Consumer smartphone apps not accurate for melanoma detection

Several popular smartphone apps designed to evaluate photographs of skin lesions and provide feedback to non-clinician users about the likelihood of malignancy are not accurate, according to a Jan. 16 article in JAMA Dermatology.

A study found the performance of these apps in assessing melanoma risk to be "highly variable," with three of four apps in the study incorrectly classifying 30 percent or more of melanomas as "unconcerning," the article reports.

"Reliance on these applications which are not subject to regulatory oversight, in lieu of medical consultation, can delay the diagnosis of melanoma and harm users," states the article. "Despite disclaimers that these applications are intended for educational purposes, they have the potential to harm users who may believe mistakenly that the evaluation given by such an application is a substitute for medical advice."

The study examined sensitivity, specificity, and positive and negative predictive values of four smartphone apps used to determine whether a skin lesion is benign or malignant. The sensitivity apps ranged from 6.8 percent to 98.1 percent; specificity, 30.4 percent to 93.7 percent; positive predictive value, 33.3 percent to 42.1 percent; and negative predictive value, 65.4 percent to 97 percent. 

The highest sensitivity for melanoma diagnosis was observed for an app that sends the image directly to a board-certified dermatologist for analysis, while the lowest sensitivity was for apps that use automated algorithms to analyze images. 

A total of 188 lesions were evaluated using the four smartphone apps. Of these lesions, 60 were melanoma, and the remaining 128 lesions were benign. Each app was presented with each of the 188 lesions in the study, and the test result was recorded as positive, negative, or unevaluable.

"Three of the four applications we evaluated do not involve a physician at any point in the evaluation," the article states. "Even the best-performing among these three applications classified 18 of 60 melanomas (30 percent) in our study as benign."

In related research, interactive electronic tools were found to improve patient performance of skin self-examinations. A computer-assisted learning tutorial and frequent telecommunications reminders to perform self-exams helped to boost patient confidence in the ability to identify melanoma, according to researchers.

To learn more:
- read the journal article