Perfusion imaging using C-Arm CT system

Perfusion imaging is an important diagnostic and treatment decision-making modality in acute brain stroke management. Thrombectomy, potentially life saving treatment, that comes together with increased risk profile, could be indicated for certain patients solely based on the perfusion scan. The aim of this project is to evaluate applicability of the perfusion imaging for acute brain stroke scanning on C-Arm CT system. This approach could be beneficial for the acute stroke patients as the C-Arm CT device is often a part of the equipment of the operating theater. Having perfusion scanning option on the site of the neurosurgery could spare time and shorten the decision-making process.

The rotational speed of the C-Arm CT device is slower in comparison to the conventional CT rotation. When estimating the velocity of the contrast agent distribution during the perfusion scan, the speed of the rotation of the C-Arm CT device could not be neglected. Therefore, we apply so called time separation technique, where we approximate contrast agent dynamic by the scalar function of the time and fit the data acquired from the scan to the preselected basis of these functions. It has been shown recently, that when the basis functions are chosen based on the prior knowledge, for example by using singular value decomposition of the data from CT perfusion scans, then this method could be used to reliably reconstruct the time attenuation curves.

The aim of this project is to develop the software tools for analysis of C-Arm CT perfusion data with arbitrarily chosen basis functions including those based on the prior knowledge and analytic ones. The software will include image registration of projection data, fitting linear models to those data, obtaining coefficients of the basis functions in projections, cone beam reconstruction of these coefficients into the volumes and the visualization of perfusion parameters (CBF, CBV, MTT, TTP, …). Programs will be implemented in C++ using multi threading approaches.

Further important part of the project is the testing of the algorithms and described methods on the software and hardware perfusion phantoms and evaluating the data. We use existing software brain perfusion phantom and the hardware phantom that was developed on OVGU. Final aim is the transfer of these results to the clinical setting and evaluation of the behavior of these algorithms on real clinical perfusion data.