Description
Module 1: Biplane overlay system The key enabling technology for this project is the accurate integration of real-time x-ray images with 3D-MR images acquired prior to the intervention. We will adapt research developed at the “Division of Imaging Sciences and Biomedical Engineering” to integrate the MR and x-ray images in a clinical environment and enable this technology to be available within a commercially available biplane x-ray system for the first time. During catheter interventions for congenital heart disease, multiple different projections are used. The need to change rapidly between different projections requires a biplane x-ray system. The focus will be on developing an accurate and robust overlay system that works in real-time in any chosen projection. Currently available technology only operates on a single-plane system at 3 frames per second but this project will extend and improve upon this. For registration we will use technqiues developed by Truong et al. (MICCAI 2012 & MICCAI 2013) within the Division. The accuracy of image overlay will be improved by adding a novel respiratory compensated module, which is based on hierarchical manifold learning (Panayiotou et al, MICCAI 2013). These approaches have been validated off-line but will be developed to operate in real-time within clinical timescales. The key developments will be interfacing to the biplane x-ray system using dual frame-grabber technology; tracking of the x-ray system gantry and table; accurate calibration of the x-ray system and accurate and robust device detection and tracking. Furthermore, we will integrate these developments into single software architecture with a robust user interface making it accessible for clinicians without requiring new skill sets. Validation will be carried out using physical phantoms prior to patient testing. Milestone 1: Development of an accurate and robust biplane overlay system with automated device detection and tracking and motion compensation. Module 2: Intelligent Control of the X-ray System The ramification of module 1 is that the interventional environment will be augmented by 3D anatomical roadmaps reducing the amount of x-ray exposure during the interventions. In fact, the x-ray fluoroscopy will only be required to visualize the interventional catheters and devices. In Module 2, techniques will be developed that couple the device detection algorithms directly to the x-ray control parameters. Exposure settings will automatically be adjusted in real-time to maintain confident detection of the devices (Panayiotou et al, Med Phys 2014). Furthermore, motion tracking of the devices will be coupled to the frame rate and radiation adjustment control of the x-ray system thus further reducing x-ray exposure. There will be an option to adjust the overlay manually. The x-ray dose will be automatically controlled to achieve the lowest possible x-ray dose for the computer to recognise the catheter (the human operator needs a much higher x-ray dosage for catheter recognition). Siemens Healthcare will be our partner for this module. The approaches will be tested using anthropomorphic phantoms and, once phantom studies are finalized, on patients undergoing interventions. Milestone 2: Development and testing of intelligent x-ray control software Module 3: Virtual guidance system The aim of module 3 is to take the developments of modules 1 and 2 and combine them into a single 3D virtual guidance system. The key development will be the simultaneous visualisation of the 3D cardiac anatomical models with the 3D device models (from module 1) in real-time complete with a robust user interface. The platform will be validated using anthropomorphic phantoms and then it is to be validated during clinical procedures. Milestone 3: development of a virtual guidance system; testing of the prototype in phantoms to achieve the required robustness and accuracy; testing of the prototype during clinical procedures| Period | 2015 → 2018 |
|---|---|
| Degree of Recognition | International |