Liver cancer, including hepatocellular carcinoma and colorectal metastases, is the second greatest cause of cancer-related death worldwide. Liver transplantation is considered the gold standard for hepatic cancer treatment. However, it is limited by the availability of liver donors and by cost. Hepatic resection is another treatment option that offers a high long-term survival rate. However, the overall resectability rate is low due to chronic liver disease and tumor location. Thermal ablation, including radiofrequency ablation as well as microwave and ultrasound ablation, has become an important alternative to liver resection and transplantation. To avoid incomplete treatments and cancer recurrence while reducing morbidity, a real-time monitoring and control approach, capable of providing consistent thermal ablation in minimal time, is needed.
Echo decorrelation imaging has been successfully employed to monitor ultrasound ablation and radiofrequency ablation both ex vivo and in vivo. In this dissertation, its utility for real-time control of ultrasound ablation was assessed in ex vivo bovine liver and in vivo rabbit liver with VX2 tumor. Ultrasound exposures and echo decorrelation imaging were performed using 5 MHz linear image-ablate array. Sonications were automatically ceased when the minimum or average cumulative echo decorrelation within a control region of interest (ROI) exceeded a predetermined threshold, corresponding to high specificity for prediction of local tissue ablation or complete ROI ablation. Ablation outcomes, treatment time and prediction performance were statistically compared between controlled and uncontrolled groups.
For controlling ex vivo focused ultrasound treatments, a small control ROI was placed at the focal zone and the selected control threshold corresponded to 90% specificity of local ablation prediction for preliminary ex vivo experiments. Controlled trials were compared to uncontrolled trials employing 2, 5 or 9 therapy cycles. For controlling ex vivo unfocused ultrasound treatments, an optimization approach was developed to determine stopping criteria for two series of controlled experiments using different echo decorrelation imaging feedback parameters and sonication sequences. Controlled trials were compared with uncontrolled trials employing 9 or 18 therapy cycles of matching sonication sequences. For controlling in vivo focused and unfocused ultrasound treatments, the selected control threshold corresponded to 90% specificity for tumor ablation prediction in previous in vivo experiments. Controlled trials were compared with a previous series of uncontrolled in vivo experiments employing focused and unfocused ultrasound ablation in rabbit liver and VX2 tumor.
In general, controlled trials resulted in higher ablation rate, smaller lesion dimensions and treatment time compared to uncontrolled trials with longer duration. However, controlled trials showed equivalent lesion dimensions and treatment time, but higher ablation rate compared to uncontrolled trials with shorter duration. Better echo decorrelation prediction capability was observed for controlled trials compared to uncontrolled trials with shorter duration, but equivalent prediction performance was observed when compared to uncontrolled trials with longer duration. For most controlled trials, integrated backscatter imaging showed similar behavior to echo decorrelation for local ablation prediction. These results indicate that control using echo decorrelation imaging may improve the reliability and duration of ultrasound-guided focused and unfocused ultrasound ablation treatments.