensity of an ultrasound image is not directly correlated with the volumetric concentration of a contrast agent, a method that can estimate the concentration, working only from the signals acquired by the ultrasound scanner, could be particularly useful in terms of precision, time consumption and cost. In this paper, a method to obtain the ultrasound image of a region of interest and the estimation of the related microbubble concentration is proposed. The mentioned tasks are performed in a unique investigation, working from the signals remotely acquired by means of an ultrasound scanner equipped with a grabber board, which is able to collect radio-frequency data. The algorithm is divided into two steps. Firstly, a signal-processing technique, based on multi-pulse transmission and recombination of the received signals, is used to obtain an image of the scene, emphasizing the bubble echoes and abating the contributions of surrounding tissue. Then, the concentration estimation method, based on a nonlinear regression approach carried out by a support vector machine, is applied. Because the training phase requires precise knowledge of the bubble concentration, a completely synthetic training set is assumed, whereas the test set is derived from real signals. In this paper, both non-specific and targeted microbubbles (able to selectively adhere to cancer cells) are considered. The results are encouraging and reveal that the proposed method can provide an accurate estimation in a small volume, which can be useful for diagnostic purposes.