examining mole on skin

First Look: Skin cancer smartphone apps

Been meaning to get a suspicious mole checked but haven’t got around to it? With a skin cancer app, you use your smartphone to take a photo of a mole and get an assessment of its risk. But are the apps accurate and easy to use? Five triallists got snapping.

×

Choose what’s right for you with confidence

Join today and get instant access to all test results and research.

Join Consumer Now

There are 2 main types of skin cancer app:

  • Algorithm apps use a mathematical algorithm to analyse the photo and provide a risk classification.
  • Teledermatology or doctor report apps give you a risk classification based on the photo you send to a skin specialist.

Some apps also require you to give a patient history, which includes your family history of skin cancer, the number of moles you have and whether your mole has changed appearance.

Our 5 triallists started by using the algorithm app SkinVision. However, only 3 were successful in uploading a photo with this app. These 3 went on to use the teledermatology app Firstcheck. For $30 you can buy a Firstcheck SkinScope – a lens attachment for your phone – which magnifies the image. This isn’t required for an assessment so we didn’t use the SkinScope. The same mole was used for each app.

The apps

Firstcheck

firstcheck.co.nz

Price: Free to download, $19.95 per consultation. Optional SkinScope $29.95.
How it works: Photo is sent to a skin specialist of your choice for a recommendation. Firstcheck providers can be dermatologists or plastic surgeons, but the majority are GPs with skin cancer qualifications. For better images, you can attach the SkinScope (smartphone lens attachment) to your phone. This gives a 20x magnified dermoscopic image. Images can be saved so you can monitor changes.
Response time: Within 72 hours.
Can check: Moles, spots, rashes, skin, hair or nail issues.

SkinVision

skinvision.com

Price: Free to download and for first photo assessment. An annual plan costs $33.99 and includes unlimited assessments and the option to have photos reviewed by a skin expert.
How it works: Uses fractal geometry, which looks for patterns in the outlines and dimensions of a mole that make it more likely to be malignant. It classifies a lesion as high-, medium- or low-risk. Images can be saved so you can monitor changes.
Response time: Immediate.
Can check: Moles and spots. However, can’t make recommendations on large lesions.

Trial results

SkinVision

Ease of use
All triallists initially found it difficult to take photos with this app – none got it first go. One said it was frustrating gauging the exact distance and angle needed to take the photo. Two had no luck: 1 took the photos OK but found the instructions confusing and lost patience; the other couldn’t get the app to take a photo.

Responses
The 3 stayers who had their photos accepted liked the immediate response. Two got a “medium risk” report and were advised to track their moles for changes. One had a “high risk” mole and was advised to see a specialist.

Firstcheck

Ease of use
Our 3 triallists found it simple to take a photo and send it to their chosen specialist (the app lists specialists in different regions).

Responses
Responses were received between 2 and 36 hours (less than the 72 hours Firstcheck promises). One triallist was told his mole was of “low suspicion”. He was advised to have another Firstcheck consultation in the next 6 months or see the specialist.

Our second triallist was told her mole was “medium risk” and she should make an appointment. After her appointment, she was advised to have the mole removed – the biopsy results confirmed it was not a concern.

Triallist 3 found her response contradictory. She was told it was likely to be a “benign naevus” but it was of “high suspicion” and she should make an appointment within 4 weeks. At the appointment, the specialist confirmed the mole was of no concern.

Firstcheck told us triallist 3 was sent a notification after her Firstcheck response advising her lesion was “low risk” and apologising for the confusion. However, she didn’t recall getting a notification alert and said she would not have made the appointment if she had.

Are they accurate?

This year the Australasian College of Dermatologists published a study on the efficacy of 3 algorithm apps. Fifty-seven lesions were assessed by the algorithm apps and compared with the clinical decisions of 2 dermatologists. According to study co-author Professor Monika Janda, none of the tested algorithm apps had high enough agreement with the dermatologists’ opinions. Of the 42 lesions classified as suspicious by the dermatologist, only 9 to 26 were classified as suspicious by the apps.

Dermatologist Dr Jennifer Pilgrim thinks the skin cancer app concept is commendable but the technology isn’t good enough yet.

“Overall there isn’t much evidence that the apps in our study offer much benefit over normal patient self-examination. In addition, a number of the apps could not analyse quite a few of the photos we sent,” Prof Janda says. “However, improvements in the algorithms may make them a useful tool in the future. App developers need to engage more with researchers and clinicians to test the performance of their product.”

Other studies have found problems with relying on algorithm apps. A 2013 study published in the Journal of the American Medical Association Dermatology analysed the performance of 3 algorithm and 1 teledermatology apps against histology results. All 3 algorithm apps incorrectly identified a large number of skin lesions – even the best algorithm app missed almost 30% of melanomas. The teledermatology app accurately identified 94% of lesions. A 2017 study, part-funded by SkinVision, found the algorithm app correctly identified 78% of pre-malignant conditions.

Dermatologist Dr Jennifer Pilgrim thinks the skin cancer app concept is commendable but the technology isn’t good enough yet.

“In my experience the apps are unreliable and either alarm the patient unnecessarily or incorrectly reassure them. I haven’t had anyone come and see me with a correct diagnosis from an app,” she says.

She is also concerned consumers only take a photo of the moles they can easily see. “About 5 to 10% of melanomas are detected during routine screening visits. These would be missed unless a person had a complete skin assessment.”

Professor Amanda Oakley believes the apps are good for raising awareness about sun protection and skin cancer, and are useful for archiving images. However, she points out images must be good quality and the danger is that small changes may not be noticed and an opportunity for early diagnosis could be missed.

“High-quality photography of skin lesions requires a dermoscopic attachment, skills and training. It’s a positive if an app doesn’t accept poor-quality photos.”

She expects the apps to get better as technology improves. Artificial intelligence offers a huge hope for the future, she says.

Skin cancer statistics

  • Skin cancer is the most common cancer affecting Kiwis – it accounts for about 80% of new cancers diagnosed each year.
  • New Zealand has one of the highest melanoma incidence rates in the world. More than 4000 people are diagnosed every year.


By Belinda Castles
Research and Testing Writer


16nov choosing wisely promo default

Choosing Wisely

Choosing Wisely is a campaign encouraging a change in thinking by health professionals and consumers to avoid unnecessary medical intervention.

Learn more

30aug promoimage default

First Looks

Cars, phones, infomercial products and anything else that takes our fancy — First Looks are trials of new or interesting products from the perspective of our product experts. Check out our latest.

Take a look