As mobile devices become increasingly powerful, their applications in healthcare can be expected to grow. Recently, an app called MelApp was launched by the Health Discovery Corportation, a “molecular diagnostics company that uses patent protected advanced mathematical techniques for personalized medicine.” The app claims to apply these techniques to skin lesions by analyzing photographs, taken by the iPhone, to assess risk for melanoma.
On the surface, the functionality, while advanced, sounds reasonable. The classic teaching for assessing malignancy risk for a lesion that looks like melanoma is ABCD – asymmetry, borders, color, diameter. Those are all features that are analyzed by looking at the lesion and all seem like parameters that could be assessed by a computer using a high-resolution image with a large reference database.
Whether the technology is feasible is something that I’m going to leave for another day. For now, lets assume it is. Before launching this kind of technology, the responsible thing to do is to validate it in an open method – unfortunately, only vague reference is made to the fact that the app was validated against a Johns Hopkins database, without any publications or data to support that claim. In fact, the press release indicates the goal of the app is to “increase awareness” of melanoma, a far cry from the diagnostic capability the app seems to imply.
In addition, this app, at best, would be comparable to only visual inspection of a lesion. As we’ve discussed previously, even when done by a physician, visual inspection alone of a single suspect lesion could lead to missing other dangerous lesions that would have been caught had the patient visited a dermatologist. Furthermore, assessing risk doesn’t stop at visual inspection – a patient’s personal history, family history, attitudes towards their own health, and medications are all critical. A patient with a history of heavy sun exposure with multiple skin cancers in the family would be high risk, almost regardless of what the lesion looks like.
Finally, its worth noting that the revenue model for the app appears to include a fee for inclusion in the “physician referral base.” Perhaps an unintentional incentive then becomes to structure the app in such a way that it increases referrals, thereby increasing the lure for physicians to sign up.
This kind of technology could eventually be an important tool when it comes to evaluation for skin cancer in particular. For MelApp, the data seems, however, to be sorely lacking at this time, despite the accomplished clinicians behind it. That being said, I suspect that MelApp is representative of what we can expect in coming months and years in mHealth technology, greatly expanding the toolkit of healthcare providers.
