Trial Review

The safety and scientific validity of this study is the responsibility of the study sponsor and investigators. Listing a study does not mean it has been endorsed by the ANZCTR. Before participating in a study, talk to your health care provider and refer to this information for consumers
Trial registered on ANZCTR


Registration number
ACTRN12621001153808
Ethics application status
Approved
Date submitted
11/05/2021
Date registered
26/08/2021
Date last updated
26/08/2021
Date data sharing statement initially provided
26/08/2021
Type of registration
Retrospectively registered

Titles & IDs
Public title
Artificial intelligence system for opportunistic screen of common eye diseases
Scientific title
Evaluating an automated system for screening common eye diseases by integration of retinal photography and artificial intelligence in adults visiting endocrinology and primary care clinics
Secondary ID [1] 304175 0
Nil Known
Universal Trial Number (UTN)
U1111-1259-8284
Trial acronym
Linked study record

Health condition
Health condition(s) or problem(s) studied:
diabetic retinopathy 319439 0
glaucoma 319440 0
age-related macular degeneration 319441 0
Condition category
Condition code
Eye 317411 317411 0 0
Diseases / disorders of the eye
Public Health 317413 317413 0 0
Epidemiology

Intervention/exposure
Study type
Interventional
Description of intervention(s) / exposure
An opportunisitc eye screening will be provided to all participants. In the waiting area of out-patient clinics, prior to seeing their non-eye physicians, eligible participants will be invited to perform opportunistic eye screening, using an automated retinal camera coupled with artificial intelligence (AI) technology. Participants who enrol in the study will perform retinal imaging on their own. With a flash illumination, fundus photo will be taken automatically without pupil dilatation. Based on the colour fundus images, the integrated AI technology will generate report on risk of glaucoma, diabetic retinopathy, and age-related macular degeneration, as well as referral recommendations. The system is operator-free and requires no prior training from participants. The automated voice instruction will guide the participants to complete the screening without help from staffs or technicians. This screening will take approximately 5 minutes to complete. No identifiable information will be entered or stored in the screening device. A QR code and unique study ID will be printed from the screening device and provided to the participants. By scanning the QR code, a full report can be available for review on participant's mobile phone, including the retinal images, the graded risk of the three eye diseases and referral recommendations. When referral to eye clinician is necessary, a referral letter will be provided by the endocrinology or general practice clinics. This multi-site study is expected to screen 500 participants in six months.
Intervention code [1] 318856 0
Early detection / Screening
Comparator / control treatment
All retinal images will be reviewed and manually re-graded by at least two, and up to three eye clinicians. This screening report by eye clinicians will serve as a standard reference to evaluate the accuracy of the AI screening report. More specifically if the first two human graders reach consensus findings, this will be used as a standard reference. Any disagreement between the first two graders will be adjudicated by a third grader.
When the AI grading differs from human grading reports, participants will be contacted to notify such discrepancy and the reports by eye clinicians. Participants will be advised to follow up with their own eye clinicians should the AI screening report be found as a false-negative.
Control group
Active

Outcomes
Primary outcome [1] 325450 0
The diagnostic accuracy of the AI screening system for diabetic retinopathy will be measured, including sensitivity, specificity, accuracy, positive/nagative predictive values and area under ROC curve. For this purpose, the AI grading report for the risk of diabetic retinopathy will be compared with the gold standard assessment by eye clinicians using the same retinal photographs.
Timepoint [1] 325450 0
At the conclusion of the study
Primary outcome [2] 328005 0
The diagnostic accuracy of the AI screening system for glaucoma will be measured, including sensitivity, specificity, accuracy, positive/nagative predictive values and area under ROC curve. For this purpose, the AI grading report for the risk of glaucoma will be compared with the gold standard assessment by eye clinicians using the same retinal photographs.
Timepoint [2] 328005 0
At the conclusion of the study
Primary outcome [3] 328006 0
The diagnostic accuracy of the AI screening system for age-related macular degeneration will be measured, including sensitivity, specificity, accuracy, positive/nagative predictive values and area under ROC curve. For this purpose, the AI grading report for the risk of age-related macular degeneration will be compared with the gold standard assessment by eye clinicians using the same retinal photographs.
Timepoint [3] 328006 0
At the conclusion of the study
Secondary outcome [1] 395252 0
The feasibility and patient experience of the DIY screening system. This is assessed by collecting responses from patient participants to a questionnaire that is specifically designed for this study. The survey aims to understand the patient satisfaction of the screening process.
Timepoint [1] 395252 0
at the time of using the DIY screening kiosk
Secondary outcome [2] 397223 0
The feasibility and clinicians' experience of the DIY screening system. This is assessed by collecting responses from clinicians to a questionnaire that is specifically designed for this study. The survey aims to understand the overall clinicians' satisfaction of the screening process.
Timepoint [2] 397223 0
At the conclusion of the study
Secondary outcome [3] 397224 0
the impact of the DIY screening system in detecting diabetic retinopathy. This is measured by the rate of true positive diagnosis of diabetic retinopathy, as validated by human graders, among all participants being screened.
Timepoint [3] 397224 0
At the conclusion of study
Secondary outcome [4] 397225 0
the impact of the DIY screening system in glaucoma. This is measured by the rate of true positive diagnosis of glaucoma, as validated by human graders, among all participants being screened.
Timepoint [4] 397225 0
At the conclusion of study
Secondary outcome [5] 397226 0
the impact of the DIY screening system in age-related macular degeneration. This is measured by the rate of true positive diagnosis of age-related macular degeneration, as validated by human graders, among all participants being screened.
Timepoint [5] 397226 0
At the conclusion of study

Eligibility
Key inclusion criteria
1. with diabetes and at least 18 years old
2. without diabetes and at least 50 years old
Minimum age
18 Years
Maximum age
No limit
Sex
Both males and females
Can healthy volunteers participate?
Yes
Key exclusion criteria
None

Study design
Purpose of the study
Diagnosis
Allocation to intervention
Non-randomised trial
Procedure for enrolling a subject and allocating the treatment (allocation concealment procedures)
Methods used to generate the sequence in which subjects will be randomised (sequence generation)
Masking / blinding
Open (masking not used)
Who is / are masked / blinded?



Intervention assignment
Single group
Other design features
Phase
Not Applicable
Type of endpoint/s
Efficacy
Statistical methods / analysis
diagnostic accuracy will be assessed by comparing the AI grading results against two independent eye clinicians. Any disagreement between the two clinicians will be ajudicated by a third one, which will serve as a gold standard reference for the AI screening results. The diagnostic accuracy include sensitivity, specificity, accuracy, positive/negative predictive values, and area under ROC curve.

Assessment of consumer acceptability will be determined by the positive response rate/screening rate. The responses to the 5-point Likert scale question will be analysed using the “document variable statistics” function in MAXQDA software.

Clinical impact of the kiosk is assessed as the detection rate, which is defined as the proportion of newly diagnosed eye disease (diabetic retinopathy, glaucoma, and age-related macular degeneration) divided by the total number of eligible individuals.

Recruitment
Recruitment status
Recruiting
Date of first participant enrolment
Anticipated
Actual
2/08/2021
Date of last participant enrolment
Anticipated
31/10/2022
Actual
Date of last data collection
Anticipated
31/10/2022
Actual
Sample size
Target
600
Accrual to date
55
Final
Recruitment in Australia
Recruitment state(s)
VIC
Recruitment hospital [1] 19397 0
St Vincent's Hospital (Melbourne) Ltd - Fitzroy
Recruitment hospital [2] 19398 0
Box Hill Hospital - Box Hill
Recruitment hospital [3] 20344 0
Endocrinology & Diabetes Specialist Centre - Box Hill
Recruitment hospital [4] 20347 0
Casey Superclinic - Berwick
Recruitment hospital [5] 20348 0
Camberwell Road Medical Practice - Hawthorn East
Recruitment postcode(s) [1] 33978 0
3065 - Fitzroy
Recruitment postcode(s) [2] 35101 0
3128 - Box Hill
Recruitment postcode(s) [3] 35104 0
3128 - Box Hill Central
Recruitment postcode(s) [4] 35109 0
3806 - Berwick
Recruitment postcode(s) [5] 35110 0
3123 - Hawthorn East

Funding & Sponsors
Funding source category [1] 306996 0
Government body
Name [1] 306996 0
National Health and Medical Research Council (NHMRC)
Country [1] 306996 0
Australia
Primary sponsor type
Other
Name
Centre for Eye Research Australia
Address
Level 7, 32 Gisborne Street
EAST MELBOURNE VIC 3002
Country
Australia
Secondary sponsor category [1] 309431 0
None
Name [1] 309431 0
Address [1] 309431 0
Country [1] 309431 0

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 307128 0
St Vincent's Hospital (Melbourne) HREC
Ethics committee address [1] 307128 0
Research Governance Unit
St Vincent’s Hopital Melbourne, 41 Victoria Parade, Fitzroy VIC 3065
Ethics committee country [1] 307128 0
Australia
Date submitted for ethics approval [1] 307128 0
09/08/2020
Approval date [1] 307128 0
02/12/2020
Ethics approval number [1] 307128 0
HREC/64640/SVHM-2020

Summary
Brief summary
Currently about 50% of patients living with diabetic retinopathy remain unaware they have the disease. Similarly, large amount of glaucoma and age-related macular degeneration cases remain undetected. To improve the detection rate. a novel system of automated fundus photography coupled with artificial intelligence (AI) will be evaluated in real world clinics to provide opportunistic screening for diabetic retinopathy, glaucoma, and age-related macular degeneration. The device is fully automated for image acquisition, analysis and report, which negates the need for clinical staffs. No prior training is required of the users. Within 2 to 3 minutes, participants can complete their own retinal photography and obtain a grading report based on the retinal photos for the risk of the above diseases. Referral recommendations are also included. This may benefit the patients in clinical settings where eye clinicians are in short supply. This study aims to test the performance of the AI screening kiosk in real world clinical settings. The AI screening results will be compared with the gold standard assessment by three eye clinicians. We hypothesize that this AI screening system is accurate, user friendly and can help improve the detection rate of some common eye diseases.
Trial website
Trial related presentations / publications
Public notes

Contacts
Principal investigator
Name 106130 0
Prof Mingguang He
Address 106130 0
Centre for Eye Research Australia
Level 7, 32 Gisborne Street
EAST MELBOURNE VIC 3002
Country 106130 0
Australia
Phone 106130 0
+61 3 99298361
Fax 106130 0
Email 106130 0
[email protected]
Contact person for public queries
Name 106131 0
Prof Mingguang He
Address 106131 0
Centre for Eye Research Australia
Level 7, 32 Gisborne Street
EAST MELBOURNE VIC 3002
Country 106131 0
Australia
Phone 106131 0
+61 3 99298361
Fax 106131 0
Email 106131 0
[email protected]
Contact person for scientific queries
Name 106132 0
Prof Mingguang He
Address 106132 0
Centre for Eye Research Australia
Level 7, 32 Gisborne Street
EAST MELBOURNE VIC 3002
Country 106132 0
Australia
Phone 106132 0
+61 3 99298361
Fax 106132 0
Email 106132 0
[email protected]

Data sharing statement
Will individual participant data (IPD) for this trial be available (including data dictionaries)?
Yes
What data in particular will be shared?
individual participant data after deidentification, which underlie the results reported in publicaiton of the study.
When will data be available (start and end dates)?
up to 5 years after the publication of results in this study.
Available to whom?
Investigators whose proposed use of the data has been approved by an independent review
committee.
Available for what types of analyses?
For individual participant data meta-analysis
How or where can data be obtained?
A methodologically sound proposal to share data by requestors should be directed to principle investigator [email protected]. Data access agreement needs to be signed by both parties involved.


What supporting documents are/will be available?

Doc. No.TypeCitationLinkEmailOther DetailsAttachment
11648Study protocol  [email protected]



Results publications and other study-related documents

Documents added manually
No documents have been uploaded by study researchers.

Documents added automatically
No additional documents have been identified.