Scientists from Birmingham University Hospital in Britain have taught artificial intelligence to diagnose skin cancer with high accuracy. In this respect reports Medscape.
Skin cancer is the general name for malignant epitheliomas of the skin such as basal cell carcinoma, squamous cell carcinoma and melanoma. Timely treatment of these diseases significantly increases the chances of successful recovery.
Scientists sought to create a new early detection tool that would detect skin cancer with high accuracy. The new technology analyzes images of moles or blemishes to determine whether they are cancerous.
To test the accuracy of artificial intelligence algorithms, scientists conducted a study involving 6,900 people with suspected skin cancer. Scientists took photographs of the areas of the skin that needed to be examined. The technology analyzed the photo and made a preliminary diagnosis. AI scores were then shared with the subjects’ dermatologists.
The algorithms were able to detect melanoma with 100% accuracy and other types of skin cancer with 99.5% accuracy overall. In addition, the technology was able to identify 541 of 585 precancerous lesions. This is a precancerous condition of the skin, in which the development of a malignant neoplasm is possible.
Scientists hope that in the future, the use of a new artificial intelligence model as a preliminary diagnosis technology can reduce the burden on dermatology centers.
Previous scientists was created A new biological product for cancer immunotherapy.