Monitoring radio-frequency thermal ablation with ultrasound by low frequency acoustic emissions--in vitro and in vivo study.

The object of this study was to evaluate the monitoring of thermal ablation therapy by measuring the nonlinear response to ultrasound insonation at the region being treated. Previous reports have shown that during tissue heating, microbubbles are formed. Under the application of ultrasound, these microbubbles may be driven into nonlinear motion that produces acoustic emissions at sub-harmonic frequencies and a general increase of emissions at low frequencies. These low frequency emissions may be used to monitor ablation surgery. In this study, a modified commercial ultrasound system was used for transmitting ultrasound pulses and for recording raw RF-lines from a scan plane in porcine (in vitro) and rabbit (in vivo) livers during radio-frequency ablation (RFA). The transmission pulse was 15 cycles in length at 4 MHz (in vitro) and 3.6 MHz (in vivo). Thermocouples were used for monitoring temperatures during the RFA treatment.In the in vitro experiments, recorded RF signals (A-lines) were segmented, and the total energy was measured at two different frequency bands: at a low frequency band (LFB) of 1-2.5 MHz and at the transmission frequency band (TFB) of 3.5-4.5 MHz. The mean energy at the LFB and at the TFB increased substantially in areas adjacent to the RF needle. These energies also changed abruptly at higher temperatures, thus, producing great variance in the received energy. Mean energies in areas distant from RF needle showed little change and variation during treatment. It was also shown that a 3 dB increase of energy at the low frequency band was typically obtained in regions in which temperature was above 53.3 ± 5° C. Thus, this may help in evaluating regions undergoing hyperthermia. In the in vivo experiments, an imaging algorithm based on measuring the LFB energy was used. The algorithm performs a moving average of the LFB energies measured at segments within the scan plane.Results show that a colored region is formed on the image and that it is similar in size to a measurement of the lesion from gross pathology, with a correlation coefficient of 0.743.