I was recently dragged kicking and screaming back into the age of technology. After moving, I had given up the idea of a smartphone until I realized I can’t live without a device that receives wireless internet wherever I go, tells me the weather when I wake up, finds the nearest Zagat-rated restaurant, acts as an alarm clock for those pesky morning classes, helps me when I get lost in Manhattan’s “off-grid” section, and detects if I have skin cancer. In the last month or so, I have been assaulted by a barrage of news segments and articles regarding the accuracy of skin cancer apps available for smartphones. However, being smartphone-less meant I never had the joy of testing these apps myself before hearing/reading about their faults. So what was the first thing I did when I was rudely hauled back into the world of perpetual Facebook-checking and emailing? I got myself some mole-detecting apps!
The ruckus created by the media centers around a study originating from the University of Pittsburgh Medical Center in which four unnamed skin-cancer/mole detecting apps were tested for the accuracy (sensitivity) of melanoma detection. The study found that three of four apps incorrectly classified many moles as not being melanomas; in fact, more than 60% of melanomas went undetected by these apps. This is not a very satisfying or reassuring statistic by any means. The fourth app, with a higher sensitivity (it detected a glorious 98% of melanomas) cost $5 to send an image into the app, which was then analyzed by board-certified dermatologists, and the diagnosis was sent back to the app user within 24 hours.
Typical smartphone apps use a photograph of a skin lesion the app user wants analyzed, sends the picture through an automated computer algorithm (that accounts for asymmetry, irregular border, variation in color, size, self-noted evolution) and “analyzes” the image. A computer algorithm is no substitute for the human eye or sensibility; many lesions that are subject to cancer progression do not fit nicely into a computer algorithm and are often “missed”, giving the user false hope that a lesion is non-cancerous.
Unfortunately, as with many items in our modern world, health apps are poorly regulated by government bodies. In 2012, the FDA Safety and Innovation Act allowed the FDA some regulatory control over medically-related smartphone apps. However, at these early stages in regulation it is not clear which apps fall under the FDA’s regulatory domain. All the apps I tried had disclaimers stating the app is for education or self-tracking of lesions, and should not be used in place of a dermatologist. In fact, many of the apps had lists of dermatologists and their office locations, which could be accessed based on the user’s current location (super helpful).
Some personal problems I had using the mole-detecting apps were the cumbersome nature of trying to take pictures of the top of your head, sides of the torso, and the entire “backside” of the body, as well as consistency both within apps and between apps. I’m sure part of my “cumbersome problem” would have been mitigated with the use of a compact iPhone instead of an iPad. Also, I found that the apps would suffer in consistency during, and between different photo-taking sessions themselves. If I changed the angle, light or distance from a mole, and took pictures of the same spot using the same app, I would get slightly different results each time. I think it would be hard to track a mole in a significantly meaningful way if the pictures you have documenting the progression of said mole are taken in different conditions from the last picture. In addition, I found inconsistencies between the apps themselves. I chose a test spot on my knee that was easy to reach and I was able to develop a pretty good way of taking a consistent picture. Some apps felt my mole was of little worry while others placed it in a “medium-risk” category (for a reference, I have had a body check and the mole I was using is a pretty normal mole that’s not going anywhere by my dermatologist’s estimation).
Furthermore, I felt the apps greatly understated the importance of visiting a dermatologist if something ”suspicious” or “high-risk” is detected, as well as the significance of early skin cancer detection. As with any cancer, if skin cancer is not caught early it becomes harder to treat and prognosis is poorer as the cancer entrenches into a location or metastasizes to distant locations in the body. Many dermatologists and health professionals in other fields of study worry that with the advent of easily accessible smartphone apps to detect medical problems (I even found apps claiming to be able to detect heart arrhythmias), populations of America in a lower-income bracket, and those uninsured or undocumented will substitute a smartphone app for an actual doctor. No matter how good an app is at its “job”, an app is not a suitable alternative for a real doctor’s visit; a smartphone app cannot take into account your personal medical history, your family’s medical background, your allergies, all your symptoms, and your daily lifestyle to make a diagnosis, unlike a doctor.
Though these mole-detecting/checking apps have their shortcomings, they are very good at (what I hope were the makers’ real intentions) raising awareness. These apps make the user aware of checking their moles (the apps send self-examination reminders), monitoring changes (self-examinations require the updating of any changes and photographic evidence), and are just generally awesome when it comes to promoting self-awareness for the early-detection of precancerous or even early stage cancerous skin lesions. Some of the apps are even built in a way so as to provide a handy skin self-check document!
In addition to being “mole-checkers”, I found many of the apps were doing double duty as a dermatologist finder, a UV index for various locations in the country on various days (with included sunscreen instructions for SPF level and reapplication times), giving cosmetic information, or containing guides for visually self-checking moles without a computer algorithm. So I say, why not download an app or two if not to just keep a tab on “skin things” and remind yourself to check moles on a regular basis. As long as it is remembered that everything the app generates regarding a diagnosis should be taken with a grain of salt, and an app is by no means a suitable substitute for true visual diagnosis and biopsy from a professional, there is no harm in snapping a photograph series of your lumps, bumps and colored spots for posterity’s sake.
Margit Lai Wun Juhasz
Mount Sinai Medical Student