Related content

Using to Realise the Potential of OCT Data



An NHS research team required a Machine Learning Ops (MLOps) tool to enable them to annotate data derived from Optical Coherence Tomography (OCT) scans, that shows two different types of cysts: intra-retinal and sub-retinal fluid cysts.

Close up of a person's eye

The data comprised 2D images of retinas gathered from patients with diabetes and fluid leak in the center of the retina (cystic macular oedema). Prior to working with, the research team had tried several home-built solutions for annotating their data – these did not always work and were not always accessible to the whole team.

The aim of the OCT data project is to create a ground-truth annotated dataset that would be utilised for the future development of fully automated AI tool capable of segmenting and quantifying different features in retina images and scans, and to assess whether these features are distributed differently based upon different diseases and illnesses and its potential role in predicting treatment outcomes in the future.


The team of medical professionals is led by Consultant Ophthalmologist and Vitreoretinal Surgeon, Mr Maged Habib, based at the Sunderland Eye Infirmary (part of the South Tyneside and Sunderland NHS Foundation Trust).

Mr Habib has specialised in retinal diseases for over 10 years. His work includes the study of diabetic eye disease and retinal vascular disorders as well as the surgical management of macular diseases and retinal disorders. Mr Habib is currently an Honorary Clinical Senior Lecturer at the Biosciences Institute at Newcastle University and his research interest is in the development of automated algorithms for the analysis of retinal images and OCT scans and assessment of novel biomarkers that help prediction of disease progression and visual outcomes.

Original image of an OCT scan showing a retina
The original image which shows the leak and cysts in the OCT scan for a patient with diabetic macular oedema.


The team is currently using’s ANNOTATE interface, part of the platform.

Using ANNOTATE, the team started by annotating a small sample data set of OCT images in order to compare and assess the correlation of feature identification and annotation made between different team members. Armed with this knowledge and a user-friendly introduction to’s MLOps tool, Mr Habib’s team are moving to annotating and curating larger data sets using the full platform. By using the platform the team are able to create highly curated, well annotated datasets that their AI collaborators can use to train new AI models.

Annotated OCT scan using's tool ANNOTATE
An OCT scan annotated using’s platform.


The annotations carried out with’s platform will form a significant step for the research group, which seeks to develop automated algorithms for early prediction of disease progression and treatment outcomes for different retinal and macular diseases. To this end, the team is creating the manual ground truth against which the performance of the automated tools will be measured.

In addition, Mr Habib’s work has played a crucial role in the testing of our ANNOTATE interface, ensuring that is providing intuitive and user-friendly products that really embed clinicians and other domain experts into the AI development cycle.


“My first impressions of the tool are great – it is really user friendly and easy to use and every button is self explanatory. It is much easier than previous tools I used that were not as interactive and friendly like this one.” 

Mr Maged Habib


The number of adults with diabetes worldwide has increased from 108 million to 422 million in the period 1980-2014. The number of affected adults worldwide is expected to rise to 592 million by 2035. About 25% of people with diabetes have some form of changes at the retina at the back of the eye (diabetic retinopathy).

Diabetic Retinopathy is one of the leading causes of blindness for working-aged adults in the United Kingdom. Diabetic macular oedema (DMO) represents the main cause of visual impairment in diabetic patients, with a prevalence of 7% and with approximately 50% of affected patients suffering from significant visual loss in 2 years if left untreated.

DMO is the accumulation of extracellular fluid in the retina secondary to inner retinal blood barrier breakdown associated with diabetes and is associated with several morphological changes in the macular retina. The fluid accumulates inside the retinal tissues (intraretinal fluid) as well as below the retina (sub retinal fluid).