Concerns About AI-Based Dermatologic Mobile Apps

Artificial intelligence (AI)–based mobile applications (apps) have demonstrated lots of promise in dermatology and skin cancer; however, an analysis published in JAMA Dermatology argued that more efforts should be dedicated to minimizing potential harms associated with this technology.¹

AI Chip | image credit: AntonKhrupinArt - stock.adobe.com

At the time of this publication (March 7, 2024), none of the available mobile apps had been granted approval by the FDA.¹ Various studies have assessed the impact of AI-based apps, including a 2023 observational study conducted by Smak Gregoor et al that evaluated the influence of mobile apps on skin cancer trends. Their findings demonstrated that users of mobile apps compared with nonusers (controls) had significantly higher annual health care costs related to dermatology (€64.97 vs €43.09; P < .001), double the number of biopsies or excisions (75 vs 34 claims per 1000 persons; P < .001), and a significantly higher proportion of claims submitted for malignant or premalignant skin lesions (OR, 1.3; 95% CI, 1.2-1.4; P < .05), among other factors.² Although the authors mention how these trends can improve outcomes in skin cancer due to an increased chance for early detection, they also allude to the risk for overdiagnosis or overtreatment.

The present authors echo this sentiment, pointing to prior research on apps created for assessing skin cancer risks.¹ At present, the literature presents varied levels of performance, sensitivity, and accuracy, despite their presented benefits for skin cancer detection. With lacking evidence to support the clinical use of this technology, they embarked on a study to overview, identify, and classify the current AI-based mobile apps used for dermatological concerns.

The terms dermatology, derm, and skin were used in the Android and Apple app stores to identify available mobile apps with AI capabilities in November and December of 2023.

A total of 41 apps were identified and eligible for this analysis. Researchers found that the primary audience of these apps was majorly patients/the public (n = 32; 78%), while the remaining apps targeted clinicians (n = 4; 9.76%) or both (n = 5; 12.2%). There were 13 apps (31.7%) specific to the Apple platform, 12 (29.3%) to Android, and 16 (39%) that were available on both.

These apps were primarily designed for detecting skin cancer (n = 14; 34.1%), identifying or diagnosing various conditions of the skin and/or hair (n = 13; 31.7%), and tracking mole changes (n = 7; 17.1%). Other purposes included tracking skin conditions, treating/diagnosing/monitoring acne, managing atopic dermatitis, or sun protection.

There were 10 (24.4%) apps that stated they had the capacity to diagnose; however, the authors pointed out that there was no scientific backing or supporting evidence in these claims. Additionally, 24 (58.5%) of apps were missing information on any testing or training data sets and only 5 included supplemental, supportive, peer-reviewed publications.

The authors noted that “the American Academy of Dermatology warns that apps claiming to diagnose or provide treatment plans can provide inaccurate information. Nevertheless, AI dermatology apps hold promise for improving access and patient outcomes, but in their current state, may cause harm due to the potential risks, lack of consistent validation, and misleading user communication.”

Among the limitations the authors encountered, they mentioned regional variability in app availability in their search on the Apple and Android platforms, as well as their limitation to AI algorithms in English—which potentially led them to exclude or miss other apps around the globe.

As their analysis revealed a number of concerns regarding this technology, as they concluded the authors emphasized: “Addressing challenges in efficacy, safety, and transparency through effective regulation, validation, and standardized evaluation criteria is essential to harness the benefits of these apps while minimizing risks. App developers should, at a minimum, disclose information on the specific AI algorithms used; the datasets used for training, testing, and/or validation; the extent of clinician input; the existence of supporting publications; the use and handling of user-submitted images; and the implementation of measures to safeguard data privacy.”

References

1. Wongvibulsin S, Yan MJ, Pahalyants V, Murphy W, Daneshjou R, Rotemberg V. Current state of dermatology mobile applications with artificial intelligence features: a scoping review. JAMA Dermatol. 2024:e240468. doi:10.1001/jamadermatol.2024.0468

2. Smak Gregoor AM, Sangers TE, Bakker LJ, et al. An artificial intelligence based app for skin cancer detection evaluated in a population based setting. NPJ Digit Med. 2023;6(1):90. doi:10.1038/s41746-023-00831-w