Development of an Aid to Melanoma Detection Using Artificial Intelligence Algorithms Based on Images From the VECTRA 3D System.

Not Applicable

Not yet recruiting

• Conditions: Naevi

• Registration Number: NCT06999499
• Lead Sponsor: Assistance Publique Hopitaux De Marseille

Brief Summary

The background to this research is that frequent medical screening of the general population for melanoma is not feasible. The real challenge of this project is to develop an automatic process for detecting any potential melanoma. To this end, the project aims to design an algorithm to build a novel diagnostic aid that makes use of the similarity and disparity of pigmented lesions in the same patient. To achieve this, we need to obtain and structure a large database of images grouping all pigmented lesions per patient according to their similarities as perceived by dermatologists.

Detailed Description

Not available

Study Timeline

• Status Verified: 2025-05
• Study First Submitted: 2025-05-20
• Study First Submitted QC: 2025-05-30
• Last Update Submitted: 2025-05-30
• Study First Posted: 2025-05-31
• Last Update Posted: 2025-05-31
• Study Start: 2025-09-01
• Primary Completion: 2027-09-01
• Study Completion: 2027-12-01

Recruitment & Eligibility

• Status: NOT_YET_RECRUITING
• Sex: All
• Target Recruitment: 1000

Inclusion Criteria

• Male or female aged 18 and over
• Patient with more than 15 nevi (moles) of various phototypes (I to III)
• Patient who has received information about the study and has not expressed any opposition
• Patient who is a beneficiary or entitled person under a social security scheme

Exclusion Criteria

• Patients with phototype V
• Patients with chronic inflammatory skin diseases
• Claustrophobic patients
• Patients who are bedridden or handicapped
• Patients who are excluded from another research protocol at the time of collection of the non-objection.
• Patients covered by articles L1121-5 to 1121-8 of the French Public Health Code (minors, adults under guardianship or trusteeship, patients deprived of their liberty, pregnant or breast-feeding women),
• Any other reason which, in the investigator's opinion, could interfere with the evaluation of the research objectives.

Study & Design

• Study Type: INTERVENTIONAL
• Study Design: SINGLE_GROUP

Primary Outcome Measures

Name	Time	Method
development and validation of algorithms to identify lesions clinically suspected of being melanoma by a dermatologist (potentially malignant and/or ugly duckling).	from enrollement to until 6 month	Comparison of the results given by the analysis of the images by 3 dermatologists or by the software. Estimation of sensitivity and specificity thresholds of at least 93% (accuracy level 5%).

Secondary Outcome Measures

Name	Time	Method
concordance rate for malignant annotations	From enrollement to 6 month after	concordance between the malignant yes-no annotations of each of the three dermatologists (2 to 2) will be tested.
concordance rate for ugly duckling	From enrollement to 6 month after	concordance between the ugly duckling yes-no annotations of each of the three dermatologists (2 to 2) will be tested.
calculation of the proportion of melanomas confirmed by anatomopathology	From enrollement to 6 month after	For each lesion identified as malignant or ugly duckling by the gold standard and removed for histological analysis: calculation of the proportion of melanomas confirmed by pathology.

Trial Locations

Locations (1)

Hôpital de la Conception; 🇫🇷 Marseille, France