AMD Model May Predict Progression

Larry Hand

November 19, 2014

A new statistical model may help ophthalmologists identify patients with age-related macular degeneration (AMD) who may progress from early or intermediate to more serious disease, according to an article published in the November 5 issue of Investigative Ophthalmology & Visual Science. The model may also be able to roughly predict when such patients will progress to wet AMD, which causes 80% to 90% of blindness associated with the disease. The researchers caution that the model will need to be validated in clinical studies before it can be used in patients.

Luis de Sisternes, PhD, from the Department of Radiology and Medicine at Stanford University, California, and colleagues developed the model based on drusen quantitative characteristics of drusen extracted from images captured through spectral domain optical coherence tomography (SD-OCT).

Drusen, yellow or white lipid deposits in the retina or on the optic nerve head, are common signs of AMD.

The researchers extracted characteristics from 2146 SD-OCT scans of 330 AMD eyes in 244 patients at Stanford's Vitreoretinal Clinic during a 5-year period. The study population consisted of consecutive patients diagnosed with AMD on their first visit and analyzed by SD-OCT. Patients had a mean age of almost 81 years, and 72.03% were women. The patients had a mean of 6.43 visits per eye and a mean of 4.82 months between visits.

The researchers extracted 11 quantitative image features, including druse shape, geometry and reflectivity, total area and volume, total number, mean area and volume per druse, maximum height, extent of retinal area affected, and density of the affected area.

The researchers defined exudative, or "wet," AMD as "any manifestation of choroidal neovascularization, fibrovascular scar formation, or fibrovascular pigment epithelial detachment."

They built a prediction model to obtain a score that would indicate likelihood of conversion to wet AMD at a given time in the future. They based the model on features including the presence of advanced wet AMD in the fellow eye, patient age and sex, and time elapsed between clinic visits. They did not include genetics because those data were not collected on visits. They validated their predictions by comparing them with known outcomes at follow-up.

The researchers found that patient age is a significant factor in AMD progression. They also found that maximum height of detected drusen was a significant factor in progression within 18 months and that the total area of drusen had significantly higher values for progressing within 30 months. They found maximum predictive performance at 11 months after a patient's first early AMD diagnosis and that the eyes predicted to progress had a much higher progression rate than eyes predicted to not progress at any given time.

Novel Research

"Our work appears to be novel in being the first to use quantitative imaging features derived from volumetric and longitudinal analysis of individual drusen in SD-OCT images for predicting AMD progression," they write.

The model "needs to be validated in clinical studies before clinicians can actually start using it on patients," senior author Daniel Rubin, MD, assistant professor of radiology and biomedical informatics at Stanford, told Medscape Medical News. "In this study, there was no way to stratify people into risk categories in terms of short-term progression. Before something can be used clinically, it needs to undergo additional study in multiple institutional settings, which is what we're undertaking now."

How Often?

He added, "Patients get these images routinely anyway. The additional thing that the ophthalmologist could do is get, in essence, a risk score that could help them segregate patients into groups, in terms of, 'Is this patient more likely to convert to AMD within the next year or less?' "

The clinician could then decide how often to see the higher-risk patients compared with the lower-risk patients, he added.

Shivan Tekwani, MD, an AMD expert and assistant professor of ophthalmology with Loyola Stritch School of Medicine in Maywood, Illinois, told Medscape Medical News that his standard is to see patients with AMD every 6 months usually, but he might see higher-risk and more serious cases every month.

"Actually, [predicting progression] is an area we're trying to learn more about," he said. "Looking at subjective data like this is actually pretty helpful, because in just about every patient that comes into my office with macular degeneration, they're going to wind up getting [SD-OCT]. If there's a way to take that and more than just merely identify dry versus wet macular degeneration, you could take a look at it and say, 'Is this patient a high-risk patient for wet [AMD],' and that's going to make it easier for them to understand where their disease is at."

Objective information such as that from this study should be incorporated with other information such as genetics and patient demographics to form a whole picture of the patient, he said.

"For me, all the screening techniques and predictive markers [lead to] the question of, How frequently do you monitor patients? That's the 6 million dollar question. How do you actually implement it in clinical practice when you get a screen result that says someone is high risk?" Matthew J. Welch, MD, an ophthalmologist and vitreoretinal specialist in private practice with Retina Specialists of Southern Arizona in Tucson, told Medscape Medical News.

"Until they actually manifest the changes of neovascular [AMD], all you're doing is following them along, recommending vitamin formulation and frequent monitoring. We know that early diagnosis and early treatment result in better visual outcomes, and that's what all of these screening methods are trying to detect," he added.

Because AMD can worsen rapidly, sometimes within a week, the question of how often to monitor patients is a big one, he said. Other patients, however, might take years.

"The fundamental problem with macular degeneration always comes back down to it's so heterogeneous, it's so hard to make broad, sweeping commentary and generalization. It's really challenging," he said.

This research was supported by the Bio-X Interdisciplinary Initiatives Program and the Spectrum-SPADA grant, both of Stanford University. The authors, Dr Tekwani, and Dr Welch have disclosed no relevant financial relationships.

Invest Ophthalmol Vis Sci. 2014;55:7093-7103. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: