The Challenge of Predicting OAG Progression from the Initial Visual Field Test

Purpose: The aim of this study was the development of a critical code in order to combine statistics with a workable diagnostic system for open angle glaucoma that could predict improvement or deterioration of the tested visual field of a glaucoma suspect, most likely after the first or second visit for the visual field test.

Methods: The study plan was to apply a set of different filters in order to select the most efficient one that could remove the most of the noise of the test printout, of which probably the greater part of this removed noise could be the component of the learning effect, as it was expected to be at the first or second session. The common mean and median filters were initially used and later on an adapted or Hybrid filter was designed in MatLab© environment and in a similar philosophy to Gardiner’s Predictor filter. Taking into account the details of the study data, an Adaptive or Hybrid filter following the deployment of the optic nerve fibre layers of the retina was tested and selecting different weight depending on the locations of possible glaucoma defects.

Results: Initially, the used mean and median filters, used to remove noise of visual field provided ambitious results. The first filter blurred the edges and the overall appearance looked fuzzy or blurry. The second one calculated the values of the neighbourhood and set these in ascending array. Then selected the median of these values to replace the original one. The result in general looks misleading. Next, applying the Hybrid Adapted filter, the end results illustrated elimination of measured noise in the visual field tests and likely the first visit outcome could predict the third or the fifth visit one.

Conclusion: This is a promising approach to identify and eliminate measurement noise in the visual field tests and to predict, after filtering the first examination outcome, the likely visual field outcome of the third or the fifth visit. The challenge of predicting the progression of open angle glaucoma from the initial visit nowadays is even more than any other the “Holy Grail” of Perimetry.