DefenseSasaGrbic

Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

THIS WEBPAGE IS DEPRECATED - please visit our new website

Doctoral Defense by Sasa Grbic

  • Date: 08.09.2014
  • Time: 10:00
  • Location: Room 03.13.010, FMI-Building, Garching, Boltzmannstr. 3

Advanced Valve Modeling and Multi-Modal Fusion for Minimally Invasive Valve Procedures

Abstract:
The treatment of valvular heart disease has shifted from open-heart surgery towards minimally invasive procedures. These procedures reduce procedural morbidity, mortality and treatment cost while accelerating patient recovery. However as there is no direct view or access to the affected anatomy, pre-operative planning in combination with procedural guidance is crucial for a successful outcome. Thus non-invasive imaging in combination with fast, precise and reproducible image analysis tools will become an essential part of clinical practice. This thesis will concentrate on the two main aspects of minimally invasive valve procedures: planning and guidance. We will present a novel dynamic patient-specific model of the complete valvular apparatus in combination with a robust machine learning framework to estimate the model parameters from computed tomography images. Advanced clinical measurements can be derived from our models and used for diagnosis, patient selection, implant selection and sizing. In addition we provide an advanced planning framework for the aortic valve implantation procedure (TAVI) where we combine a volumetric model estimation method with a virtual valve deployment framework. Further we introduce a robust model-based fusion framework to fuse high-quality pre-operative modalities with low-quality intra-operative images. In the first phase the pre-operative modality will be used to improve the model estimation accuracy in the low-quality intra-operative image. The second part involves developing novel multi-modal model based registration approaches to register pre-operative and intra-operative images. We employ robust machine learning techniques during the estimation. The methods are extensively validated on a large number of patients from multiple medical centers around the world. Both the planning and guidance frameworks have the potential to improve the treatment outcome while lowering procedural risks and treatment costs.

WebEventForm
Title: Doctoral Defense by Sasa Grbic: Advanced Valve Modeling and Multi-Modal Fusion for Minimally Invasive Valve Procedures
Date: 8 September 2014
Location: MI 03.13.010
Abstract:  
Imageurl: gradhat_small.jpg
Type: News
Videourl:  
Conferencelink:  


Edit | Attach | Refresh | Diffs | More | Revision r1.1 - 08 Sep 2014 - 06:04 - MaximilianBaust

Lehrstuhl für Computer Aided Medical Procedures & Augmented Reality    rss.gif