In situ guidance for MRI interventions using projected feedback.

PURPOSE: To develop and evaluate an augmented reality instrument guidance system for MRI-guided needle placement procedures such as musculoskeletal biopsy and arthrography. Our system guides the physician to insert a needle toward a target while looking at the insertion site without requiring special headgear. METHODS: The system is comprised of a pair of stereo cameras, a projector, and a computational unit with a touch screen. All components are designed to be used within the MRI suite (Zone 4). Multi-modality fiducial markers called VisiMARKERs, detectable in both MRI and camera images, facilitate automatic registration after the initial scan. The navigation feedback is projected directly onto the intervention site allowing the interventionalist to keep their focus on the insertion site instead of a secondary monitor which is often not in front of them. RESULTS: We evaluated the feasibility and accuracy of this system on custom-built shoulder phantoms. Two radiologists used the system to select targets and entry points on initial MRIs of these phantoms over three sessions. They performed 80 needle insertions following the projected guidance. The system targeting error was 1.09 mm, and the overall error was 2.29 mm. CONCLUSION: We demonstrated both feasibility and accuracy of this MRI navigation system. The system operated without any problems inside the MRI suite close to the MRI bore. The two radiologists were able to easily follow the guidance and place the needle close to the target without any intermediate imaging.

A meta registration framework for lesion matching.

A variety of pixel and feature based methods have been proposed for registering multiple views of anatomy visible in studies obtained using diagnostic, minimally invasive imaging. A given registration method may outperform another depending on anatomical variations, imaging conditions, and imaging sensor performance, and it is often difficult a priori to determine the best registration method for a particular application. To address this problem, we propose a registration framework that pools the results of multiple registration methods using a decision function for validating registrations. We refer to this as meta registration. We demonstrate that our framework outperforms several individual registration methods on the task of registering multiple views of Crohn's disease lesions sampled from a Capsule Endoscopy (CE) study database. We also report on preliminary work on assessing the quality of registrations obtained, and the possibility of using such assessment in the registration framework.