Glaucoma diagnosis is currently based on opportunistic case finding, which makes the case
for up to 50% of patients remaining undiagnosed. Diagnostic technology has been deemed
efficient in diagnosing, but cost and (hospital) location acts as a barrier for effective
screening for this asymptomatic disease with a low population-based prevalence
Portuguese National Strategy for Visual Health, published in 2018 asks for a pilot study
aimed at a one-time intervention at the primary care setting at the age of 60 years to do
both an optic disc analysis and an intraocular pressure (IOP) measurement as a screening
system for glaucoma
Two Portuguese centers have applied for this pilot. An urban-based center (Lisbon) and a
countryside center (Minho) will conduct an invitation-based screening for those
registered in their global primary care area. Age range was increased to 55 to 65 to
capture a spectrum of data, enabling to later refine the target population.
Screening Intervention will be the same in both centers. Both will assess reasons for
undergoing or rejecting screening and demographic and ophthalmological-related
parameters, including glaucoma family history and a known personal glaucoma diagnosis.
IOP will be non-invasively checked by rebound tonometry and optic disc retinographies
will be performed on both centers. The Minho arm will additionally perform an optical
coherence tomography (OCT) on all subjects.
Decision to refer will be made on one (or both) two findings:
IOP of 24mmHg or higher
referrable glaucoma
The Lisbon arm will use an Artificial intelligence (AI) system to rank the retinography
findings into a binary referrable vs non-referrable system, based on a pre-defined
threshold (0.73).
The Minho arm will base their referral on the existence of optical coherence tomography
(OCT) retinal nerve fibre layer abnormality (defined as one sector thickness being
outside normal limits)
Subjects flagged as referrable will be sent to the hospital Glaucoma Clinic, where
functional and structural examinations will be performed and a standard of care clinical
decision will be made by the glaucoma expert.
A reading center, with two masked experts, will be established to determine an unbiased
ground truth for the comparison analysis. Reading center will conduct analysis on 3 tier
level:
First round will include masked fundus pictures of all recruited subjects from both
centers (both positive and negative referrals).
Second Round will include all subjects from Minho arm, who will have both fundus picture
and OCT data.
Third round will include data analysis from all positive referrals from both arms, which
includes all clinical data from the CRF plus all hospital exams (excluding clinical
impression).
Outcomes of the reading center will be twofold:
"Glaucoma diagnosis label"
"normal"
"glaucoma suspect"
"definite glaucoma" (which implies a glaucomatous visual field) 2) Score (0-100) of
likeliness of definitive glaucoma Disagreements on any category will be solved by a
third masked referee (in the case of the scoring system, a difference larger than
10% would be considered a disagreement)