P1F-7 On the Influence of Microbubble Concentration on the Assessment of Perfusion Curves Using Contrast Echography

Background: contrast ultrasound allows a perfect depiction of vessels and has shown an improvement in diagnostic accuracy for lesion characterization. Using quantitative parametric curve analysis, perfusion patterns showed to be correlated to lesion malignancy. However, the assessment of perfusion has always been performed by assuming a linear relationship between video intensity and concentration. Our purpose is to evaluate, in-vitro, the efficiency of the perfusion patterns assessment from RF signals by varying ultrasound contrast agent (UCA) concentrations. Methods: Perfusion was evaluated using the replenishment technique with a Siemens Antares scanner and a VF13-5 probe transmitting at 4 MHz in THI mode. Microbubble destruction was realized at maximum MI during 2 seconds with a frame rate of 8 fps. Perfusion was assessed with low MI (0.2) during 16 seconds. A nonrecirculating flow phantom containing a 4mm diameter vessel was used. The liquid flow varied from 40ml/h to 120ml/h. SonoVue (Bracco) microbubbles were diluted from 0.05/1000 to 9/1000. A rectangular region of interest was located at the center of the vessel. Perfusion curves were extracted from RF signals using a research interface toolbox. The experimental results were fitted to an exponential function y = A*(1-exp(-beta*t)) where A represents the steady state and beta the replenishment rate (microbubble flow velocity). Results: the results showed that for a given flow rate and using dilutions up to 1/1000, the coefficient beta remains constant. At 80ml/h, beta was estimated to be 0.68 (SD = 0.03). This result is in accordance with the constant flow velocity. The plateau A increases linearly with the microbubble concentration, indicating a linear relationship between intensity and concentration. However, for higher concentrations (> 1.2/1000), the coefficient beta showed important variations despite a constant flow velocity. Compared to the low concentrations, the estimated flow velocity varies by- up to 60%. Moreover, coefficient A reaches a saturation value although the UCA concentration has increased. Conclusions: this study shows that above a certain concentration of contrast agent, the assessment of perfusion can be biased since the coefficient beta does no longer reflect the flow velocity. Furthermore, the signal intensity cannot be related to the microbubbles concentration