Fifty-sixth annual meeting of the American association of physicists in medicine
SU-C-18C-07: Quantitative Measurement of Arterial Arrival Time in Pial Collaterals: Effect of Physical Parameters On Performance
During Middle Cerebral Artery Occlusion (MCAO), the presence pial collateral capillaries can affect a stroke's severity and subsequent recovery Arterial arrival time (AAT) is the difference in contrast arrival time between a reference artery and a measured vessel, and can serve as an indirect angiographic measure of pial collateral recruitment during acute ischemic stroke. No automated method for AAT assessment is currently available on digital subtraction angiography. We report on an automated method for AAT maps derived from DSA and potential pitfalls that confound the accuracy of these maps.
Our dataset consisted of digital angiographic images of five canines obtained on a GE OEC 9800 Plus Cardiac system using, 490 × 492 pixel images and iohexol, Images had been obtained after an induced MCAO. Our automatic method corrects for drift in x-ray output and measured noise in a region with no contrast. For each pixel, contrast arrival time was calculated using thresholding. AAT is the difference in contrast arrival time between a reference artery and a measured vessel, and yields information on the extent of pial collateralization in a measured vessel. This method was also validated against a manual determination of AAT. Deviations between manual and automatic methods were assessed across various physical parameters.
Factors that degrade the automatic determination of the AAT included motion, vessel orientation, the partial volume effect, pixel location from the vessel's center, and contrast agent dispersion. For example, contrast agent dispersion caused errors in the AAT from 33 to 891 milliseconds, across three dogs, for arrival times between a reference artery of the middle cerebral artery and a pial collateral vessel.
For parameters studied, temporal distance can be validated down to milliseconds, thus allowing for the objective assessment of AAT and the potential appraisal of a stroke's severity.