Trial Review

Did you know?

The ANZCTR now automatically displays published trial results and simplifies the addition of trial documents such as unpublished protocols and statistical analysis plans.

These enhancements will offer a more comprehensive view of trials, regardless of whether their results are positive, negative, or inconclusive.

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
ACTRN12619000398101
Ethics application status
Approved
Date submitted
27/02/2019
Date registered
12/03/2019
Date last updated
14/03/2022
Date data sharing statement initially provided
12/03/2019
Type of registration
Prospectively registered

Titles & IDs
Public title
A study to assess whether an Artificial Intelligence (AI) technology can be used to assess images of suspicious skin lesions to determine the likelihood of skin cancer; and reduce the number, and cost assocated with, biopsies of skin lesions in primary care
Scientific title
Assessing the effectiveness of an Artificial Intelligence (AI) algorithm to identify skin cancer lesions: The Deep Ensemble for Recognition of Malignancy (DERM) in Primary Care Study
Secondary ID [1] 297643 0
Nil known
Universal Trial Number (UTN)
1111-1229-3233
Trial acronym
Linked study record

Health condition
Health condition(s) or problem(s) studied:
Skin Cancer 311773 0
Condition category
Condition code
Skin 310386 310386 0 0
Dermatological conditions
Cancer 310503 310503 0 0
Malignant melanoma
Cancer 310504 310504 0 0
Non melanoma skin cancer

Intervention/exposure
Study type
Interventional
Description of intervention(s) / exposure
All participants will have their suspect skin lesion(s) photographed on a smartphone camera with a dermosopic lens attachment by their General Practitioner (GP). The process should take 2-3 minutes of their standard GP visit.
The GP will decide whether to biopsy each lesion. The images will then be analysed by an Artificial Intelligence (AI)-based tool, Deep Ensemble for Recognition of Malignancy (DERM) in real time, and immediately (within 1 minute) return a response to the GP. Where the GP has decided to biopsy the lesion, the response will be for the GP to continue as planned. Where the GP decided not to biopsy a lesions, but DERM recommends this, the GP may decide to change the patient management plan. The final decision on whether to biopsy each lesion is made by the GP (in agreement with the patient).
Intervention code [1] 313782 0
Diagnosis / Prognosis
Comparator / control treatment
No control group
Control group
Uncontrolled

Outcomes
Primary outcome [1] 319267 0
Number of lesions GP chose to biopsy per skin cancer diagnosed, as captured in the patient record held by the GP
Timepoint [1] 319267 0
During GP appointment
Primary outcome [2] 319361 0
Number of biopsies recommended by DERM per skin cancer diagnosed, as captured by the DERM system
Timepoint [2] 319361 0
During GP appointment
Secondary outcome [1] 367514 0
Healthcare system costs, or cost-savings, associated with usage of DERM in primary care setting by calculating difference between resource use and costs from GP medical records
Timepoint [1] 367514 0
3 month post enrolment
Secondary outcome [2] 367515 0
Sensitivity of DERM to detect skin cancer when the biopsy result used as the gold standard
Timepoint [2] 367515 0
During GP appointment
Secondary outcome [3] 367516 0
Sensitivity of GP to detect skin cancer when the biopsy result used as the gold standard
Timepoint [3] 367516 0
During GP appointment
Secondary outcome [4] 367518 0
The concordance between GP decision to biopsy, as recorded in the medical records, and DERM recommendation to biopsy as captured by the DERM system
Timepoint [4] 367518 0
3 month post enrolment
Secondary outcome [5] 367519 0
Number of skin lesions identified as changed over time, as assessed by DERM
Timepoint [5] 367519 0
During GP appointment, 3 months post enrolment
Secondary outcome [6] 367520 0
Proportion of poor quality images, as determined by the number of images that fail the image quality check step in DERM
Timepoint [6] 367520 0
During GP appointment, 3 months post enrolment
Secondary outcome [7] 367895 0
Specificity of DERM in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [7] 367895 0
During GP appointment
Secondary outcome [8] 367896 0
False Positive Proportion of DERM in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [8] 367896 0
During GP appointment
Secondary outcome [9] 367897 0
False Negative Proportion of DERM in detecting skin cancer when the biopsy result used as the gold standard.
Timepoint [9] 367897 0
During GP appointment
Secondary outcome [10] 367898 0
Positive Predictive Value of DERM in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [10] 367898 0
During GP appointment
Secondary outcome [11] 367899 0
Negative Predictive Value of DERM in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [11] 367899 0
During GP appointment
Secondary outcome [12] 367900 0
Specificity of GP in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [12] 367900 0
During GP appointment
Secondary outcome [13] 367901 0
False Positive Proportion of GP in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [13] 367901 0
During GP appointment
Secondary outcome [14] 367902 0
False Negative Proportion of GP in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [14] 367902 0
During GP appointment
Secondary outcome [15] 367903 0
Positive Predictive Value of GP in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [15] 367903 0
During GP appointment
Secondary outcome [16] 367904 0
Negative Predictive Value of GP in detecting skin cancer when the biopsy result used as the gold standard
Timepoint [16] 367904 0
During GP appointment
Secondary outcome [17] 367905 0
Quality Adjusted Life Years (QALYs) as calculated using EQ-5D
Timepoint [17] 367905 0
During GP appointment, 3 months post enrolment

Eligibility
Key inclusion criteria
Adult patients (18 years or older) who attend a primary care practice and have a clinical assessment of at least one suspicious skin lesion that is suitable for photographing (<15mm, in an anatomical site unsuitable for photographing, has previously been biopsied or in an area of visible scarring).
Minimum age
18 Years
Maximum age
No limit
Sex
Both males and females
Can healthy volunteers participate?
No
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
The population will be split into two for comparison of the number of biopsies undertaken:
Control group - those lesions which the GP decides to biopsy, irrespective of DERM recommendation
Intervention group - those lesions which the GP would not have biopsied without DERM recommendation (irrespective of whether these are then biopsied)
We expect that 46.3% of lesions examined by GPs will be biopsied, and that 16.7% of lesions will be skin cancer postitive. In order to show that DERM reduces the percentage of unnessessarily biopsied lesions by one quarter (from 19.4% according to GP recommendation to 14.5% according to DERM recommendation), with 80% power, a sample size of 426 disease negative lesions are required. Consequently we will need to study 514 lesions.
Allowing for a 15% loss to follow up / poor image quality, the sample size for the study is expected to be between 750 and 1,000 patients.

Recruitment
Recruitment status
Recruiting
Date of first participant enrolment
Anticipated
31/03/2022
Actual
25/02/2022
Date of last participant enrolment
Anticipated
30/09/2022
Actual
Date of last data collection
Anticipated
31/12/2022
Actual
Sample size
Target
750
Accrual to date
1
Final
Recruitment in Australia
Recruitment state(s)
QLD

Funding & Sponsors
Funding source category [1] 302104 0
Commercial sector/Industry
Name [1] 302104 0
Skin Analytics Pty
Country [1] 302104 0
Australia
Primary sponsor type
Commercial sector/Industry
Name
Skin Analytics Pty
Address
C/- 126a Herbert Street
Doubleview
WA6018
Country
Australia
Secondary sponsor category [1] 301933 0
None
Name [1] 301933 0
Address [1] 301933 0
Country [1] 301933 0
Other collaborator category [1] 280576 0
University
Name [1] 280576 0
Griffiths University
Address [1] 280576 0
170 Kessels Road
Nathan
Queensland 4111
AUSTRALIA
Country [1] 280576 0
Australia

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 302782 0
Royal Australian College of General Practitioners National Research and Evaluation Ethics Committee
Ethics committee address [1] 302782 0
The Royal Australian College of General Practitioners
100 Wellington Parade, East Melbourne, VIC 3002
Ethics committee country [1] 302782 0
Australia
Date submitted for ethics approval [1] 302782 0
09/11/2018
Approval date [1] 302782 0
26/05/2020
Ethics approval number [1] 302782 0
18-016

Summary
Brief summary
The purpose of this study is to investigate the impact of an Artificial Intelligence (AI) algorithm - Deep Ensemble for Recognition of Malignancy (DERM) - on skin cancer diagnosis services in primary care.

Who is it for?

You may be eligible for this study if you are over the age of 40 and are visiting your GP with at least one suspicious skin lesion that is suitable for photographing.

Study details

All participants in this study will have a photograph taken of their suspicious lesion(s) during their regular GP consult, which DERM will analyse. The GP will decide whether to biopsy the lesion before DERM provides a recommendation on whether the lesion should be tested further. Where DERM recommends a lesion is biopsied that the GP otherwise would not have done, the GP may choose to follow DERMs recommendation or their own assessment.

It is hoped that this research will help determine if DERM can help identify cancerous skin lesions and reduce the number of unnecessary biopsies.
Trial website
Trial related presentations / publications
Public notes

Contacts
Principal investigator
Name 91386 0
Prof Scott Kitchener
Address 91386 0
Griffith School of Medicine
Griffith University
Parklands Dr
Southport,
Queensland 4215
Country 91386 0
Australia
Phone 91386 0
+61 0458476386
Fax 91386 0
Email 91386 0
[email protected]
Contact person for public queries
Name 91387 0
Scott Kitchener
Address 91387 0
Griffith School of Medicine
Griffith University
Parklands Dr
Southport,
Queensland 4215
Country 91387 0
Australia
Phone 91387 0
+61 0458476386
Fax 91387 0
Email 91387 0
[email protected]
Contact person for scientific queries
Name 91388 0
Scott Kitchener
Address 91388 0
Griffith School of Medicine
Griffith University
Parklands Dr
Southport,
Queensland 4215
Country 91388 0
Australia
Phone 91388 0
+61 0458476386
Fax 91388 0
Email 91388 0
[email protected]

Data sharing statement
Will individual participant data (IPD) for this trial be available (including data dictionaries)?
No
No/undecided IPD sharing reason/comment
Consent does not allow for images or patient data to be shared


What supporting documents are/will be available?

No Supporting Document Provided



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.