Pranav Bhagirath

117 Interventional cardiac magnetic resonance imaging in electrophysiology: advances toward clinical translation resolution T 2 -mapping techniques in 2D 42 and even in 3D 43 can be used to visualize edema. Necrosis can be visualized using the commercially available T1-weighted PSIR turbo-GRE 31 or using the experimental 3D respiratory-navigated IR-GRE sequence 24 . Direct feedback during the ablation procedure may contribute to a complete isolation during the first attempt. In case there is reoccurrence of arrhythmia’s due to incomplete isolation, this information can be utilized to perform targeted gap ablation. Detailed MRI-derived information with respect to lesion formation has therefore the potential to improve single procedure success rates 12 . IMAGE RECONSTRUCTION Implementing an iCMR suite is significantly different from the conventional diagnostic imaging set-up and requires an optimization of the imaging workflow. Performing an interventional study requires 1) real-time acquisition and near real-time reconstruction, 2) rapid sequence changes and 3) integrationwith device hardware for tracking purpose ( figure 3 ). It is has been shown that standard sequences, present on every diagnostic scanner, can be used for performing and evaluating interventions 31 . However, there are three specifically designed visualization frameworks targeted at the EP community (i.e. Gadgetron, Vurtigo and RTHawk), that are user-friendly and can facilitate procedural guidance. An overview of these frameworks and their respective functionalities is provided below. Gadgetron Gadgetron is an open source, modular framework designed to facilitate rapid development and easier sharing of image analysis and reconstruction algorithms 44 . Modules (gadgets) can be developed using standard toolkits present in the framework. These gadgets are programmed to perform specific post-processing tasks and can be configured as part of the scanner data pipeline. Multiple gadgets can be used for simultaneous processing. Furthermore, the individual gadgets canbe re-used indifferent workflows thereby enabling thorough evaluation of consistency and performance. However, this dynamic framework does have amajor limitation. There are no possibilities to control theMRI scanner directly using the tools provided. Therefore, image acquisition can only be performed using the existing scanner front-end. Due to the restrictions of the in-built interfaces, this will require deployment of third-party solutions.