A dataset comprising close to 1 million images linked to patient clinical records has been published by Moorfields and UCL with hopes to support the development of AI tools and research on anterior segment conditions including cataracts.
Moorfields states this type of eye condition is a leading cause of blindness, however less than ten percent of ophthalmic imaging datasets currently include them. The CADMUS dataset reportedly holds 945,243 images from 22,482 unique patients, collected from visits to Moorfields between December 2019 and September 2024. Data from follow-up visits allows the study of disease progression and long-term outcomes, the trust notes.
Data includes raw DICOM images from the MS-39 anterior segment OCT tomographer; derived quantitative indices including keratometry values, pachymetry, wavefront aberrometry, and AI-generated classifier scores for keratoconus and related conditions; and linked EHR data covering information such as demographics and diagnoses.
Researchers can apply for access to the dataset through INSIGHT, Moorfields’ Eye and Oculomics Health Data Research Hub, with the trust hoping it can help support research into earlier disease detection, surgical outcome prediction, and the development of AI tools.
Moorfields added: “Researchers wishing to access CADMUS can do so through INSIGHT’s established Data Use Application process, which includes oversight from an independent patient and public advisory board, and applies the internationally recognised “Five Safes” framework evaluating safe projects, safe people, safe data, safe settings, and safe outputs.”
Shafi Balal, lead author, highlights that use of the data is already producing promising results, being used to establish “precision limits” for keratoconus progression measurement, and providing a basis for understanding disease progression. “We have also trained deep learning models on CADMUS. One model can predict patient age and biological sex from anterior segment scans, demonstrating that routine clinical images carry rich biological signals invisible to the human eye,” Balal shares.
Find the CADMUS datasheet in Ophthalmology Science here: https://doi.org/10.1016/j.xops.2026.101203
Wider trend: Health data
The European Innovation Council and SMEs Executive Agency has announced three winners following an open call for innovative projects looking to make health data “usable, interoperable, and clinically meaningful across fragmented systems”. The Regional Innovation Valley project UNITE is funded by the EU to further European digital health innovation and collaboration. Its first open call, centred on sharing health data and personalising remote care, reportedly prompted engagement from more than 1,000 organisations, and received a total of 19 proposals across universities, startups, hospitals, and healthcare providers. Three projects were ultimately selected for funding, and will enter an implementation phase in Spring 2026.
London-based social enterprise management consultancy, PPL, has launched a free-to-use online tool designed to support neighbourhood health across the capital, bringing a wealth of data together to identify local population trends and needs. The London Neighbourhood Public Data Explorer tool enables users to view and compare neighbourhoods on a map, explore indicators across different themes, and export information in PDF format for use elsewhere.
Cambridge University Hospitals NHS Foundation Trust has welcomed the launch of the Zenith supercomputer funded by the Department for Science, Industry and Technology. Said to offer researchers the ability to study health data on an “unprecedented scale”, the supercomputer will also support and inform the development of AI tools for patient care across the NHS. The AI Centre for Value-Based Healthcare is reportedly collaborating with Zenith on the project to help ensure the supercomputer’s power is deployed “responsibly and securely”.
