Artificial intelligence app shows promise in detecting skin changes

The integration of artificial intelligence (AI) into dermatology is rapidly evolving, with hopes of improving the diagnosis and monitoring of skin conditions ranging from benign conditions such as acne and eczema to serious problems such as skin cancer.¹ With the increasing prevalence of skin changes and increasing public awareness about skin health, the demand for efficient and reliable tools to monitor these changes has become more important than ever.² The rise of smartphone applications designed for skin checks reflects a consumer-driven interest in innovative solutions. However, these applications must meet specific criteria, including usability, accessibility, affordability and scientific validity.³

A recent study examined the findings of the SkinChange.AI (SCAI) app, which uses a novel AI-assisted approach to identify skin changes over time. The study provides insight into the potential of such technologies in helping dermatologists and other physicians in their practices.⁴

Study design and methods

The pilot study, conducted at the Danish Skin Cancer Research Center, aimed to assess the feasibility of the SCAI app in detecting simulated skin changes. The study involved 24 healthy adults, aged 19 to 62 years, with Fitzpatrick skin types I-III. In a controlled environment, the researchers applied adhesive test spots of different colors (black, brown and red) to the participants’ backs and legs. The SCAI app captured images before and after the application of these spotlights, ensuring standardized conditions through a tailor-made lighting setup.

The SCAI app features an interface for standardized image capture and an AI backend that facilitates image alignment and comparison. The app uses advanced AI algorithms for background identification and spot detection, allowing comparisons between paired images.

Researchers stated that the primary outcome measures were the sensitivity and specificity of the app in identifying the test sites applied. Statistical analyzes provided descriptive data, including predictive values ​​for app performance.

Results

The results of the pilot study showed that the SCAI app demonstrated a sensitivity of 92.0% and a specificity of 95.5%. Researchers said the positive predictive value (PPV) was recorded at 38.0%, while the negative predictive value (NPV) reached 99.7%. This indicates that while the app successfully identified real skin changes, it also generated a notable number of false positives.

Researchers said a classification of the results based on anatomical location showed increased sensitivity on the back compared to the legs, attributed to differences in image capture distance and background complexity. Additionally, the study found that the app showed better performance in identifying dark test spots than brown ones, suggesting that color differentiation could be a crucial factor in the app’s efficacy.

Discussion

The study shows that the SCAI app has potential as a tool for improving the monitoring of skin changes, especially in populations that may not have regular access to dermatological care. The app’s ability to provide a high NPV means it can reliably rule out non-changes, a crucial aspect in clinical settings to avoid unnecessary anxiety and healthcare burden associated with false positives.

However, researchers said the low PPV raises important considerations. They noted that a high false positive rate may lead to overdiagnosis and increased workload for dermatologists, as they will have to evaluate numerous marked changes that may not require clinical intervention. Researchers say this balance between sensitivity and specificity is a common challenge in AI applications and should be addressed in future iterations of the app.

Implications for clinical practice

The research shows that integrating AI tools such as the SCAI app into dermatology could increase patient engagement in self-monitoring skin changes, especially in populations at risk for skin cancer. The app’s design allows users to capture high-quality images while maintaining standardized conditions, which they say is crucial for consistent monitoring over time.

Furthermore, as AI tools evolve, continued training and validation of algorithms with diverse data sets will be necessary to improve the detection accuracy of different skin lesions beyond the simulated changes in the study. The study stated that future studies should focus on real-world applicability, assessing the app’s performance on actual clinical lesions and broader demographics to ensure robustness across different skin types and conditions.

Conclusion

In conclusion, the study shows that the SCAI app shows promising capabilities in detecting simulated skin changes with high sensitivity and specificity. While the app’s potential as a clinical tool in dermatology is clear, researchers believe that addressing the challenges of false positives and broadening its application to real clinical settings are imperative for its successful integration into daily routines. dermatological practice.

References

1. Liopyris K, Gregoriou S, Dias J, et al. Artificial intelligence in dermatology: challenges and perspectives. Dermatol Ther (Heidelb). 2022;12(12):2637-2651. doi:10.1007/s13555-022-00833-8
2. Furriel BCRS, Oliveira BD, Prôa R, et al. Artificial intelligence for the detection and classification of skin cancer for a clinical setting: a systematic review. Front Med (Lausanne). 2024;10:1305954. Published January 8, 2024. doi:10.3389/fmed.2023.1305954
3. Freeman K, Dinnes J, Chuchu N, et al. Algorithm-based smartphone apps to assess skin cancer risk in adults: systematic review of diagnostic accuracy studies [published correction appears in BMJ. 2020 Feb 25;368:m645. doi: 10.1136/bmj.m645]. BMJ. 2020;368:m127. Published February 10, 2020. doi:10.1136/bmj.m127
4. Grove GL, Reedtz G, Vangsgaard B, et al. Artificial intelligence smartphone application for simulated skin change detection: an in vivo pilot study. Skin examination technology. doi: 10.1111/srt.70056