Identifying and monitoring GA: Technologic improvements over time
Srinivas R. Sadda, MD, described the advantages and disadvantages of the various imaging technologies used to evaluate GA at the AAO 2024 conference.
(Image credit: AdobeStock/Fabio)
Keeping a sharp eye on geographic atrophy (GA) associated with age-related macular degeneration (AMD) is the goal, and reaching it has gotten easier over the years with the introduction of more advanced technologies.
Srinivas R. Sadda, MD, a professor of ophthalmology at Doheny Eye Institute, University of California Los Angeles, described the advantages and disadvantages of the various imaging technologies used to evaluate GA at the 2024 American Academy of Ophthalmology annual meeting in Chicago.1
“There currently are numerous modalities to monitor atrophy, and the currently favored ones are confocal color images, green autofluorescence, and optical coherence tomography,” Sadda said.
Running the gamut: Flash color photography to microperimetry
Flash color fundus photography had been the decades-old gold standard technique for diagnosing GA. Based on the color imaging, GA is observed to have sharply demarcated borders, depigmentation, and increased visibility of choroidal vessels.
However, the downside of using color photographs to quantify atrophy is that good stereopsis is required to reliably determine the borders for quantification because of insufficient contrast. In addition, the technology is not practical for use with all patients and within large trials.
Despite those drawbacks, scanning laser ophthalmoscopy (SLO) confocal color imaging has resulted in improved contrast for visualizing GA. This technology can provide very reliable accurate determinations of the atrophic size similar to autofluorescence. Other approaches for imaging GA are autofluorescence and optical coherence tomography (OCT).
Sadda and colleagues conducted a study2 that compared confocal color to flash color imaging. Confocal color achieved an inter-grader reproducibility for measurement of GA that was similar to that of fundus autofluorescence (intraclass correlation [ICC] = 0.998) and significantly better than flash color (ICC = 0.961).
He demonstrated the marked difference in GA visualization in a comparison of blue light autofluorescence to color fundus imaging. While the atrophic area was hard to distinguish on the color fundus image, the GA was readily identified as large patches of decreased autofluorescence on an SLO image. As a result of this capability to sharply delineate GA, autofluorescence has become a very useful tool as an outcome measure in clinical trials.
The potential drawbacks associated with autofluorescence is patient discomfort. A question also has arisen regarding repeated exposure to blue light and any toxic effects in patient with an atrophic macular disease, Sadda pointed out.
Green light autofluorescence is another technologic option that has emerged. Sadda recounted a study3 that compared green light and blue light autofluorescence and found that the former may be more comfortable for patients, and the estimate of the atrophy size was similar between the 2 technologies.
A recent study,4 of which Sadda was the senior author, compared the 2 types of autofluorescence and reported that the GA measurements using Optos ultrawidefield green autofluorescence were highly reproducible and were correlated with the 30-degree blue autofluorescence measurements. A caveat is that clinicians should use the same device for repeated measurements of patients because values can differ from machine to machine.
Using OCT to measure atrophy is more patient-friendly compared with green light autofluorescence. “The atrophy is well visualized on OCT and the increased signal transmission can be appreciated into the choroid in areas in which the retinal pigment epithelium and photoreceptors have been lost,” Sadda said.
He explained that the “hypertransmission” can be exploited and provides a higher sum value of pixels in A-scans in which there is evidence of atrophy. On the scans, this appears as a bright area, ie, hypertransmission on OCT fundus images; this can be used to quantify atrophy. “We and others have shown a strong correlation between the measurement of atrophy and fundus autofluorescence versus on OCT. However, they may not be measuring the same thing,”5 Sadda said.
Defining atrophy
He further explained that until recently, there was no consensus on the classification of atrophy and the choroidal hypertransmission may not be definitive evidence of atrophy.
A consensus panel6 defined atrophy as the presence of the following: hypertransmission of 250 or more microns, a zone of attenuation/disruption of the retinal pigment epithelium plus or minus the basal lamina complex of 250 or more microns, and evidence of overlying photoreceptor degeneration, the features of which include outer nuclear layer thinning, external limiting membrane loss, and ellipsoidal zone and interdigitization zone loss.
When measuring atrophy on autofluorescence images, the hypertransmission defect remains the reliable parameter. Sadda explained that instrumentation tools are available to this end but can be susceptible to segmentation errors.
“We are fortunate that in this era of deep learning the performance of the tools7 is quite good for both fundus autofluorescence and for segmenting of atrophic lesions,” he commented.
Functional testing
Visual acuity may not adequately assess the early progression of GA. The Othera Study8 found that while the topical drug being evaluated improved the visual acuity, the secondary measures of changes in the area of GA showed that the progression was unaffected by the drug. The microperimetry evaluation proved to be more useful to determine the GA progression.
“GA can be diagnosed with multiple imaging modalities. We currently are in the era of confocal color, green autofluorescence, and OCT imaging. These modalities are useful for both the initial diagnosis, assessing the pace of GA progression, and monitoring therapeutic responses. Microperimetry is not useful for the initial diagnosis but may aid the monitoring of functional progression. Automatic quantification added by artificial intelligence will make a difference in this new era,” Sadda concluded. •
