Melanoma Image Analysis Algorithm (MIAA) Validation Study

Effectiveness of an Image Analysing Algorithm to Diagnose Melanoma Compared to Gold Standard Histological Determination

Prospective, single-arm, cross-sectional, study to establish the effectiveness of MIAA to detect melanoma in pigmented lesions, compared to gold standard histological determination.

Study Overview

• Status: Completed
• Conditions: Malignant Melanoma
• Intervention / Treatment: Device: Melanoma Image Analysis Algorithm (MIAA)

Detailed Description

Skin Analytics Limited have developed an algorithm (MIAA) which reviews photographs of pigmented lesions to determine whether melanoma is likely to be present. This study aims to establish how well MIAA determines the presence or absence of melanoma, compared to a biopsy.

Pigmented lesions that a dermatologist has decided to biopsy, and are suitable for photographing, will be photographed up to five times in a single visit. Three different camera will be used, and two different dermoscopic lens attachments will be used on smartphone cameras. Images will be analysed by MIAA and the results compared to the biopsy result. Clinicians, patients and the statistical analysis team will be blinded to the result of MIAA.

At least 65 pigmented lesions positive for melanoma (as determined by biopsy) are required, which is predicted to mean 1250 patients will be recruited.

• Study Type: Interventional
• Enrollment (Actual): 514
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United Kingdom
  • Bristol, United Kingdom, BS2 8AE
    • Bristol Royal Infirmary
  • London, United Kingdom, NW3 2QG
    • Royal Free London NHS Foundation Trust
  • Oxford, United Kingdom, OX3 7LE
    • Churchill Hospital
  • Devon
    • Exeter, Devon, United Kingdom, EX2 5DW
      • Royal Devon and Exeter
  • London
    • Leytonstone, London, United Kingdom, E11 1NR
      • Whipps Corss Hospital
  • Staffordshire
    • Stoke, Staffordshire, United Kingdom, ST4 6QG
      • Royal Stoke University Hospital
  • West Midlands
    • Dudley, West Midlands, United Kingdom, DY1 2HQ
      • Russells Hall Hospital

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years and older (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

Description

Inclusion Criteria:

• Participant is willing and able to give informed consent for participation in the study,
• Male or Female, aged 18 years or above,
• Have at least 1 lesion suitable for photographing that is scheduled for biopsy to determine the presence of melanoma,
• In the Investigators opinion, able and willing to comply with all study requirements.

Exclusion Criteria:

• Any other significant disease or disorder which, in the opinion of the Investigator, may either put the participants at risk because of participation in the study, or may influence the result of the study, or the participant's ability to participate in the study.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Diagnostic
• Allocation: N/A
• Interventional Model: Single Group Assignment
• Masking: None (Open Label)

Number of Arms

1

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: All patients; Each patient will have any pigmented lesions (PLs) which are due to be biopsied, two PLs not due for biopsy and one patch of healthy skin photographed. Each will be photographed using three cameras: A standard DSLR and two smartphones with a dermoscopic lens attachment. Photographic images will be analysed by Melanoma Image Analysis Algorithm (MIAA)	Device: Melanoma Image Analysis Algorithm (MIAA); The images will be analysed by MIAA (Melanoma Image Analysis Algorithm). The MIAA result for PLs that are biopsied will be compared to the diagnosis made from the biopsy. The MIAA result for PLs not biopsied will be compared to clinician diagnosis.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
The Area Under the Curve of a Receiver Operating Characteristic (AUROC) curve of MIAA result, using a maximum likelihood estimation (MLE) from all of the available images of biopsied lesions, compared to the biopsy result	Compare the MIAA result with the biopsy result	Study completion, on average 2 weeks

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
The sensitivity of MIAA, using a MLE from all of the available images of biopsied lesions, compared to the biopsy result	%true positives, as determined by biopsy, identified	Study completion, on average 2 weeks
The sensitivity of MIAA, using a MLE from all of the available images of non-biopsied lesions, compared to clinical assessment	%true positives, as determined by clinician, identified	Study completion, on average 2 weeks
The specificity of MIAA, using a MLE from all of the available images of biopsied lesions, compared to the biopsy result	%true negatives, as determined by biopsy, identified	Study completion, on average 2 weeks
The specificity of MIAA, using a MLE from all of the available images of non-biopsied lesions, compared to clinical assessment	%true negatives, as determined by clinician, identified	Study completion, on average 2 weeks
The positive predictive value of MIAA, using a MLE from all of the available images of biopsied lesions, compared to the biopsy result	Probability that subjects with a positive MIAA result have a positive biopsy result	Study completion, on average 2 weeks
The positive predictive value of MIAA, using a MLE from all of the available images of non-biopsied lesions, compared to clinical assessment	Probability that subjects with a positive MIAA result are thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The negative predictive value of MIAA, using a MLE from all of the available images of biopsied lesions, compared to the biopsy result	Probability that subjects with a negative MIAA result have a negative biopsy result	Study completion, on average 2 weeks
The negative predictive value of MIAA, using a MLE from all of the available images of non-biopsied lesions, compared to clinical assessment	Probability that subjects with a negative MIAA result are not thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The false positive rate of MIAA, using a MLE from all of the available images of biopsied lesions, compared to the biopsy result	Probability that subject with positive MIAA result has a negative biopsy result	Study completion, on average 2 weeks
The false positive rate of MIAA, using a MLE from all of the available images of non-biopsied lesions, compared to clinical assesment	Probability that subject with positive MIAA result are not thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The false negative rate of MIAA, using a MLE from all of the available images of biopsied lesions, compared to the biopsy result	Probability that subject with negative MIAA result has a positive biopsy result	Study completion, on average 2 weeks
The false negative rate of MIAA, using a MLE from all of the available images of non-biopsied lesions, compared to clinical assessment	Probability that subject with negative MIAA result are thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The AUROC of MIAA, using images of biopsied lesions from each of the image capture apparatus combinations, compared to the biopsy result	Comparing the MIAA result, using images taken by each device, with the biopsy result	Study completion, on average 2 weeks
The AUROC of MIAA, using images of non-biopsied lesions from each of the image capture apparatus combinations, compared to clinical assessment	Comparing the MIAA result, using images taken by each device, with the clinician's diagnosis	Study completion, on average 2 weeks
The sensitivity of MIAA, using images of biopsied lesions from each of the image capture apparatus, compared to the biopsy result	%true positives, as determined by biopsy, identified by each imaging device	Study completion, on average 2 weeks
The sensitivity of MIAA, using images of non-biopsied lesions from each of the image capture apparatus, compared to clinical assessment	%true positives, as determined by clinician, identified by each imaging device	Study completion, on average 2 weeks
The specificity of MIAA, using images of biopsied lesions from each of the image capture apparatus, compared to the biopsy result	%true negatives, as determined by biopsy, identified by each imaging device	Study completion, on average 2 weeks
The specificity of MIAA, using images of non-biopsied lesions from each of the image capture apparatus, compared to clinical assessment	%true negatives, as determined by clinician, identified by each imaging device	Study completion, on average 2 weeks
The positive predictive value of MIAA, using images of biopsied lesions from each of the image capture apparatus, compared to the biopsy result	Probability that subjects with a positive MIAA result, using images taken by each device, have a positive biopsy result	Study completion, on average 2 weeks
The positive predictive value of MIAA, using images of non-biopsied lesions from each of the image capture apparatus, compared to clinical assessment	Probability that subjects with a positive MIAA result, using images taken by each device, are thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The negative predictive value of MIAA, using images of biopsied lesions from each of the image capture apparatus, compared to the biopsy result	Probability that subjects with a negative MIAA result, using images taken by each device, have a negative biopsy result	Study completion, on average 2 weeks
The negative predictive value of MIAA, using images of non-biopsied lesions from each of the image capture apparatus, compared to clinical assessment	Probability that subjects with a negative MIAA result, using images taken by each device, are not thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The false positive rate of MIAA, using images of biopsied lesions from each of the image capture apparatus, compared to the biopsy result	Probability that subject with positive MIAA result, using images taken by each device, has a negative biopsy result	Study completion, on average 2 weeks
The false positive rate of MIAA, using images of non-biopsied lesions from each of the image capture apparatus, compared to clinical assessment	Probability that subject with positive MIAA result, using images taken by each device, are not thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The false negative rate of MIAA, using images of biopsied lesions from each of the image capture apparatus, compared to the biopsy result	Probability that subject with negative MIAA result, using images taken by each device, has a positive biopsy result	Study completion, on average 2 weeks
The false negative rate of MIAA, using images of non-biopsied lesions from each of the image capture apparatus, compared to clinical assessment	Probability that subject with negative MIAA result, using images taken by each device, are thought likely to have melanoma by the clinician	Study completion, on average 2 weeks
The concordance of MIAA result between each of the image capture apparatus	The extent to which the image capture apparatus generate the same MIAA results as the other devices	Study completion, on average 2 weeks
The number of adverse events, including adverse device events and serious adverse events.	The number of adverse events, including adverse device events and serious adverse events.	Study completion, on average 2 weeks
The proportion of lesions with 4 images that can be analysed by MIAA	The proportion of lesions with 4 images that can be analysed by MIAA	Study completion, on average 2 weeks
The proportion of lesions with at least 1 readable images that can be analysed by MIAA,	The proportion of lesions with at least 1 readable images that can be analysed by MIAA,	Study completion, on average 2 weeks
The AUROC curve of the MIAA result, using MLE from all of the available images of non-biopsied PLs, compared to the clinical assessment	Compare the MIAA result with the clinicians assessment	Study completion, on average 2 weeks

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
Impact of patient characteristics on the AUROC assessment of MIAA	A statistical model will test the whether patient characteristics, such as age, gender, and skin type, affect the overall result and if so by how much	Study completion, on average 2 weeks
Impact of the image variables on the AUROC assessment of MIAA	A statistical model will test the whether image variables, such as manufacturer and lens type, affect the overall result and if so by how much	Study completion, on average 2 weeks
Impact of the assessing clinician's level of experience on the AUROC assessment of MIAA	A statistical model will test the whether the clinicians' level of experience affect the overall result and if so by how much	Study completion, on average 2 weeks
The concordance between the referring clinician's level of confidence for biopsy and the MIAA result	The extent to which the clinician's assessment of melanoma is the same as biopsy and MIAA results	Study completion, on average 2 weeks

Collaborators and Investigators

• Sponsor: Skin Analytics Limited
• Investigators: Principal Investigator: Ioulios Palamaras, MD PhD, Royal Free London NHS Foundation Trust

Study record dates

Study Major Dates

• Study Start (Actual): January 11, 2017
• Primary Completion (Actual): July 11, 2018
• Study Completion (Actual): July 11, 2018

Study Registration Dates

• First Submitted: November 18, 2015
• First Submitted That Met QC Criteria: November 19, 2015
• First Posted (Estimate): November 23, 2015

Study Record Updates

• Last Update Posted (Actual): January 25, 2019
• Last Update Submitted That Met QC Criteria: January 24, 2019
• Last Verified: January 1, 2019

More Information

Terms related to this study

• Keywords: Diagnosis; skin imaging
• Additional Relevant MeSH Terms: Neoplasms by Histologic Type; Neoplasms; Neuroectodermal Tumors; Neoplasms, Germ Cell and Embryonal; Neoplasms, Nerve Tissue; Neuroendocrine Tumors; Nevi and Melanomas; Melanoma
• Other Study ID Numbers: SA-001-4ev

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: UNDECIDED

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No