Well maybe not hats off, we are a company that encourages people to be sun smart after all! So we shall leave our hats on and offer a round of applause instead.
The University of Heidelberg has published a paper in the Annals of Oncology that demonstrates the important role that artificial intelligence (AI) can play in melanoma detection. At Skin Analytics this is something that we have believed for a long time. This paper is another example of high quality research that backs us up.
In simple terms, The University of Heidelberg customised the Google Inception Network and trained it to recognise melanoma. They then prepared 100 images of skin lesions, some of which were melanoma, so that that could compare the diagnostic ability of AI and dermatologists.
On one side there was the specifically trained AI network and on the other side 58 qualified dermatologists. Each side reviewed the images, provided diagnoses and the results were then compared. The outcome showed that AI was better at identifying melanoma than the dermatologists.
It is probably important to say that in our opinion the test that was run by the University of Heidelberg was probably a little biased towards AI. Clinicians don’t only view an image to make their decision, they use many other factors too and removing these factors is a little like asking you to do up a button without using your thumbs.
By any measure, this study is rather exciting because it shows that AI has a huge role to play in finding melanoma earlier to help improve survival rates. (If want you can read the paper here).
When it is all said and done we believe that dermatologists and AI can work in tandem to help improve the health outcomes for skin cancer patients. That is why at Skin Analytics we are focused on delivering a solution that works with clinicians, not in place of them. A small but important difference.