Abstract
Purpose: :
Six radial sampling lines are utilized to determine the retinal map thickness in an ETDRS grid map display in the Stratus OCT. It has been recognized that increased sampling with less extrapolation occurs in the center grid compared to the more peripheral grids. Correlation of these zonal differences to error is explored.
Methods: :
We developed a software model demonstrating the error differences possible with thickness variations mimicking diabetic macular edema. We constructed a grid around the six sampling lines and OCT ETDRS macular grid output display. We then developed a computer algorithm where edema was simulated by varying amount of edema, area of edema, and number of edematous areas . We then analyzed the errors of reported results by zone.
Results: :
The inner foveal zone had the least error rate as expected. In addition, the outer zones had the greatest error rate. For instance, a run of 200 separate simulations of a single lesion 250 microns additional thickness with a lesion length of 250 microns, resulted in a reported mean additional thickness of 115.90 microns for the inner foveal circle with a standard deviation of 59.08 and the mean thickness of the outer quadrant was reported 44.67 microns with a standard deviation of 62.03.
Conclusions: :
There is a large difference in error possible from the central zone to the more peripheral zone on the ETDRS grid in models that mimic diabetic macular edema. Certain patterns of diabetic macular edema are likely to create greater errors than others. The error rates are so large that studies evaluating resolution of edema should best be performed on the center fovea grid when using time domain OCT sampling algorithms. These errors would not be present in spectral OCT sampling methods.
Keywords: imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)