SVP detection using machine learning (SVP-ML)
Research type
Research Study
Full title
Automated detection of spontaneous venous pulsations within fundal videos using machine learning
IRAS ID
318322
Contact name
Timothy Jackson
Contact email
Sponsor organisation
King's College London
Clinicaltrials.gov Identifier
clinicaltrials.gov registration number, NCT05731765
Duration of Study in the UK
1 years, 0 months, 0 days
Research summary
Elevated intracranial pressure (ICP) is associated with serious neurological conditions and can lead to blindness, brain injury and death. Determining the ICP is clinically important but challenging because it requires an invasive lumbar puncture (LP) procedure to test the fluid around the brain and spine.
Raised ICP has been shown to cease spontaneous venous pulsations (SVPs), which are seen as rhythmic calibre variations occurring in the retinal veins at the back of the eye (fundus). Thus the presence of SVPs is clinically used as a sensitive marker for the exclusion of raised ICP. Prior studies have demonstrated that SVPs are identifiable in high-quality optical coherence tomography (OCT) videos, but expert interpretation is required.
Automated SVP detection would remove the need for expert video analysis and could reduce the need for invasive LPs, transforming ICP testing in multiple settings. The automated system would have wide ranging appeal in clinical and community settings given that OCT are non-invasive, rapidly capture fundal videos, and are widely available.
We aim to develop an automated, machine learning (ML) enabled detection system for SVPs using widely available OCT fundal videography and create a novel test for suspected raised ICP. Positive results from this proof-of-concept study would support future prospective clinical trials and further system development to expand the use of the automated detection system to portable imaging devices, in turn broadening ambulatory and community applications.REC name
North of Scotland Research Ethics Committee 1
REC reference
23/NS/0042
Date of REC Opinion
11 Apr 2023
REC opinion
Favourable Opinion