WCD 2019 | Melanoma diagnosis: first data from artificial intelligence

Susana Puig, Professor of the Hospital Clínic di Barcelona, introduces an innovative study published on Lancet Oncology.

Artificial intelligence is becoming more effective when it comes to melanoma diagnosis, more effective than single experts. This is the results of a study published on Lancet Oncology and introduced at the 24th World Congress of Dermatology by Professor Susana Puig, from Hospital Clínic of Barcelona.

The research project, at which have participated expert dermatoscopists from all around the world, has compared the experts’ diagnosis with the ones done by the computers. They have used some of artificial intelligence algorithms: the top three algorithms together work better than the elaborated diagnosis done by the single experts. 

“This means that the algorithms of artificial intelligence have reached results as good as the ones by the experts – declares Prof. Puig – we have seen that when experts work together as a team, like the machines when they pull together algorithms, they reach even results”. 

“This represents an enormous step forward because until now artificial intelligence was working well with its algorithms, but only as a test in the learning phases and on lesions already selected and diagnosed. But they were not accurate enough for the real diagnosis and for the patients – explains Prof. Puig – now we have been able, through algorithms, to take such artificial intelligence to real diagnosis and proper patient, with a diagnostic accuracy for melanoma and other cutaneous lesions, just like an expert dermatologist”. 

The study published on Lancet Oncology is a great step forward for the diagnosis of cutaneous tumours and shows the importance of team work. Nonetheless, considering that the everyday clinical activity only concerns a doctor alone during the diagnosis, and a team diagnosis only in rare occasions, artificial intelligence seems to have an important role in future dermatology.

“The results of this study show the importance of human team work, but not always doable in everyday clinical practice”, concludes Susana Puig.