New AI App Enhances Skin Cancer Diagnosis via Smartphones

A recently developed application utilizing artificial intelligence is transforming the diagnosis of skin cancer by allowing quick and precise examinations through smartphones.

This innovation is expected to substantially decrease waiting times within the UK's National Health Service (NHS). It follows the earlier use of a previous version of the technology, known as Derm AI, which aided in detecting approximately 20,000 cancer cases among over 230,000 patients in the British healthcare system. The earlier version required a special camera lens, whereas the latest iteration operates directly via the phone without additional equipment and has recently received the highest level of medical device certification in Europe.

The app, created by the British healthcare company Skin Analytics, analyzes images of moles and skin lesions using AI trained on thousands of images of confirmed medical cases. The system accurately identifies non-concerning cases while flagging suspicious ones for further medical review.

Accuracy and Early Detection Benefits

According to available data, the program achieved 99.8% accuracy in detecting melanoma, one of the deadliest and most dangerous types of skin cancer.

The primary warning sign of the disease is the appearance of a new mole or changes in the shape or size of an existing mole. These signs can manifest anywhere on the body but are more common in sun-exposed areas.

Clinical Impact and Accessibility

Dr. Alexandra Kemp, a dermatology consultant and clinical director for cancer treatment at Amersham Hospital, stated that integrating this technology into the diagnostic pathway has clearly improved clinical efficiency and the quality of patient care.

She emphasized that early detection of skin cancer increases the chances of successful treatment. Dr. Kemp also highlighted that making this technology accessible via smartphones, without the need for special equipment, broadens access to early screening and accelerates the diagnostic process.