Please login or create an account. If you do not have access to this content, you will be shown a 30 second preview and licensing options.
- Presentation
Technology in Dermatologic Surgery
Description
The presentation discusses advancements in technology within dermatologic surgery, focusing on patient safety, satisfaction, and application efficiency. It highlights the growing trend of direct-to-consumer applications that allow patients to evaluate skin lesions quickly through their smartphones, facilitated by improved algorithms. The presenter notes that while there's an increasing trust among consumers in such apps, performance can be unreliable, with many studies showing low diagnostic accuracy. The need for transparency and improved data reporting is emphasized, particularly regarding training datasets which often lack diversity and may not perform equally across different skin tones. Additionally, artificial intelligence's role in frozen section pathology for basal cell carcinoma diagnosis is explored, revealing potential but also consistent performance issues in segmentation tasks. The talk concludes with practical applications in communication among surgical teams, noting the effectiveness of headsets for enhancing interaction and efficiency during procedures, despite some minor challenges associated with their use.
View moreConclusions
- The demand for direct-to-consumer applications for dermatologic assessment is on the rise, driven by convenience and increasing skin cancer rates.
- Consumer trust in these applications is high, but the actual performance of many apps is suboptimal, with significant accuracy issues.
- Most direct-to-consumer skin cancer detection apps exhibit poor design and high bias, leading to a mean accuracy of only 59%.
- While some applications show promise, with accuracies up to 81%, there remains a critical need for improved transparency and reliable evidence.
- There is an urgent need for larger, more comprehensive training datasets to ensure the performance of algorithms is equitable across diverse skin tones.
- Efficiency improvements in dermatologic surgery can be achieved through technologies such as headsets, which enhance communication and optimize team performance.
- Image-based AI offers substantial potential in dermatology, particularly for diagnosing basal cell carcinoma, but more research is needed to improve reliability and clinical applicability.
- Daneshjou R et al. Disparities in dermatology Al performance on a diverse, curated clinical image set. Sci Adv. 2022 Aug 12;8(32)
- Sohn et al
- Campanella et al
- van Zohn et al