Abstract
In peripheral revascularization procedures, an obstructed vessel is unblocked to restore the blood flow to the tissue. Currently, treatment assessment is carried out by angiography which allows only for a qualitative inspection of the blood flow in arteries. Periprocedural assessment of tissue perfusion would be a valuable tool to assess
... read more
the effective restoration of blood flow toward the tissue and thus to evaluate the technical success of the revascularization procedure. Perfusion quantification however requires the knowledge of the dynamic contrast evolution in 3D which is only available on tomographic modalities as CT, PET and MR. This thesis investigates different possibilities to measure quantitative perfusion with interventional C-arm systems. The current use of C-arm systems is merely for 2D angiographic acquisitions and live guidance during interventions, however recent developments have enabled 3D imaging also on these systems. Nevertheless the dynamic 3D performance of C-arms are limited in comparison to CT which makes perfusion imaging a challenging task. In this thesis we propose a set of perfusion estimation methods based on the combination of 2D and 3D angiographic images acquired on commercially available C-arm systems. To tackle the lack of temporal and spatial information in the acquired data, we exploit prior knowledge on the spatial and temporal distribution of contrast in the tissue and integrate this knowledge in the estimation procedure. In particular we assume that in legs the perfusion is spatially homogeneous over volumetric regions which resemble muscle blocks and vessels. Due to the assumption on homogeneity the spatial resolution achieved with the proposed methods is slightly inferior to that of fully 3D estimation methods. The thesis focuses on the technical implementation of the methods and on their evaluation on simulated and clinical data. Results on phantom data showed that the contrast resolution and the spatial resolution are sufficient to differentiate hypoperfused tissue from healthy tissue in patients affected by peripheral arterial disease. The evaluation on clinical data proved the feasibility of the method in a clinical environment. However a number of limitations still needs to be addressed in order to apply the methods in the clinical routine. In particular the amount of contrast agent and radiation need to be reduced. In conclusion, in this thesis we show that quantification of tissue perfusion from planar (i.e. 2D) angiographic images can be achieved by adding anatomical information derived from 3D images. Based on the contributions made in this thesis, direct and quantitative assessment of peripheral revascularization procedures could be made available in the future and could be used as a tool to optimize revascularization procedures.
show less
