Use of Artificial Intelligence in Histological Detection of Melanocytic Lesions in Sun-damaged Skin

This study is an attempt at evaluating the use of artificial intelligence in detecting melanocytic lesions in skin tissue with prominent sun damage.

Places available:

  • One Master of Biomedical Science
  • One Honours

Pigmented lesions of the skin are some of the most commonly biopsied skin conditions, which require histological assessment by a pathologist for accurate diagnosis.  Some of these lesions are melanocytic, whilst a large proportion of them are of non-melanocytic origin. Pigmented lesions are often biopsied because of their suspicious morphology and melanomas are on the top of the list of conditions to be excluded or confirmed. Therefore, assessing skin biopsy material for a potential melanoma is one of the most common requests pathologists receive on a daily basis. One of the difficulties in assessing these lesions microscopically arises from the fact that a subtle melanocytic proliferation can be difficult to interpret in a severely sun-damaged skin, which melanomas commonly arise from.

This project is an attempt at evaluating the use of artificial intelligence in detecting melanocytic lesions in skin tissue with prominent sun damage. This will be investigated both on routine stains and special stains for melanocytic markers and Deep Learning techniques will be used for assessment of the scanned histological images. This area is novel, topical and exciting, with only limited research done to date in this field.

The project would suit someone with strong interest in both pathology as a science and artificial intelligence as a tool and its promising potential applications. A good grasp of statistical analysis and previous research experience is highly desirable.

Contact and more information

Associate Professor Kais Kasem 
kais.kasem@unimelb.edu.au