Validation of a predictive model for retinal vascular resistance and blood flow by means of Laser Speckle Flowgraphy

Abstract

AbstractPurpose(1) To mathematically predict retinal vascular resistance and blood flow from minimal input; (2) to validate the model predictions in a healthy population using Laser Speckle Flowgraphy (LSF).MethodsFundus photographs, OCT and OCT‐angiography scans, and systolic/diastolic blood pressure (SBP/DBP) and intraocular pressure (IOP) measurements were performed in 32 healthy subjects. Predicted vascular resistance (PVR) was determined from the central retinal artery and vein equivalents, the fractal dimension of the vasculature, and population‐based hematocrit values, according to the Poiseuille law and an adapted version of the fractal model proposed by Takahashi et al. (2009). Predicted blood flow (PBF) was calculated as OPP/PVR, where OPP is the ocular perfusion pressure. For validation, the mean blur rate (MBR; measure of velocity) of large vessels inside the optic disc and waveform parameters (heart rate [HR], flow acceleration index [FAI], skew, acceleration time index, blowout score and time, fluctuation, rising rate, falling rate [FR]) were recorded by means of LSF. Linear models reduced by the Akaike Information Criterion were used to assess the relationship of PVR and PBF with the LSF parameters.ResultsIn the reduced multivariable model, PVR was higher with higher DBP (p < 0.001), FAI (p < 0.001), and FR (p = 0.042), as well as with lower skew (p < 0.001), MBR (p = 0.001), and fluctuation (p = 0.103). PBF was higher with higher skew (p < 0.001) and MBR (p = 0.040), as well as with lower FAI (p < 0.001) and HR (p = 0.055). The R2 of the models was 0.83 and 0.58, respectively. PVR correlated with retinal nerve fiber layer thickness (RNFLT), but not with macular volume (r = −0.53, p = 0.002; r = −0.218, p = 0.23). PBF correlated with macular volume, but not with RNFLT (r = 0.382, p = 0.031; r = 0.326, p = 0.068).ConclusionsPVR can be used as a surrogate of vascular resistance. PBF provided a lesser fit with the LSF parameters and partially describes retinal blood flow.ReferenceTakahashi T, Nagaoka T, Yanagida H, et al. (2009). A mathematical model for the distribution of hemodynamic parameters in the human retinal microvascular network. J Biorheol 23:77‐86