Key Takeaways

  • CADMUS includes 945,243 anterior segment images from 22,482 patients with linked clinical records, making it one of the largest longitudinal datasets of its kind
  • The resource captures a broad range of anterior segment diseases, imaging modalities, surgical data, and quantitative measurements to support AI development and clinical research
  • Researchers can access the anonymized dataset through the INSIGHT Health Data Research Hub under established ethical and governance frameworks

Clinical researchers at Moorfields Eye Hospital NHS Foundation Trust and the UCL Institute of Ophthalmology have introduced CADMUS, a large-scale anterior segment imaging dataset comprising more than 945,000 ophthalmic images linked to clinical records, with the goal of accelerating research into diseases affecting the front of the eye.

The dataset, published in Ophthalmology Science, addresses a longstanding gap in ophthalmic data resources.1 While anterior segment diseases—including cataract, corneal opacity, and keratoconus—are among the leading causes of visual impairment and blindness worldwide, publicly available datasets for these conditions have historically been limited in size and scope.

According to the researchers, CADMUS includes 945,243 images from 22,482 unique patients collected during routine clinical care at Moorfields Eye Hospital between December 2019 and September 2024. Approximately 96% of patients have longitudinal follow-up data, enabling researchers to study disease progression and long-term clinical outcomes.

The dataset was developed through INSIGHT, the Eye and Oculomics Health Data Research Hub at Moorfields, which links ophthalmic imaging with electronic health records. Researchers will be able to apply for access to CADMUS through INSIGHT's established data access process.

CADMUS includes more than 40,000 surgical records from over 12,000 patients, covering cataract surgery, corneal cross-linking, corneal transplantation, keratectomy, and other corneal procedures. Linked clinical records also include demographic information, diagnoses, visual acuity, refraction measurements, and systemic comorbidities such as hypertension and type 2 diabetes. The resource combines three complementary data types: raw DICOM images from the MS-39 anterior segment OCT tomographer, quantitative biometric and topographic measurements, and linked electronic health record data. Derived metrics include keratometry, pachymetry, wavefront aberrometry, and artificial intelligence-generated classifier scores for keratoconus and related conditions.

The investigators said the dataset is intended to support research in early disease detection, prediction of surgical outcomes, health equity, and the development of artificial intelligence tools suitable for clinical practice.

"Early research using CADMUS data has already produced promising results," said lead author Shafi Balal, MBBS, an ophthalmic surgeon at Moorfields Eye Hospital and NIHR doctoral fellow at UCL. "We have used the dataset to establish precision limits for keratoconus progression measurement, providing a scientifically grounded basis for defining 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."

The researchers noted that all patient data within CADMUS has been irreversibly anonymized using cryptographic methods and that the project operates under ethics approval from the West of Scotland Research Ethics Service. Access requests are reviewed through INSIGHT's Data Use Application process, which includes oversight from an independent patient and public advisory board and follows the internationally recognized Five Safes framework for data governance.

The release of CADMUS also aligns with broader efforts to expand anterior segment research resources. The European Society of Cataract and Refractive Surgeons, which supported development of the dataset, has committed to maintaining an annual catalogue of available anterior segment datasets through 2030.

Reference

Balal S, et al. CADMUS: A large-scale anterior segment imaging and clinical dataset from routine ophthalmic care. Ophthalmology Science. 2026. doi:10.1016/j.xops.2026.101203.