Imaging and predictive modelling of proliferative vitreoretinopathy V1

  • Research type

    Research Study

  • Full title

    Identification of imaging biomarkers and predictive modelling of proliferative vitreoretinopathy using deep learning.

  • IRAS ID

    360601

  • Contact name

    Mahiul MK Muqit

  • Contact email

    Mahi.Muqit1@nhs.net

  • Sponsor organisation

    University College London

  • Clinicaltrials.gov Identifier

    Z6364106/2025/07/145, UCL DPO

  • Duration of Study in the UK

    1 years, 0 months, 4 days

  • Research summary

    Proliferative vitreoretinopathy (PVR) complicating rhegmatogenous retinal detachments (RRD) remain a significant cause of surgical failure despite advances in surgical instrumentation and techniques. However, despite major advances in freely accessible and non-invasive imaging techniques in the recent years, we have yet to identify if imaging biomarkers can help predict PVR in day-to-day clinical practice, therefore help tailor patients’ treatments and improve outcomes for them.

    This is a PhD study to investigate imaging biomarkers of PVR from multimodal imaging techniques and develop deep learning models to predict PVR in collaboration with an artificial intelligence (AI) expert team at UCL Institute of Ophthalmology.

    This study will use anonymised imaging obtained from two separate studies:
    1. Cohort-MMC - phase 1 dose-finding trial called MORPH-1 as part of its ancillary studies to study 'complicated' RRD due to PVR.
    2. Cohort-NHS - control set of imaging obtained from patients with uncomplicated ‘simple’ RRD.
    2. Cohort-fellow eye - another control set of fellow eyes of participants without RD or PVR.

    Above imaging will be obtained at Moorfields Eye Hospital Clinical research Facility and anonymised imaging data will be analysed at UCL.

  • REC name

    Wales REC 4

  • REC reference

    26/WA/0092

  • Date of REC Opinion

    10 Apr 2026

  • REC opinion

    Favourable Opinion