Detecting skin cancer: Ah, if there were only an app for that. Turns out, there is; in fact, the number of apps claiming to detect suspicious skin lesions is growing. It seems simple – just take a picture of your spot, and an app backed by an algorithm (or dermatologist) will scan it for signs of melanoma, like size, asymmetry or a strange color.
Problem is, can you trust an app over an in-person doctor visit with something as potentially life-threatening as cancer? Is it even responsible to take the chance? A new study analyzing cancer-identifying smartphone apps reports that only one in four of the apps claiming to evaluate a person’s risk for skin cancer was found to be statistically accurate. Doctors from the University of Pittsburgh School of Medicine conducted the research, which was published in JAMA Dermatology.
Of the four apps they tried, using photos of 188 moles (some confirmed with melanoma), three apps used algorithms and were free. The fourth cost $5 per photo and beams the photo directly to a dermatologist. According to the researchers, the dermatologist app, SkinVision, correctly identified suspicious moles 98 percent of the time; the others missed melanomas in nearly one-third of cases – bad news for a disease that is often fast moving and deadly.
The Centers for Disease Control and Prevention reports that skin cancer is the most common form of cancer in the U.S., with (highly curable) basal cell and squamous cell carcinomas most common. Melanoma, on the other hand, can be far more dangerous. Of the 61,646 people in the US diagnosed with melanoma in 2009, 9,199 people died from it.
The new smartphone apps certainly seem appealing – particularly as resources for the un- or underinsured. But lead researcher Dr. Laura Ferris, assistant professor at the Department of Dermatology, University of Pittsburgh School of Medicine, told NPR that the technology is "not good enough to trust your life with.”
Would you use a smartphone app to help you decide whether to be concerned about a skin lesion, or leave that diagnosis to a (human) pro?