Prediction of systemic biomarkers from retinal photographs: development and validation of deep-learning algorithms.

BACKGROUND: The application of deep learning to retinal photographs has yielded promising results in predicting age, sex, blood pressure, and haematological parameters. However, the broader applicability of retinal photograph-based deep learning for predicting other systemic biomarkers and the generalisability of this approach to various populations remains unexplored. METHODS: With use of 236 257 retinal photographs from seven diverse Asian and European cohorts (two health screening centres in South Korea, the Beijing Eye Study, three cohorts in the Singapore Epidemiology of Eye Diseases study, and the UK Biobank), we evaluated the capacities of 47 deep-learning algorithms to predict 47 systemic biomarkers as outcome variables, including demographic factors (age and sex); body composition measurements; blood pressure; haematological parameters; lipid profiles; biochemical measures; biomarkers related to liver function, thyroid function, kidney function, and inflammation; and diabetes. The standard neural network architecture of VGG16 was adopted for model development. FINDINGS: In addition to previously reported systemic biomarkers, we showed quantification of body composition indices (muscle mass, height, and bodyweight) and creatinine from retinal photographs. Body muscle mass could be predicted with an R2 of 0·52 (95% CI 0·51-0·53) in the internal test set, and of 0·33 (0·30-0·35) in one external test set with muscle mass measurement available. The R2 value for the prediction of height was 0·42 (0·40-0·43), of bodyweight was 0·36 (0·34-0·37), and of creatinine was 0·38 (0·37-0·40) in the internal test set. However, the performances were poorer in external test sets (with the lowest performance in the European cohort), with R2 values ranging between 0·08 and 0·28 for height, 0·04 and 0·19 for bodyweight, and 0·01 and 0·26 for creatinine. Of the 47 systemic biomarkers, 37 could not be predicted well from retinal photographs via deep learning (R2≤0·14 across all external test sets). INTERPRETATION: Our work provides new insights into the potential use of retinal photographs to predict systemic biomarkers, including body composition indices and serum creatinine, using deep learning in populations with a similar ethnic background. Further evaluations are warranted to validate these findings and evaluate the clinical utility of these algorithms. FUNDING: Agency for Science, Technology, and Research and National Medical Research Council, Singapore; Korea Institute for Advancement of Technology.

Areas and factors associated with patients' dissatisfaction with glaucoma care.

OBJECTIVE: The aim of this study was to evaluate patients' dissatisfaction with overall and specific aspects of a tertiary glaucoma service and to determine their independent factors, including intraocular pressure (IOP) and visual acuity (VA). METHODS: Patients, aged ≥21 years, from a specialist glaucoma service in a tertiary eye hospital in Singapore for at least 6 months, were recruited for this cross-sectional study between March and June 2014. All consenting patients completed a 7-area glaucoma-specific satisfaction questionnaire and one item related to satisfaction with overall glaucoma care. We determined the top three areas of dissatisfaction and overall dissatisfaction with the glaucoma service. We also explored the independent factors associated with overall and specific areas of patients' dissatisfaction with their glaucoma care, including VA and IOP by using logistic regression models. RESULTS: Of the 518 patients recruited, 438 (84.6%) patients completed the study. Patients' dissatisfaction with the overall glaucoma service was 7.5%. The three areas of glaucoma service with the highest dissatisfaction rates were as follows: 1) explanation of test results (24.8%); 2) explanation of glaucoma complications (23.7%); and 3) advice on managing glaucoma (23.5%). Patients who were dissatisfied with the overall service had a worse mean VA compared with satisfied patients (logarithm of the minimum angle of resolution =0.41±0.43 vs 0.27±0.49, p=0.005), whereas mean IOP remained well-controlled in both the groups (13.55±2.46 mmHg vs 14.82±2.86 mmHg, p=0.014). In adjusted models, factors associated with overall dissatisfaction with glaucoma care included a pre-university education and above (odds ratio [OR] =8.06, 95% CI =1.57-41.27) and lower IOP (OR =0.83, 95% CI =0.71-0.98). CONCLUSION: Although less than one tenth of glaucoma patients were dissatisfied with the overall glaucoma service, one in four patients were dissatisfied with three specific aspects of care. A lower IOP, ironically, and education level were associated with overall dissatisfaction. Improving patients' understanding of glaucoma test results, glaucoma complications, and disease management may increase patient satisfaction levels.