HealthCert Blog

Will artificial intelligence reduce the need for skilled dermatologists in the future? In a commentary recently published in the Conference Review of the 76th American Academy of Dermatology Annual Meeting 2018, Dr S. Swetter discussed whether artificial intelligence will “do us out of a job”.

One of the great skills of dermatologists is differentiating benign from malignant melanocytic and non-melanocytic skin lesions. Dr Swetter highlighted that computer programs have now been developed to do this task. A group of Stanford University researchers have created an artificial intelligence program based on a patterned recognition algorithm (a deep “convolutional neural network” or CNN) to differentiate cutaneous malignancies.

The two main questions are (1) Is the lesion benign or malignant? (2) Should you biopsy, treat or reassure? The algorithm was validated against biopsy-labelled images and a panel of 21 board-certified
dermatologists. The CNN performed as well as the dermatologists.

While this study has limitations, it represents the dawning of a new dermatology diagnostic tool. An “app” has already been created to study this algorithm prospectively in real time validation in the clinical setting. However, it is comforting that artificial intelligence will aid, but not replace, the dermatologist.

Read more recent research in dermatology.

Source:
Swetter, S. Will artificial intelligence replace dermatologists? 76th American Academy of Dermatology (AAD) Annual Meeting 2018 Conference Review. February 2018.

Interested in General Dermatology?

The HealthCert Professional Diploma of General Dermatology will teach you how to manage all major dermatological conditions in primary care. The courses are university quality-assured, CPD-accredited and count towards multiple Master degree pathways and clinical attachment programs in Australia and overseas. The program is delivered online.