Your smartphone has delivered all sorts of professionals into a state of semi-obsolescence: the mapmaker, the food critic, the weatherman. But how about the doctor?
An iPhone app called Skin Scan ($4.99 in the App Store) purports to analyze your moles for evidence of malignancy–and it recently nabbed €50,000 in seed funding from Seedmoney. Dermatologists are surely paying attention.
How does the app work? Snap a photo (at a distance of 10-15 cm, in good lighting) of a mole, upload it, and clearly label it–you’ll want a different file on each mole you’re going to analyze. You may want to snap several photos over time, since one predictor of whether a mole is really melanoma is the manner in which it grows. Add your age and gender, which are also relevant to your iDiagnosis, and then click analyze. The app crunches the images of your mole through its proprietary software, and decides whether it thinks your mole is growing normally or abnormally.
Do dermatologists need to worry about the end of their profession? Far from it. Two dermatologists are on the board of Skin Scan, first of all, so skin doctors at the very least have roles supervising the diagnostic robots of the future. But even that scenario is far-fetched. Skin Scan is the first to claim that its app isn’t a replacement for a doctor–it’s just meant to get more people who should see the doctor to actually go make the visit. A disclaimer on the app’s site says the app is for “informational purposes only,” and that it “is not intended to be a substitute for professional medical advice, diagnosis or treatment. Never disregard professional medical advice, or delay in seeking it, because of something you have read on Skin Scan.”
Could machine intelligence ever really replace the doctor’s diagnostic ability? Perhaps in some tasks (and I’m certainly not expert enough to judge if melanoma-detection is one of them), but in others, it’s not likely to happen anytime soon. I recently spoke with an AI-researcher-turned-entrepreneur who told me about his experience teaching machines to recognize cancerous biopsies. “In reality it was incredibly hard to get close to how well humans do even using the most sophisticated algorithms,” Ian Hogarth told me. “My overriding view now is that we’re a long, long way off from machine intelligence outstripping us.”
Still, that’s not to say an app like this isn’t useful. “Skin Scan has potential for good,” Yale dermatologist Irwin Braverman tells Technology Review, calling such an app “a boon to non-dermatologic physicians who can use it to screen out suspicious lesions during routine exams and refer the patient to a dermatologist and other appropriate specialists.” He hastens to add, though, that such an app could never replace an actual, breathing doctor: “There are some morphologic clues to melanomas that may not be quantifiable but are recognized by the human eye of well trained dermatologists.”
Skin Scan includes the option of uploading your location–partly so it can help you find actual, licensed dermatologists near you, but also so it can compile data of potential epidemiological value. The app’s homepage contains a map of scans, pegged to location and color-coded by the degree of cancer risk. The data probably has to do more with the incidence of iPhone ownership than anything else at this point, but there’s an intriguing possibility for a statistically useful dataset sometime in the future.
There has been much talk of telemedicine of late, but sometimes, the intersection of technology and medicine can happen in the home, before the doctor enters the picture. And even if Skin Scan doesn’t wind up being the first great dermatological app–reviews on the App Store are decidedly mixed–it seems to be doing fine as an awareness raising tool. Its blog, for instance, links to this affecting and effective video called “Dear 16-year-old Me,” full of testimonials from melanoma survivors.
Maybe save the five bucks on this app, then, and put it towards a new bottle of sunscreen instead.