Smartphone-based Retinal Imaging to Characterize Early Functional Retinal Vascular Changes in Diabetic Retinopathy
Other conference paper




Abnormal autoregulation of retinal blood flow is one of the subclinical pathophysiologic responses that precedes onset of clinical diabetic retinopathy (DR). We aim to examine changes in quantitative retinal vascular parameters from sitting to supine positions measured via smartphone-based retinal imaging in patients with diabetes.

We conducted a case-control study of 30 subjects with diabetes (20 with any DR and10 without DR) and 21 healthy volunteer controls. Each subject had retinal images capatured from a clip-on adapter lens (OphthoLens, RainsOptics Ltd, HK) over the camera and flashlight on a mobile smartphone, first in sitting and then in supine positions (Figure). Retinal vascular parameters were measured using the Singapore I Vessel Assessment (SIVA) software from the smartphone retinal images. For each parameter, the postural change was calculated as percentage change in the parameters from sitting to supine positions. Paired t-test was first performed to compare the parameters between positions in each group (normal controls, diabetics without DR, diabetics with any DR). Analyses of covariance (ANCOVA) was then performed to estimate mean postural changes across the three groups, after controlling for age, BP and IOP.

In the control group, arteriolar caliber (75.8μm vs. 88.8μm, p<0.001), venular caliber (93.0μm vs. 107.8μm, p=0.002) and arteriolar fractal dimension (1.197 vs. 1.207, p=0.006) were significantly decreased from sitting to supine positions. In contrast, in patients with diabetes, there were no significant changes in the retinal vascular parameters between the positions (all p>0.05). In the ANCOVA, changes in arteriolar caliber (p-trend=0.028), and arteriolar tortuosity (p-trend=0.021) were significantly different across three groups. Among the subjects with diabetes, the calibers of retinal arterioles were less likely to constrict, but the tortuosity of retinal arterioles were more likely to be increased from sitting to supine positions (Table). Persons with diabetes without DR also had similar alterations in retinal arterioles.

Physiological postural changes in retinal vasculature can be assessed and quantified using smartphone-based retinal imaging which may be used as a portable means to characterize early subclinical functional retinal vascular changes and thus identify diabetic individuals at risk of DR.

Layman Abstract:
Diabetic retinopathy (DR) is a common eye disease caused by diabetes that lead to progressive visual impairment and even blindness. Currently, individuals with diabetes in many countries are recommended to undergo photography of the back of the eye (i.e. retina) for DR screening. However, annual screening was found to be unnecessarily expensive. Biomarkers that can help to predict the development of DR more accurately are urged to help efficient allocation of healthcare resources, for example, those with lower risk may not need annual screening.

In this study, we propose to use smartphone-based retinal imaging to assess function of retinal vessels by analyzing retinal photographs captured at sitting and supine positions. We found that the function of retinal vessels was impaired in patients with diabetes (lack of vessel constriction and higher vessel curvature from sitting to supine positions) as measured by a software, compared with that in healthy volunteers. Such alterations were also observed in persons with diabetes without DR.

Our study demonstrated that postural changes in retinal vessels can be assessed using smartphone-based retinal imaging which may be used as a portable means to characterize early subclinical changes and thus identify diabetic individuals at risk of DR.
著者Carol Y. Cheung, Y. Wang, J. H. Liu, K. Lai, R. Chang, T. Y. Wong
會議名稱Association for Research in Vision and Ophthalmology Annual Meeting (ARVO) 2019

上次更新時間 2021-19-10 於 00:45