Interactive Visualization System for Endoscopes

Rapid developments in minimally invasive surgery have promoted the effective use of endoscopes for a wide variety of diagnostic and therapeutic procedures, using the natural orifices of the body or small, sub-centimeter incisions. The smaller incisions made during resections and ablations reduce morbidity by minimizing tissue damage, reducing pain and improving convalescence. If an operative approach is restricted due to inaccessible or sensitive areas, minimally invasive endoscopic surgery is able to leave surrounding structures untouched. The endoscope has its limitations, however: endoscopes can show only visible surfaces the presence of fog, blood and/or fluids can obscure the exposure of the region of interest endoscopes are generally disorienting due to the lack of directional clues. Virtual endoscopy systems are now widely used to visualize internal anatomy. Using previously-acquired CT, MR or ultrasound images, a virtual endoscopic view from anywhere in or around the dataset may be generated. The advantages of virtual endoscopy are: the absence of physical geometrical constraints the ability to vary the opacity of any structure and thus to see through it One of its main disadvantages can be seen in the fact that tissue deformations during interventions misaligns the preoperative images and the real-time view and decrease accuracy of biospsies. Our long-term goal is to develop new technology to improve rigid and flexible endoscopic biopsy and aspiration procedures by enhancing visualization, orientation and exposure, increasing accuracy, reducing procedure duration, and reducing patient discomfort. The objective of the proposed research is to integrate an endoscope, a three- dimensional (3D) ultrasound system, data reconstruction and registration, correction of tissue deformation, and a real-time interactive navigation engine to create an interactive image-guided navigation system using real-time ultrasound data. The envisaged system will compensate for the limitations of endoscopic interventions, fundamentally advancing this minimally-invasive surgical intervention and improving surgical outcomes. It will enable surgery to be performed using the same set of tools and basic procedures as current endoscopic biopsies, but with the added benefit of greater precision and a commensurate increase in the probability of accurate biopsy. Contemporary image guidance systems have limited positional accuracy, provide only two-dimensional (2D) image updates from pre-operative data, do not provide six degrees-of-freedom (DOF) tracking of flexible endoscopes and do not regard tissue deformation during interventions. These are the limitations that we aim to address.

We developed a navigation system for flexible endoscopes which improves endoscopic biopsy procedures by enhancing visualization and orientation, increasing accuracy, reducing procedure duration, and reducing patient discomfort. Beside the development of several key procedures which were necessary for our image guidance system we invented a registration method for patient re-positioning without the use of ionising radiation. This method was used to transfer information from pre-interventional, high-resolution CT images to real-time but noisy ultrasound images. The same method was applied in radiation therapy to ensure the same alignment of the patient on the treatment table as given at the planning CT. The registration method introduces a 3D ultrasound to 3D ultrasound image registration which shows best result among image registration methods because same image modalities are applied. The first 3D US image is immediately taken after the pre-operative CT acquisition, the second US image immediately before (endoscopic or radiation) intervention takes place. This also allowed for detection and correction of tissue deformation which is a main challenge in image guidance.