As well as a 100% sensitivity for detecting melanoma, the new software correctly detected 99.5% of all skin cancers and 92.5% of pre-cancerous lesions.

The latest artificial intelligence (AI) software reached a 100% detection rate for melanoma, according to research presented at the European Academy of Dermatology and Venereology (EADV) Congress in Berlin.

The study assessed 22,356 patients with suspected skin cancers over a 2.5-year period.

As well as a 100% (59/59 cases identified) sensitivity for detecting melanoma, the new software correctly detected 99.5% (189/190) of all skin cancers and 92.5% (541/585) of pre-cancerous lesions.

The third version of the AI software marks a significant improvement from the first model, tested in 2021, which detected 85.9% (195/227) of melanoma, 83.8% (903/1078) of all skin cancers and 54.1% (496/917) of pre-cancerous lesions.

"This study has demonstrated how AI is rapidly improving and learning, with the high accuracy directly attributable to improvements in AI training techniques and the quality of data used to train the AI,” says study author Dr Kashini Andrew, specialist registrar at University Hospitals Birmingham NHS Foundation Trust. “The latest version of the software has saved over 1,000 face-to-face consultations in the secondary care setting between April 2022 and January 2023,3 freeing up more time for patients that need urgent attention."

The data is incredibly encouraging, but AI should not be used as a standalone detection tool without the support of a dermatologist, the researchers stress. Of the basal cell carcinoma cases, a single case was missed out of 190, which was later identified at a second read by a dermatologist 'safety net'. 

This further demonstrates the need to have appropriate clinical oversight of the AI.  “Further research with appropriate clinical oversight may allow the deployment of AI as a triage tool,” says study  co-author, Dr Irshad Zaki, a consultant dermatologist at University Hospitals Birmingham NHS Foundation Trust. “ However, any pathway must demonstrate cost- effectiveness, and AI is currently not a stand-alone tool in dermatology.”