Skin cancer is an example of how AI can revolutionise cancer diagnostics.
Skin cancer is the most common type of cancer in the United States, with 9,500 people receiving a diagnosis every day. Standard diagnosis involves visual inspection and histological analysis, which can often take several months. Misdiagnosis rate is around 20%, and around 28% of misdiagnosed cases are life-threatening.
Innovations in medicine often come from unexpected places. Smartphones are promising in skin cancer detection as they have high-performance digital cameras which can be fitted with dermoscope adapters (see figure). ML approaches to skin cancer classification have already been implemented in smartphone apps. These developments raise the possibility of accurate and rapid home tools for skin cancer screening.
At CoolClinical we are exploring the application of ML algorithms to cancer detection. For the technical papers, click below...