[2 min read] Accuracy of computer-aided melanoma diagnosis | Prof David Wilkinson

This month, I was fascinated to read a meta-analysis of the accuracy of computer-aided melanoma diagnosis, by Vincent Dick et al.

As we all know, computer-aided diagnosis is becoming more common across many aspects of medicine, including dermatology and skin cancer. We have heard about recent studies that use artificial intelligence and machine learning to achieve very high rates of diagnostic accuracy.

The current study is a useful summary, showing (as so many such meta-analyses do) that many studies done to date are of poor or limited quality, and that only a small number are of sufficient quality to be considered reliable.

This report shows sensitivity around 75% and specificity around 85% for melanoma diagnosis, and equivalent accuracy with dermatologists’ diagnosis.

As we move into a new era of computer-aided skin cancer diagnosis, researchers need to ensure that the studies they do reflect real-world practice.

Read the paper here.

Professor David Wilkinson


Read more from Professor David Wilkinson on recent research:


Learn more about skin cancer medicine in primary care at the next Skin Cancer Certificate Courses:

Skin Cancer Certificate Courses in Australia

One comment on “[2 min read] Accuracy of computer-aided melanoma diagnosis | Prof David Wilkinson

Leave a Reply

Your email address will not be published. Required fields are marked *