|Title||REDUCING RNFL THICKNESS VARIABILITY MEASURED USING THE STRATUS OCT|
|Author, Co-Author||Brittany Adams, William Swanson|
Stratus OCT is commonly used to assess retinal nerve fiber layer (RNFL) thickness. These measurements aid in the detection and management of glaucoma, although many artifacts including eye movements can make image acquisition difficult and yield unreliable results. Signal strength (SS) is widely used as an exclusion factor to reduce measurement variability. However, a new analytical program has been developed that estimates the amount of scan circle misalignment for the Stratus OCT. We used the program to identify eccentric scan circles and determine the displacement amounts.This program can potentially improve the identification of unreliable RNFL thickness measurements. Comparing both traditional and novel methods by implementing specific exclusion criteria, it is our goal to reduce RNFL thickness measurement variability in order to better monitor the progression of loss in glaucoma.
26 people ages 45-85 were used as subjects, each with one eye tested. Each subject had reproducible visual field defects, glaucomatous optic nerve damage, and was free from other ocular disease. We employed a SS exclusion criterion of <7 (out of 10), qualifying an OCT scan as unsatisfactory, and examined 6 RNFL sectors (T, ST, IT, N, SN, IN) as well as overall RNFL thickness. Secondly, we used the new analytical method based on scan circle displacement and employed an exclusion criteria by discarding scans with >0.3mm displacement. For both methods, we evaluated standard deviation (SD) decreases across RNFL sectors and the percent of scans lost.
Both the traditional and novel approach discarded a number of scans resulting in subjects losing data from multiple visits over time, with SS exclusion losing 25% of subject visits as well as two subjects losing all visits and displacement exclusion losing 10%. Displacement exclusion tended to provide both larger SD decreases and increases than SS exclusion. SS exclusion provided more instances of RNFL sectors decreasing in SD (44%) over displacement exclusion (35%).
Both SS and displacement methods decreased average SD per subject. Further studies are needed to fully understand both methods.
T35-EY013937 and R01EY007716 grants
|Affiliation of Co-Authors|