Changes in retinal tissues known to be associated with Parkinson’s disease (PD) may occur up to 7 years before clinical symptoms of the disease appear, a new study suggests.
Researchers used artificial intelligence (AI) to analyze data from two population-level data sets and the world’s largest database of retinal images and associated clinical data to detect the retinal changes in patients with PD and in healthy individuals who developed the disease years later.
Prior research had shown that PD is associated with a thinning of the ganglion cell-inner plexiform layer (GCIPL) in the retina, something that investigators confirmed in this new study. But they also identified changes in the inner nuclear layer (INL), which is a new finding.
The study is the largest to date on retinal markers in PD and the first to show these changes in living patients.
“I think we are still several years away from converting these findings into individual level prediction for patients,” lead author, Siegfried Wagner, MD, MsC, Honorary Clinical Senior Research Fellow at Moorfields Eye Hospital and University College of London Institute of Ophthalmology in London, told this news organization. “The most important takeaway is that there are observable differences in the retina of individuals who go on to develop Parkinson’s disease.”
The findings were published online in Neurology.
Another look at OCT
Researchers used data from retinal eye scans taken by optical coherence tomography (OCT), a noninvasive three-dimensional imaging technology that is widely used by opticians.
Other studies have used OCT to detect retinal changes in multiple sclerosis and cognitive decline.
For this research, investigators identified markers in people with PD using ophthalmic imaging data from 700 patients and 105,770 controls who participated in the retrospective AlzEye study.
After adjustment for age, sex, ethnicity, hypertension, and diabetes, individuals with PD had significantly thinner GCIPL and reduced thickness of the INL.
To evaluate retinal changes in patients before a PD diagnosis, researchers then turned to 50,405 participants in the UK Biobank with no history of PD who received a retinal scan as part of their baseline visit. Of that group, 53 were diagnosed with PD during the study period.
Researchers found an association between new diagnoses of PD and reduced thickness of the GCIPL (hazard ratio [HR], 0.62; P = .002) and thinner INL, especially at the inferior subfield (HR, 0.66; P = .002). That association persisted even in people whose clinical symptoms developed within 2 years of the retinal scan.
“We wonder if the reduced INL thickness is indicating a direct dopaminergic impairment occurring within the inner retina,” Dr. Wagner said. “Dopaminergic amacrine cells only account for a small proportion of the cells in this layer but previous work in the laboratory shows observable abnormalities in Parkinson’s disease.”
Too early for diagnostics?
Commenting on the findings, Rebecca Gilbert, MD, PhD, chief scientific officer, American Parkinson Disease Association, noted that the changes in the retinal thickness identified in the study were too small to be useful in the clinic as a screening tool for early PD.
“In order for that to happen, the specificity and sensitivity needs to be established,” she said. “Both specificity and sensitivity need to be high enough so that the test can be used to give clinically meaningful results – and reliably tell an individual with PD that he or she does have the disease and individual without PD that he or she doesn’t have the disease.”
Authors of an accompanying editorial agreed. Valeria Koska, MD, and Philipp Albrecht, MD, both of Heinrich Heine University Düsseldorf in Germany, noted that though the effect sizes of retinal changes were small, the study “sets new standards for the role of retinal morphology as potential biomarker in neurodegenerative disease.”
The study was funded by Fight for Sight UK, Medical Research Council, UK Research & Innovation, Basque Health Department, and the Wellcome Trust Study. Dr. Wagner reported funding from the Medical Research Council and the Rank Prize. Dr. Gilbert is employed by the American Parkinson Disease Association. Dr. Albrecht has received grant and personal fees and nonfinancial support from Allergan, Biogen, Celgene, Ipsen, Janssen Cilag, Merck, Merz Pharmaceuticals, Novartis, Roche, and Teva, outside the submitted work. Dr. Koska reported no relevant disclosures.
A version of this article appeared on Medscape.com.