Differential Attenuation Imaging for the Characterization of High Intensity Focused Ultrasound Lesions

High intensity focused ultrasound (HIFU) is an effective technique for creating coagulative necrotic lesions in biological tissue, with a view to treating localized tumors. Although good results have already been obtained, notably in urology, current systems lack a real time monitoring system to check the efficacy of the treatment procedures. This study describes the development and assessment of a noninvasive system for making local measurements of attenuation variations during HIFU treatment procedures. An apparatus (Ablatherm, Edap-Technomed, France), combining a 2.5 MHz therapeutic transducer and a 5.5 MHz twin plane imaging probe (connected to an ultrasound scanner), was used to produce lesions. The rf signals needed to calculate the attenuation were recorded as outputs from the ultrasound scanner, before and after the high intensity firing sequences, which were performed on ten pieces of porcine liver. Each firing sequence involved producing a lesion volume comprising 42 individual lesions. A number of recordings were also made without producing lesions, in order to test the reproducibility of the measurements. The attenuation function was evaluated locally using the centroid and the multinarrowband methods. Initially, changes in the integrated attenuation αbar; (mean attenuation in the 4–7 MHz range) and the attenuation slope β were examined for the lesion volume. β values did not vary significantly within this range, whereas α values varied significantly (in the region of 86% of the initial level) in comparison to measurements performed without forming lesions. The differential attenuation Δα (representing local variations in α) was subsequently used to generate images revealing the lesion areas. There was a strong similarity between these ‘Δα images’ and the lesion volumes defined by the operator. ‘Δα images’ offer several advantages over existing attenuation imaging techniques. Any problems related to the heterogeneity of the medium are eliminated, since only the change in attenuation is taken into account. Furthermore, there is no need to compensate for diffraction when estimating Δα, as the rf signals are captured in exactly the same positions before and after treatment. This technique can be used during in vivo treatment procedures. It can be implemented in real time, since the computational algorithms (based primarily on FFT calculations) are very fast. The technique should provide clinical practitioners with valuable qualitative and quantita