Publication
Article
Digital Edition
Author(s):
Using active learning approach can help gauge visual function loss in patients
A study using this method found that contrast sensitivity function measured in both standard and low luminance conditions decreased with increasing severity of dry AMD and was better in normal controls compared with the dry AMD subgroup representing the earliest stage of disease.
Related: AMD, DED most common retinal conditions in the United States
“AMD is a leading cause of vision loss in the elderly, although visual acuity tends to be spared until late in the disease course,” said William C. Ou, a researcher at the Retina Foundation of the Southwest in Dallas, Texas. “However, patients with earlier stage AMD often have visual complaints, especially under low luminance conditions. Thus, there is a need to identify alternative measures of visual dysfunction that may better reflect the true degree of visual impairment.”
According to Ou, the study suggests that the tool investigators evaluated may be useful as a supplemental visual function measure in patients with AMD.
Although contrast sensitivity has been considered a more sensitive indicator of visual dysfunction in early AMD compared with visual acuity, measuring contrast sensitivity with traditional methods is difficult outside research settings.
The method investigated in the present study uses an active learning approach to estimate a model of the contrast sensitivity function’s global shape.
The algorithm, known as the quick contrast sensitivity function (qCSF), measures the full spatial contrast sensitivity function in just 5 to 10 minutes, which makes it practical for use in clinical settings.
Related: Targeting AMD patient treatment burden
For the testing, patients are asked to identify 25 consecutive optotype triplets varying in size and contrast. The software generates contrast sensitivities at individual spatial frequencies.
Testing for the study was done under standard photopic conditions and then repeated at low luminance created using a 2.0 log neutral density filter.
A total of 65 patients with dry AMD and 23 age-matched controls participated in the study. Those with AMD had disease in at least 1 eye and a visual acuity of 20/80 or better. To investigate correlations between AMD severity and contrast sensitivity, the patients with dry AMD were divided into 3 subgroups defined based on the presence of soft drusen only (ie, intermediate AMD, n = 26), subretinal drusenoid deposits (n = 19), and geographic atrophy (n = 20).
A plot of the qCSF data from testing at standard photopic luminance showed decreasing function with increasing AMD severity across all spatial frequencies. Losses in contrast sensitivity were greatest at low to intermediate spatial frequencies (3 and 6 cycles per degree, equivalent to Snellen sizes of 20/200 and 20/100).
Related:Vision-related quality of life: Impacts of advanced AMD
“Our finding that the group with subretinal drusenoid deposits fell in between the intermediate AMD and geographic atrophy groups is consistent with previous studies that indicated patients with subretinal drusenoid deposits appear to be phenotypically distinct from patients with soft drusen alone,” Ou said.
The results from testing under low luminance conditions were similar, although some significant floor effects were seen at the highest spatial frequencies.
To allow for a statistical analysis of the results, the investigators calculated the area under the log CSF (AULCSF). The AULCSF gives a single value for each contrast sensitivity function.
Qualitatively, the data showed that the AULCSF decreased with increasing AMD severity under both standard photopic and low luminance conditions.
Pairwise comparisons between the AMD severity groups showed statistically significant differences for most but not all comparisons.
Read more by Cheryl Guttman Krader
--
Karl Csaky, MD, Phd
e:kcsaky@retinafoundation.org
William C. Ou presented this research at the 2020 Association for Research in Vision and Ophthalmology virtual annual meeting. He has no financial interests to disclose.