Editor: Bolysbek Dana
Author: Tulkibaeva Nursulu
Translator: Tursunova Balkadisha
Melanoma is a particularly aggressive type of cancer that develops from melanocytes-cells containing the dark pigment melanin. One of the main signs of this tumor is the "ugly duckling" symptom. The appearance of a formation on the skin that is not similar to other moles characterizes the concept of this symptom.
The system, created by scientists from the Massachusetts Institute of Technology and the Wyss Institute of Biological Engineering, is able to recognize suspicious changes in the skin of patients from photos, with an accuracy of up to 90%. Also, this automated diagnostic system, by analyzing images, notices the "ugly duckling" symptom, and classifies it as cancerous.
To test the system, three dermatologists examined 135 images from 68 patients, identifying "suspicious" changes from the images. The same images were analyzed and evaluated by the algorithm. When the results were compared, the researchers found that the algorithm was consistent with the dermatologists ' choice in 88% of cases.
"We were able to achieve dermatologist - level accuracy in the diagnosis of melanoma from images, images can be taken by anyone with a smartphone, this opens up great opportunities for the early detection and treatment of this form of cancer," said study co-author Jim Collins.
Source: https://www.sciencedaily.com/releases/2021/02/210217151108.htm
