A deep learning model for glaucoma diagnosis based on optical coherence tomography (OCT) : a multicenter study
- Conditions
- OCT data of glaucoma patients and normal population for developing deep learning modelglaucomadeep learning modeloptical coherence tomography
Recruitment & Eligibility
- Status
- Pending (Not yet recruiting)
- Sex
- All
- Target Recruitment
- 500
Inclusion Criteria
1. Patients with age over 18 years old
2. Patients who diagnosed glaucoma 400 eyes
3. Normal subject without glaucoma 100 eyes
4. Each subject should have OCT result and HVF 24-2 result within the same period (not more than 6 months apart)
Exclusion Criteria
1. Patients with ocular comorbidities which obscure OCT data gathering
2. OCT signal strength less than 6
3. Patient with other comorbidities which interfere HVF 24-2 result e.g. retinal disease or neuro-ophthalmologic disease
4. Unreliable HVF result (fixation loss > 20%, false positive > 20%, false negative > 20%, rim artifact, lid effect, clover leaf)
Study & Design
- Study Type
- Observational
- Study Design
- Not specified
- Primary Outcome Measures
Name Time Method efficacy of the model end of study AUC, true positive, true negative, sensitivity, specificity
- Secondary Outcome Measures
Name Time Method correlation between predicted outcome from the model based on OCT versus HVF 24-2 result end of study comparison of visual field index (VFI) using independent sample T-Test