A dynamic reconstruction approach for cerebral blood flow quantification with an interventional C-arm CT

Tomographic perfusion imaging is a well accepted method for stroke diagnosis that is available with current CT and MRI scanners. A challenging new method, which is currently not available, is perfusion imaging with an interventional C-arm CT that can acquire 4-D images using a C-arm angiography system. This method may help to optimize the workflow during catheter-guided stroke treatment. The main challenge in perfusion C-arm CT is the comparably slow rotational speed of the C-arm (approximately 5 seconds) which decreases the overall temporal resolution. In this work we present a dynamic reconstruction approach optimized for perfusion C-arm CT based on temporal estimation of partially backprojected volumes. We use numerical simulations to validate the algorithm: For a typical configuration the relative error in estimated arterial peak enhancement decreases from 14.6% to 10.5% using the dynamic reconstruction. Furthermore we present initial results obtained with a clinical C-arm CT in a pig model.