Age-Related Macular Degeneration in Ethnically Diverse Australia: Melbourne Collaborative Cohort Study

Development and validation of a deep-learning algorithm for the detection of neovascular age-related macular degeneration from colour fundus photographs

IMPORTANCE

Detection of early onset neovascular age-related macular degeneration (AMD) is critical to protecting vision.

BACKGROUND

To describe the development and validation of a deep-learning algorithm (DLA) for the detection of neovascular age-related macular degeneration.

DESIGN

Development and validation of a DLA using retrospective datasets.

PARTICIPANTS

We developed and trained the DLA using 56 113 retinal images and an additional 86 162 images from an independent dataset to externally validate the DLA. All images were non-stereoscopic and retrospectively collected.

METHODS

The internal validation dataset was derived from real-world clinical settings in China. Gold standard grading was assigned when consensus was reached by three individual ophthalmologists. The DLA classified 31 247 images as gradable and 24 866 as ungradable (poor quality or poor field definition). These ungradable images were used to create a classification model for image quality. Efficiency and diagnostic accuracy were tested using 86 162 images derived from the Melbourne Collaborative Cohort Study. Neovascular AMD and/or ungradable outcome in one or both eyes was considered referable.

MAIN OUTCOME MEASURES

Area under the receiver operating characteristic curve (AUC), sensitivity and specificity.

RESULTS

In the internal validation dataset, the AUC, sensitivity and specificity of the DLA for neovascular AMD was 0.995, 96.7%, 96.4%, respectively. Testing against the independent external dataset achieved an AUC, sensitivity and specificity of 0.967, 100% and 93.4%, respectively. More than 60% of false positive cases displayed other macular pathologies. Amongst the false negative cases (internal validation dataset only), over half (57.2%) proved to be undetected detachment of the neurosensory retina or RPE layer.

CONCLUSIONS AND RELEVANCE

This DLA shows robust performance for the detection of neovascular AMD amongst retinal images from a multi-ethnic sample and under different imaging protocols. Further research is warranted to investigate where this technology could be best utilized within screening and research settings.

Ethnic Differences in the Association Between Age-Related Macular Degeneration and Vision-Specific Functioning

Importance

Understanding the link between ethnicity and health is critical to making appropriate public policy decisions. Few population-level data are available about this connection, however, including the influence of ethnicity on the association between age-related macular degeneration (AMD) and vision-specific functioning (VSF).

Objective

To identify the influence of ethnicity on VSF among Chinese, Malay, and Indian patients with AMD.

Design, Setting, and Participants

This cross-sectional, population-based study relied on patients and their data from 3 population-based studies in 3 ethnic groups: Chinese, Malay and Indian. Of 10 033 Chinese, Malay, and Indian adults who participated in the study, 9962 (99.3%) who had gradable fundus images and Visual Function Index (VF-11) data available were included in the analyses for the present study. Uniocular presenting distance visual acuity was measured using the logMAR chart. Separate multiple linear regression models examined the association between AMD and VSF in the 3 ethnic groups, adjusting for age, sex, presenting visual acuity in the better-seeing eye, educational level, income, smoking status, hypertension, diabetes, cardiovascular disease, total cholesterol level, and other eye conditions. Data were collected between January 20, 2004, and December 19, 2011; data analysis was conducted between November 12, 2015, and December 28, 2016.

Exposures

Age-related macular degeneration according to fundus photographs graded using a modified Wisconsin Age-Related Maculopathy Grading System.

Main Outcomes and Measures

Rasch analysis was used to convert VF-11 questionnaire scores to estimated interval measures of VSF.

Results

Of the 9962 participants, the mean (SD) age was 58.8 (10.4) years; 4909 (49.3%) were male; 590 (5.9%) had early AMD (241 Chinese, 161 Malays, and 188 Indians) and 60 (0.6%) had late AMD (25 Chinese, 21 Malays, and 14 Indians). In the adjusted models, compared with no AMD, early AMD was associated with a small reduction in VSF (2.9%; β = -0.12; 95% CI, -0.23 to -0.00; P = .046) in the Chinese group but not in the Indian and Malay groups. Moreover, Chinese participants with late AMD had a clinically significant 19.1% loss of VSF (β = -0.78; 95% CI, -1.13 to -0.43, P < .001). In the Malay group, those with late AMD had a 13.5% drop in VSF (β = -0.49; 95% CI, -1.01 to 0.04; P = .07) compared with their counterparts without AMD. Similarly, late AMD was not associated with VSF in the Indian group.

Conclusions and Relevance

Early and late AMD negatively affected VSF in Chinese but not in Indian and Malay participants. This finding suggests that there is an independent ethnic influence in the association of the disease with VSF in multiethnic Asian populations, thus warranting ethnicity-based strategies to delay the onset or progression of AMD.

Reticular Pseudodrusen and Their Association with Age-Related Macular Degeneration: The Melbourne Collaborative Cohort Study

PURPOSE

To determine the prevalence of reticular pseudodrusen (RPD) and its association with age-related macular degeneration (AMD) and AMD risk factors in a large sample.

DESIGN

Community-based cohort study in Melbourne, Victoria, Australia.

PARTICIPANTS

A total of 21,130 participants 48 to 86 years of age available for ophthalmic assessment at follow-up from 2003 through 2007.

METHODS

Lifestyle, diet, and anthropometric measurements were obtained at baseline and follow-up. At follow-up, digital macular color photographs were graded for early, intermediate, and late AMD as well as the presence of RPD. Data were analyzed using multinomial logistic regression controlling for age, gender, smoking, country of birth, and diet.

MAIN OUTCOME MEASURES

Detection of RPD based on color fundus photographs.

RESULTS

Prevalence of RPD was 0.41% (87 of 21,130 participants), with 51% having bilateral RPD. Patients with RPD were older compared with patients with large drusen (>125 μm; 76±4 vs. 68±9 years; P < 0.001). Increasing age, female gender, being a current smoker, as well as focal pigmentary abnormalities and large drusen (>125 μm) were associated with a higher prevalence of RPD. Presence of geographic atrophy (GA) was associated with the highest odds of having RPD (odds ratio [OR], 153; 95% confidence interval [CI], 53-442), followed by choroidal neovascularization (CNV; OR, 90; 95% CI, 26-310), intermediate AMD (OR, 33; 95% CI, 14-77), and early AMD (OR, 12; 95% CI, 5-31) compared with those with no AMD. The ARMS2 single nucleotide polymorphism (SNP) rs10490924, HTRA1 SNPs rs11200638 and rs3793917, and CFH SNPs rs393955, rs1061170, and rs2274700 were associated with increased prevalence of RPD (all P < 0.05).

CONCLUSIONS

Reticular pseudodrusen are highly concurrent with AMD and have similar associations with known AMD risk factors such as age, gender, smoking, and genetic risk factors. Reticular pseudodrusen are associated more strongly with GA than with CNV. Although RPD are not specific to AMD, they are likely to be a strong risk factor for progression to late-stage AMD, similar to focal pigmentary abnormalities and large drusen.