Artificial Intelligence in Skin Cancer Diagnosis: The pros and Cons

Artificial intelligence (AI) is integrating into all facets of life — from education and art to business and healthcare. In the last decade there has been a rise in the number of AI-powered skin check apps which show a high level of accuracy in classifying skin lesions. Yet, while the use of AI for something as serious as skin cancer diagnosis is plausible, can it really outperform a professional skin check with dermatologist diagnosis?

The increasing risk of skin cancer

Melanoma is the most serious type of skin cancer. According to the Melanoma Research Alliance, the rates of melanoma (particularly in younger people) are quickly rising. Between 2013 and 2023 the number of melanoma cases diagnosed annually grew by 27%. This is even though the rates for other types of cancers have declined.

The good news is that melanoma treatment has come a long way. If detected early, before the skin cancer progresses past stage two, 90% of melanomas can be successfully treated. 

How do AI-powered skin check apps work?

The truth about skin check apps is that they offer a simple and quick way to check and track moles. Typically, they use a built-in camera or require a clip-on lens to photograph the skin. Using an AI-based algorithm, they analyse the mole image to identify abnormalities in the cells. The mole may be classified as high or low risk for skin cancer, offering insight into the need for clinical diagnosis. 

Skin check apps may also offer the following features:

  • Ability to take full body images
  • Mole mapping
  • Image library to assist with tracking and comparing changes in the skin
  • Measurement tools
  • Information on what to look for during a skin check
  • Reminder alerts to prompt a professional skin cancer check

Currently, no skin check apps provide a skin cancer diagnosis

What skin check apps are available?

From start-up to more advanced, there are a range of skin check apps available to download, including:

How effective are skin cancer apps for early detection?

The American Academy of Dermatology reports that when put to the test, 41% of melanomas were missed by the skin check apps used. This has led dermatologists to raise concerns that apps could cause an increase in misdiagnoses and delay in skin cancer diagnoses.

According to the authors of a 2020 study exploring the accuracy of skin check apps, ‘apps cannot be relied on to detect all cases of melanoma or other skin cancers,’ due to ‘variable and unreliable’ results.

Advantages and disadvantages of skin check apps

AI has shown promising results in assisting clinicians with early skin cancer detection and diagnosis. Yet, like all technology, it has its pros and cons. 

PROS CONS
Acts as a tool in the clinician’s repertoire for earlier diagnosis May give way to false reassurance and delay diagnosis
Assists dermatologist to make faster, informed decisions Accuracy is questionable in a clinical setting
Can reduce unnecessary referrals for benign lesions Data-collecting limitations (i.e. obtaining images or rare skin cancers and benign disorders) impact AI training
Offers a more standardised approach to skin cancer detection Most rely on the built-in camera lens, which is less accurate than a professional’s trained eye and imaging tools
Over-reliance on AI, reducing clinician’s critical judgement 

 

The downside of skin check apps is mostly linked to insufficient testing, lack of expert input during the development process, and problems with the technology itself. Of course, over time, these are issues that can be fixed. 

Conclusion

From a patient/consumer viewpoint, skin check apps should not replace a professional  skin check. The ease and accessibility of AI-powered skin check apps may encourage more people to perform a self-skin-check. They offer valuable assistance in the early detection of a suspicious mole or lesion, but they shouldn’t be relied upon for absolute certainty. Even apps claiming high percentage accuracy are not a substitute for a professional skin cancer check.

In dermatology, the use of AI can assist with diagnosis, potentially improving the speed and accuracy of diagnosis. However, it’s important to acknowledge that the skills and expertise of the clinician play a prominent role in getting the most from AI. So, while AI assistance is a useful tool for dermatologists, for the time being it does not replace the clinical diagnosis of melanomas.