Artificial Intelligence: The Next Big Thing in Dermatology?

 

 

Artificial Intelligence: The Next Big Thing in Dermatology?
Randie Kim, MD

From smart phones to smart homes, technological advancements powered by artificial intelligence (AI) now permeate our daily lives. Currently, AI is making headlines with text-to-image generators, chat bots that simulate human speech, and programs that compose essays, poems, and even music. The medical field, which is rich in vast amounts of data, is primed for an AI revolution. Are dermatologists ready?

First, it is helpful to understand the basics of AI. In its simplest form, AI is nothing more than an algorithm, or a set of instructions that is provided to a computer to complete a task. An “if/then” statement, for example, is a simple algorithm. But as computer science advances, so have the algorithms. Called “machine learning” or “deep learning” algorithms, these programs are capable of identifying patterns from data and making predictions or outputs with minimal human supervision.

In 2017, researchers from Stanford trained a type of machine learning algorithm called a convolutional neural network (CNN) to classify clinical images of skin lesions as "benign” or “malignant”. This landmark study showed the CNN performed as well as dermatologists on distinguishing benign nevi from melanoma¹. Many of the subsequent AI-based research papers in dermatology follow this general framework in that an AI-based algorithm (of which there are many types) analyzes a digitized image (which can be a clinical image, a dermoscopic image², or histopathological image³) and provides an output (e.g. benign vs. malignant, high-risk vs. low-risk, mutation vs. wild type). When trained on a large high-quality dataset, it’s becoming clear that AI has the potential of performing these tasks as well or even better than humans.

Now, there are mobile apps that use AI to analyze user-submitted photos of skin lesions in order to triage suspicious lesions for skin cancer⁴. DermAssist, a Google-developed app, has already attained a CE mark for use as a Class I medical device in the European Union. However, none of these apps are currently FDA approved in the United States.

But the real question is: To what extent will AI impact our daily clinical practice?  None of us know the answer to that yet, but as a field, we are putting ourselves in a position to be prepared. Rather than using the term “artificial intelligence,” the American Academy of Dermatology “supports the development of augmented intelligence (AuI) technology provided that it is designed and evaluated in a manner that enables the delivery of high-quality care to patients.” In their position statement, they defined AuI as “a concept that focuses on AI’s assistive role, emphasizing that AuI is designed to enhance human intelligence and the physician/patient relationship rather than replace it.” Or to put it more plainly, as happy as I am that I can FaceTime my mom - who lives 3,000 miles away - nothing will ever beat an in-person hug!

 

References:  

1. Esteva, A., Kuprel, B., Novoa, R. et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature 542, 115–118 (2017). https://doi.org/10.1038/nature21056
   
   
2. Combalia M, Codella N, Rotemberg V, Carrera C, Dusza S, Gutman D, Helba B, Kittler H, Kurtansky NR, Liopyris K, Marchetti MA, Podlipnik S, Puig S, Rinner C, Tschandl P, Weber J, Halpern A, Malvehy J. Validation of artificial intelligence prediction models for skin cancer diagnosis using dermoscopy images: the 2019 International Skin Imaging Collaboration Grand Challenge. Lancet Digit Health. 2022 May;4(5):e330-e339. doi: 10.1016/S2589-7500(22)00021-8. PMID: 35461690; PMCID: PMC9295694.
   
   
3.  Chen SB, Novoa RA. Artificial intelligence for dermatopathology: Current trends and the road ahead. Semin Diagn Pathol. 2022 Jul;39(4):298-304. doi: 10.1053/j.semdp.2022.01.003. Epub 2022 Jan 14. PMID: 35065872.
    
4. Kränke T, Tripolt-Droschl K, Röd L, Hofmann-Wellenhof R, Koppitz M, Tripolt M. New AI-algorithms on smartphones to detect skin cancer in a clinical setting-A validation study. PLoS One. 2023 Feb 15;18(2):e0280670. doi: 10.1371/journal.pone.0280670. PMID: 36791068; PMCID: PMC9931135.