Individuals with diabetes Nutritional, Metabolic, Endocrine
Conditions
Interventions
The ODISC study is a observational cohort study that includes type 2 diabetes mellitus (T2DM) patients from a part of a nationwide community screening program under China’s Basic Public Health Serv
lifestyle factors such as smoking, alcohol use, physical activity
medical history and medication adherence) , health data such as blood sugar levels, kidney function, and heart health
and ophthalmic evaluation (slit-lamp biomicroscopy and digital fundus photography). The study leverages artificial intelligence (AI)-driven computational methods to extract quantitative features fro
Sponsors
State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University
Eligibility
Sex/Gender
All
Inclusion criteria
Inclusion criteria: 1. Individuals aged over 18 years. 2. T2DM diagnosis was confirmed by primary care physicians using internationally accepted thresholds: fasting plasma glucose (FPG) =7.0 mmol/L, 2-hour postprandial glucose (2h-PG) =11.1 mmol/L during a 75-g oral glucose tolerance test (OGTT), or hemoglobin A1c (HbA1c) =6.5%.
Exclusion criteria
Exclusion criteria: 1. Patients with type 1 diabetes, gestational diabetes, severe comorbidities (e.g., advanced cardiovascular disease like stage III-IV cardiovascular disease, chronic kidney disease with eGFR <30 mL/min/1.73 m², or active malignancy, etc. 2. Inability to complete ophthalmic assessments. 3. Unable to give their own informed consent.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| The 5- and 10-year risks of diabetic retinal and systemic complications (e.g., cardiovascular events, nephropathy) are predicted using oculomics mainly extracted by digital fundus photography and other clinical data (at baseline, 1-year visit, 2 year-visit, 3-year visit). | — |
Secondary
| Measure | Time frame |
|---|---|
| The distinct diabetic subtypes were identified based on retinal phenotypic patterns. DDRTree (Discriminative Dimensionality Reduction via Learning a Tree) will be applied to project high-dimensional oculomics data into a low-dimensional latent space while preserving pseudotemporal trajectory patterns. Unsupervised clustering algorithms—including k-means (hard clustering) and Gaussian mixture models (soft clustering) are also applied. | — |
Countries
China
Contacts
Public ContactWei Wang
Outcome results
None listed