Study Title:

Effectiveness of an image analysing algorithm (MIAA) to diagnose melanoma compared to gold standard histological determination.

Version & Date:

Version 1.3; 27 April 2018

Type of study:

Efficacy study

Trial Design:

Prospective, multi-centre, single-arm, cross-sectional, blinded study. Selected pigmented lesions (PL) will be photographed three times using different combinations of camera and dermoscopic lenses in a single visit. Images will be analysed by MIAA and the results compared to the biopsy result. A feasibility pilot study is planned.

Trial Participants:

Patients attending a dermatology clinic, with at least 1 PL that is scheduled for biopsy.

Planned Sample Size:

At least 100 PLs positive for melanoma, as determined by biopsy. Additionally, 1000 non-biopsied lesions are planned, which is predicted to require at least 500 patients.

The feasibility pilot study will recruit all suitable patients that attend the clinic on the designated days. These patients will form part of the study population.

Follow-up duration:

None.

Planned Trial Period:

Patients will complete all study related activities in one visit. Results of biopsies, where ordered, will be collected.

Primary Objective:

To demonstrate effectiveness of MIAA to diagnose melanoma in patients presenting with PLs scheduled for biopsy.

Secondary Objectives:

Improve the predicted sensitivity and specificity of MIAA by training the algorithm with real world data

To demonstrate the effectiveness of MIAA to assess PLs as compared to the clinical assessment;

To demonstrate safety of MIAA;

To demonstrate concordance between images taken using different camera types, manufacturers and lenses;

To demonstrate feasibility of image collection by patients.

Primary Endpoint:

Area Under the Curve of a Receiver Operating Characteristic (AUROC) curve of the MIAA result, using a maximum likelihood estimation (MLE) from all of the available images, compared to the biopsy result.

Secondary Endpoints:

The AUROC curve of the MIAA result, using MLE from all of the available images of non-biopsied PLs, compared to the clinician’s assessment of melanoma likelihood is a key, powered secondary endpoint.

The sensitivity, specificity, false positive and false negative rate, positive predictive value and negative predictive value of MIAA, using a MLE from all of the available images of biopsied PLs, compared to the biopsy result;

The sensitivity, specificity, false positive and false negative rate, positive predictive value and negative predictive value of MIAA, using a MLE from all the available images of non-biopsied PLs, compared to the clinician’s assessment of melanoma likelihood;

The AUROC, sensitivity, specificity, false positive and false negative rate, positive predictive value and negative predictive value of MIAA of each camera / lens combination, compared to the biopsy result and clinician’s assessment of melanoma likelihood;

The concordance of results from each of the camera / lens combinations;

The proportion of PL with 3 images that can be analysed by MIAA;

The proportion of PL with at least 1 readable images that can be analysed by MIAA;

The number of adverse events, including adverse device events and serious adverse events.

Device Name:

Melanoma Image Analysis Algorithm (MIAA)

Manufacturer Name:

Skin Analytics Ltd

Principle intended use:

To aid detection of melanoma