If you would like to submit a blog post for consideration, please email firstname.lastname@example.org
[3 min read] Can smartphone apps really detect skin cancer?
Do smartphone apps that assess moles for skin cancer actually work? Some new research suggests these popular apps might not be a reliable method of detecting all forms of skin cancer. Read on for an additional expert comment from Professor David Wilkinson!
A recent review of nine different studies found that some skin cancer detection apps were missing melanomas and incorrectly flagging harmless moles as suspicious.
The assessment of SkinVision and SkinScan revealed that the European apps performed poorly and could not safely be recommended for use.
The apps were found to be limited in their ability to reliably identify melanomas, which need to be detected early for the best chances of successful treatment and cure.
The authors of the review said that prompt and accurate detection of skin cancer is crucial, which is why unreliable methods of diagnosis such as smartphone applications need to be used with caution. However, the apps do have merit in that they encourage people to proactively take notice of their skin.
Around 800,000 skin cancers are diagnosed every year in Australia, with melanoma being the most common cancer affecting young Australians aged 15-39.
Living in the country with the highest skin cancer rate in the world, it’s vital for all Australians to have easy access to a reliable and qualified doctor who is trained in skin cancer detection.
Comment from Professor David Wilkinson:
If you have a smartphone – who doesn’t? – and if you are interested in skin cancer, you have probably downloaded one of the many apps available that claim to help detect melanoma and other skin cancers. How good are these apps? Are they useful, are they safe? This important paper, published in the highly reputable British Medical Journal provides an important insight. Basically the paper shows that the accuracy of the algorithms in the tested apps is low. The authors also point out that performance of these apps will surely be worse when used widely by the general public, and hence there is real risk of missed and incorrect diagnoses. So, time to delete the app from that phone.
It is important to note that this study is NOT looking at the role of image recognition, artificial intelligence and machine learning algorithms that are being used in formal research settings and evaluated in clinical studies. These sophisticated algorithms do work and are impressive. Maybe they will move into our smartphones over time. They are not there yet.
Professor David Wilkinson
Learn more with HealthCert Certificate Courses in Skin Cancer:
Source: CNN Health