Towards 4DCT-US image fusion for liver motion monitoring

Abstract

High-intensity focused-ultrasound (HIFU) is a promising technique for treating liver tumors. However, liver motion due to breathing imposes a real-time monitoring of the treatment. To reach this goal and following on previous work, we propose to encase an ultrasound (US) imaging probe into an extracorporeal HIFU device, such that the imaging plane is aligned with the HIFU acoustic axis. Because the tumor itself may not always be visible on US images, we plan to rely on a pre-operative 4D-Computed Tomography (CT) model to infer the tumor location during treatment, using intensity- or feature-based registration techniques. In order to study the feasibility of US guidance according to a pre-operative planning image, we decided to perform US acquisitions on patients after informed consent. These patients undergo radiotherapy treatment and have a 4D-CT image for planning (Philips Brillance), with injected contrast-agent. A hand-held US imaging probe was used to provide 2D images sequences in the coordinate system of the CT scanner thanks to an optical tracking system. No device was available to record a breathing signal during US acquisitions. A breathing signal was estimated for each US sequences using Principal Component Analysis. Inhalation versus exhalation was identified by the user, and the respiratory phase was estimated from the breathing signal using the Hilbert transform. Using this signal and the information from the tracking device, it was possible to approximately register both modalities spatially, and throughout the respiration cycle.