Evaluation of a systematic review of self‐administered apps used to diagnose skin cancers

Skin cancers are some of the commonest human cancers. Of the skin cancers, melanoma is the most serious diagnosis and has the highest chance of potentially being fatal. Early detection of melanoma has therefore been a goal of dermatologists and public health experts for many years. The development of smartphone applications (apps) that patients can use to survey or scan lesions (such as dark lumps or patches on their own skin) and that are enhanced by new technologies, such as artificial intelligence, and improve the chances of catching melanomas early to avoid poor outcomes would be a welcome addition. This paper is an assessment of a recently published systematic review into the accuracy of studies that have assessed these new apps. Most of the apps can be used to scan pigmented lesions for early identification of melanomas, although some also include other types of skin lesion that might indicate other skin cancers. Overall, the authors concluded that there was not enough evidence that the currently available self‐administered apps used to check skin lesions for melanoma are accurate or effective. Several of the studies were also carried out by researchers with a vested interest in the product being tested. The authors of this paper make several recommendations to improve the value and quality of future test devices. These include declaring any conflict of interest by the authors, the use of a measure called a positive predictive value in order to provide information on the performance of the test and the inclusion of an assessment of the proportion of patients with different skin pigment types in the test population, as this may affect the performance of the app. Linked Article :  Malhi and Yiu. Br J Dermatol 2021; 184 :638–639.