Melanoma detection: There’s an app for that (but best to skip it)
Some of us may not be able to live without our smart phones and their apps, but are we willing to bet our lives on them? A study this week likely answered that question when it evaluated several apps' ability to determine whether or not a mole is melanoma.
It found that apps are, at best, unreliable in differentiating cancerous lesions from benign ones.
The research – published in the Jan. 16 edition of JAMA Dermatology – assessed four apps’ ability to evaluate lesions using 188 images that had already been evaluated and diagnosed by a dermatopathologist.
The upshot: "Sensitivity of the 4 tested applications ranged from 6.8% to 98.1%; specificity, 30.4% to 93.7%; positive predictive value, 33.3% to 42.1%; and negative predictive value, 65.4% to 97.0%. "
Those are far from reliable results. In fact, the study found, 3 of the 4 apps incorrectly classified 30% or more of melanomas as unconcerning.
"The highest sensitivity for melanoma diagnosis was observed for an application that sends the image directly to a board-certified dermatologist for analysis; the lowest, for applications that use automated algorithms to analyze images," the researchers wrote.
City of Hope dermatologist Jae Jung, M.D., Ph.D., who was not involved with this study, said that the findings should raise red flags for anyone considering an app in lieu of a professional examination.
“Currently, apps don’t have a place in detecting melanoma, and it is very risky to rely on them to say whether or not a mole is cancerous,” Jung said.
The study’s authors agreed, noting that such apps are not subject to regulatory oversight and could delay people from seeking medical consultation. Such a delay could postpone the diagnosis and treatment of actual melanomas, the most deadly of skin cancers. According to the American Cancer Society, there will be more than 76,000 new cases – and 9,000 deaths – from melanoma this year.
In addition to missing melanomas, the apps also incorrectly identified benign lesions as cancerous, thus creating unnecessary alarm, Jung said.
Jung's advice: Skip the app. Instead, she advised, people worried about a lesion should follow the current guidelines for detecting melanoma – using self-examination, assessing a mole’s ABCDE (asymmetry, irregular borders or coloring, large diameters and evolution) profile, and asking a dermatologist to further analyze suspicious lesions.
Jung also stressed the need for context in screenings; people with risk factors such as a family history of melanomas or frequent, prolonged sun exposure should consider more frequent and thorough screenings.
Healthcare may be going increasingly digital, but for now, it's best to leave your mole evaluation to expert—and human—eyes.