Exploration of blood flow patterns in cerebral aneurysms during the cardiac cycle

Abstract

This paper presents a method for clustering time-dependent blood flow data, represented by path lines, in cerebral aneurysms using a reliable similarity measure combined with a clustering technique. Such aneurysms bear the risk of rupture, whereas their treatment also carries considerable risks for the patient. Medical researchers emphasize the importance of investigating aberrant blood flow patterns for the patient-specific rupture risk assessment and treatment analysis. Therefore, occurring flow patterns are manually extracted and classified according to predefined criteria. The manual extraction is time-consuming for larger studies and affected by visual clutter, which complicates the subsequent classification of flow patterns. In contrast, our method allows an automatic and reliable clustering of intra-aneurysmal flow patterns that facilitates their classification. We introduce a similarity measure that groups spatio-temporally adjacent flow patterns. We combine our similarity measure with a commonly used clustering technique and applied it to five representative datasets. The clustering results are presented by 2D and 3D visualizations and were qualitatively compared and evaluated by four domain experts. Moreover, we qualitatively evaluated our similarity measure.