Angiographic Signatures of Brain Collateralization

The severity of ischemic strokes, the likelihood of successful endovascular treatment, and subsequent clinical patient outcomes are all strongly influenced by the degree of collateralization of the brain arterial network. Although several scoring systems based on different imaging modalities have been developed to assess brain collateralization, they are subjective and quite coarse and do not capture many nuances of the underlying collateral vessel distribution. This article introduces a methodology for quantitatively and objectively assessing the degree of collateralization of the brain. The strategy is based on a regional analysis of time-density curves extracted from dynamic angiographic image sequences. To investigate the dependence of the extracted metrics on the underlying collateralization in a controlled manner, a computational study is carried out by generating virtual patients with prescribed collateralization characteristics and conducting blood flow and transport simulations to create virtual angiographic sequences. The virtual angiograms of patients with large vessel occlusions and different collateral vessel distributions are used to generate "angiographic signatures" reflective of the underlying collateralization. The results show that angiographic signatures created from virtual angiograms exhibit consistent characteristics with angiographic signatures of actual stroke patients with different levels of collateralization. These findings indicate that these angiographic signatures can be used to assess the underlying collateralization of the brain and identify patients and conditions associated with adverse treatment outcomes.