Using a specially developed application, a regular smartphone will detect dangerous growths on the skin.
Some skin cancers, such as melanoma, are very aggressive. But distinguishing a malignant form from a mole is not so easy: you need to regularly examine your body, monitor the appearance of moles, remember their location, shape and colour in order to be able to see the changes that occur. Oncologists recommend doing these “self-examinations” regularly. And it is ideal to periodically photograph your moles to make it easier to see the changes.
If the growths behave in an uneasy manner (the size, shape and colour of the moles change), they should be examined by a dermatologist, and if there is a threat of cancer detection, by an oncologist. Cancer surveillance is important, even if in most cases the examination shows that a mole or other skin formation is harmless. These examinations, if carried out regularly, are the only way to “detect” cancer at an early stage.
The About Moles app, developed within the framework of the Melanoma-Free Region project, will help the general practitioner or therapist to carry out a visual analysis of the patient’s skin condition. This medical decision support service can be used by primary care physicians. The patient himself can use the app if he notices suspicious growths on his skin.
“The About Moles app works on the basis of two neural networks. Using artificial intelligence, the image of a mole is analyzed, i.e. the neural network detects a mole in the photo, the second one classifies it into six types of neoplasms,” one of the app developers, AIMED development director Anton Melnikov, explained to RG.
To date, the neural network has been trained to recognize melanoma, squamous cell skin cancer, basal cell skin cancer, seborrheic keratoma, nevus and hemangioma. In addition, the accuracy of the determination is high.
“We conducted clinical studies at the Blokhin National Medical Research Center in July 2023 and they confirmed the high accuracy of our complex. The results of clinical studies showed a sensitivity of 93.5% and a specificity of 84.3%. This figure exceeds the quality of work of even the most prominent doctors – oncologists, who account for only 92%, not to mention dermatologists – 83%, and general practitioners – 65%,” Melnikov says.
Of course, we are talking only about a preliminary, initial diagnosis. The request points out a possible problem and the person understands that an urgent examination is necessary.
“Anyone can use the app. You need to download the app, take a photo of the moles and get the result of the diagnosis. In addition, we now have a separate app called “About Moles Honey”, created specifically for doctors. A doctor, using a mobile phone, can also take a photo of a patient’s mole and receive the result of the examination,” concluded Anton Melnikov.
