Can your smartphone diagnose skin cancer?
Not long ago, a cellphone was merely a portable device that could transmit and receive phone calls. It has evolved into smartphones, which can be enhanced with a myriad of apps ranging from game playing to serious functions. A number of apps are currently available for medical use, such as monitoring diabetes, displaying X-ray images, and measuring heart rates. The ability of a smartphone to take photographs has been enhanced to develop apps that can determine whether a skin lesion is a benign mole or a melanoma, which is an extremely dangerous form of skin cancer. The question becomes: how good are these apps? Researchers affiliated with the University of Pittsburgh Medical Center addressed that question. They published their findings online on January 16 in the journal JAMA Dermatology.
The objective of the study was to measure the performance of smartphone apps that evaluate photographs of skin lesions and provide the user with feedback about the likelihood of malignancy. The investigators used four unidentified apps to analyze images of 188 moles, including 60 melanomas, which had already been evaluated by a board certified dermopathologist (a pathologist with special training in diagnosing skin diseases). Three of the apps, priced at $5 or less, used algorithms to analyze moles.
The main outcome measures were the sensitivity, specificity, and positive and negative predictive values of the four apps designed to aid non-clinician users in determining whether their skin lesion is benign or malignant. The researchers found that the sensitivity of the four tested apps ranged from 6.8% to 98.1%; the specificity ranged from 30.4% to 93.7%; the positive predictive value ranged from 33.3% to 42.1%; and the negative predictive value, ranged from 65.4% to 97.0%. Sensitivity measures the proportion of actual positives which are correctly identified as such (i.e., the percentage of patients who are correctly identified as having a melanoma). Specificity measures the proportion of negatives which are correctly identified (i.e., the percentage of individuals who are correctly identified as not having a melanoma). Predictive value of tests is the probability of a target condition (i.e., melanoma) given by the result of a test, often in regard to medical tests. A positive predictive value refers to the predictive value of the presence of a condition, whereas, the negative predictive value refers to the predictive value of its absence.
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 was found for applications that use automated algorithms to analyze images. The investigators concluded that the performance of smartphone applications in assessing melanoma risk is highly variable; furthermore, three of the four smartphone applications incorrectly classified 30% or more of melanomas as unconcerning. They cautioned that reliance on these applications, which are not subject to regulatory oversight, in lieu of medical consultation can delay the diagnosis of a melanoma and harm users.
Take home message:
Interestingly, the only app that had any real value was the one that sent images to a dermatologist; thus, it was comparable to taking a photograph with your smartphone—or other camera—and e-mailing it to a dermatologist. The other apps that attempted to diagnose a melanoma via an algorithm fell far short of the mark. It behooves one who has a suspicious mole to be evaluate by a board-certified diagnostic. Also, anyone who has suffered a melanoma should have regular follow-up with a dermatologist.
Reference: JAMA Dermatology