Skin Lesion Detection in a Clinical Environment

  • Research type

    Research Study

  • Full title

    Clinical Decision Support System for Skin Monitoring and Skin lesion Classification using Machine Learning

  • IRAS ID

    284034

  • Contact name

    Janusz Kulon

  • Contact email

    j.kulon@southwales.ac.uk

  • Sponsor organisation

    University of South Wales

  • Clinicaltrials.gov Identifier

    21430, KESS

  • Duration of Study in the UK

    1 years, 9 months, 31 days

  • Research summary

    There are often limited number of dermatologists available, which means General Practitioners (GPs) are often the first to diagnose a patient’s skin condition. Both the written observations and an image of the skin lesion is submitted to a dermatologist to confirm the diagnosis. The patient would then be referred to a dermatologist who will confirm the diagnosis in person and recommend the necessary course of treatment to prevent it from progressing.

    Diagnostic procedures are used in clinical environments to increase the accuracy and reliability of diagnoses. The method used by GPs is called the ABCD rules, which includes measuring the Asymmetry, Border, Colour and Diameter. These rules however, are ambiguous and unreliable when diagnosing rarer types of melanoma or distinguishing some types of seborrheic keratosis.

    The objective of this study is to increase the reliability of GP observations by developing a Computer-Aided Diagnosis system (CAD) that can support the diagnostic procedure by standardising referrals to dermatologists with the assistance of machine learning algorithms. This would require a large amount of image data accompanied by clinical observations to train and test the effectiveness of the algorithms within a clinical environment.

  • REC name

    N/A

  • REC reference

    N/A