ORIGINAL PAPER Using fuzzy logic for morphological classification of IVUS-based plaques in diseased coronary artery in the context of flow-dynamics Ryan Beaumont Æ Kiran Bhaganagar Æ Bruce Segee Æ Ozer Badak Ó Springer-Verlag 2009 Abstract Plaque morphology in a diseased coronary artery plays a significant role in the modification of the fluid flow characteristics. The plaque morphology of 42 patients who underwent IVUS (intravascular ultrasound) procedure was quantified by degree of membership in four fuzzy logic sets, which we refer as type I: protruding, type II: ascending, type III: descending, and type IV: diffuse. Of 42 cases, 28% were of type I, 18% type II, 20% type III and 23% type IV, 6% belonged to hybrid types (partial mem- bers of multiple types) and the remaining 5% did not fit in any category. The degree of membership is of significance as the inter-class blood flow patterns (those strongly members of the same set) are similar to each other com- pared to the intra-class behavior, indicating plaque morphology (shape of blockage) is an important metric in addition to the degree of stenosis to represent the flow characteristics in a diseased stenotic coronary artery. Keywords Fuzzy logic  Classification  Fluid dynamics  Flow simulation  Stenosis  Coronary artery 1 Introduction During the course of the coronary artery disease, the atherosclerotic plaques form on the walls of the coronary artery resulting in a localized narrowing or stenosis of the artery lumen. To date, for computer modeling of the flow in a diseased coronary artery, it is a common practice to approximate the geometry of the plaque formed on the walls of the artery as an idealized protrusion, where the degree of stenosis is represented by the height of this protrusion (e.g. Stroud et al. 2000; Ryval et al. 2004; Mittal et al. 2001; Long et al. 2001; Mallinger and Drikakis 2002). The flow in the diseased arteries is characterized based on the degree of stenosis and it has been the only measure to represent the extent of flow alterations due to the presence of plaque in an artery. One of the reasons for this approximation in numerical modeling of the flow is due to the difficulty in acquiring realistic geometrical information of the plaque. In reality, the plaques are rough and asymmetric, and it is an over- simplification to approximate them as smooth protrusion (Nissen and Yock 2001; Chandran et al. 1996; Krams et al. 1997; Liu et al. 2001; Long et al. 2001; Gibson et al. 1993; Tasciyan et al. 1993) from a flow perspective. Funda- mental studies in fluid dynamics have been performed to study the effect of surface-roughness (e.g. obstruction such as plaques) on the wall of the channel have revealed the importance of the geometrical nature of the surface roughness (shape of the blockage) in altering the flow characteristics (e.g. Antonia and Krogstad 2001; Bhag- anagar et al. 2004, 2007; Perry et al. 1987). It is becoming clear that degree of stenosis is not the only measure of the intensity or characteristics of flow, and suggests a need to explore the dependence of flow features on the morpho- logical features of the plaque. R. Beaumont  K. Bhaganagar (&) Department of Mechanical Engineering, University of Maine, Orono, ME, USA e-mail: kiran.bhaganagar@maine.edu B. Segee Department of Electrical Engineering, University of Maine, Orono, ME, USA O. Badak Department of Cardiology, School of Medicine, Dokuz Eylul University, 35340 Inciralti, Izmir, Turkey 123 Soft Comput DOI 10.1007/s00500-009-0401-9