Sensitivity of the gradient oscillatory number to flow input waveform shapes.

The sensitivity of the gradient oscillatory number (GON), which is a potential hemodynamic indicator for cerebral aneurysm initiation, to flow input waveform shapes was examined by performing computational fluid dynamics (CFD) simulations of an anatomical model of a human internal carotid artery under three different waveform shape conditions. The local absolute variation (standard deviation) and relative variation (coefficient of variation) of the GON calculations for three waveform shapes were computed to quantify the variation in GON due to waveform shape changes. For all waveform shapes, an elevated GON was evident at a known aneurysm site, albeit occurring at additional sites. No significant differences were observed among the qualitative GON distributions derived using the three different waveform shapes. These results suggest that the GON is largely insensitive to the variability in flow input waveform shapes. The quantitative analysis revealed that GON displays an improved relative variation over a relatively high GON range. We therefore conclude that it is reasonable to use assumed flow input waveform shapes as a substitute for individual real waveform shapes for the detection of possible GON elevations of individual clinical cases in large-scale studies, where the higher values of GON are of primary interest.