Background Retinal optical coherence tomography findings in Lewy body diseases and their implications for visual outcomes remain controversial. We investigated whether region‐specific thickness analysis of retinal layers could improve the detection of macular atrophy and unravel its association with visual disability in Parkinson's disease. Methods Patients with idiopathic Parkinson's disease (n = 63), dementia with Lewy bodies (n = 8), and E46K mutation carriers in the α‐synuclein gene (E46K‐SNCA) (n = 4) and 34 controls underwent Spectralis optical coherence tomography macular scans and a comprehensive battery of visual function and cognition tests. We computed mean retinal layer thicknesses of both eyes within 1‐, 2‐, 3‐, and 6‐mm diameter macular discs and in concentric parafoveal (1‐ to 2‐mm, 2‐ to 3‐mm, 1‐ to 3‐mm) and perifoveal (3‐ to 6‐mm) rings. Group differences in imaging parameters and their relationship with visual outcomes were analyzed. A multivariate logistic model was developed to predict visual impairment from optical coherence tomography measurements in Parkinson's disease, and cutoff values were determined with receiver operating characteristic analysis. Results When compared with controls, patients with dementia with Lewy bodies had significant thinning of the ganglion cell–inner plexiform layer complex within the central 3‐mm disc mainly because of differences in 1‐ to 3‐mm parafoveal thickness. This parameter was strongly correlated in patients, but not in controls, with low contrast visual acuity and visual cognition outcomes (P < .05, False Discovery Rate), achieving 88% of accuracy in predicting visual impairment in Parkinson's disease. Conclusion Our findings support that parafoveal thinning of ganglion cell–inner plexiform complex is a sensitive and clinically relevant imaging biomarker for Lewy body diseases, specifically for Parkinson's disease. © 2019 The Authors. Movement Disorders published by Wiley Periodicals, Inc. on behalf of International Parkinson and Movement Disorder Society.

No abstract