Go to AAD Home
Donate For Public and Patients Store Search

Go to AAD Home

Dermatology and dermatopathology awake: now is the time to embrace the artificial intelligence revolution

DII small banner

By Jason B. Lee, MD
June 23, 2021
Vol. 3, No. 25

Dr. Jason Lee - DermWorld Insights and Inquiries
In 1991, the late Dr. A. Bernard Ackerman sounded the alarm about the decline in the interest in dermatopathology by dermatologists in an effort to preserve this foundational subspecialty within dermatology. (1) Although not with the same urgency, Zakhem and co-workers are pleading dermatologists to take a leadership role in the development and implementation of artificial intelligence (AI) in dermatology. (2) In their analysis of studies related to AI and skin cancer, they point out that only 41% of authors were dermatologists and that greater leadership by dermatologists is needed for issues of data collection, data set biases, model assessment, and application of this fast-changing technology.

For many of us, movies have shaped our imagination of AI, from HAL 9000 in 2001: A Space Odyssey to androids in Blade Runner to Ava in Ex Machina. Although self-aware machines may not be a reality anytime soon, the recent advances in transformative AI technology are poised to dramatically change the practice of medicine in the very near future. Sixty years ago, Alan Turing, considered the founding father of AI, posed the question “Can machines think?” in his seminal work. (3) Computer and cognitive scientist John McCarthy coined the term “artificial intelligence” in 1956. (4) He and his coworkers Herbert Simon, Allen Newell, and Marvin Minsky built on the work of Turing, firmly ushering in the era of AI in the late 1950s. Progress has been limited for several decades until recently with the advances in machine learning, a subset of AI in which the computer algorithms make predictions without being explicitly programmed. In 2012, algorithms in deep learning, one of more than a dozen different kinds of machine learning, revolutionized image recognition, the basis for the application of AI to perceptual specialties such as radiology, pathology, and dermatology. (5)

Illustration for DWII on AI

Image from JAAD DOI:https://doi.org/10.1016/j.jaad.2020.01.028.

Large amounts of data are required in the development of AI products. In fact, data is now the new “oil” for the AI revolution of which we are currently in the midst. As radiologists embraced digitalization over 20 years ago, their specialty is rife for the implementation of AI. Recent AI studies in radiology have received prominent attention in the lay press because they show that the performance of AI is equivalent to or better than that of radiologists. (6, 7) The specialty is now at the forefront of AI implementation in the real-world setting as FDA approved AI products are beginning to emerge. (8)

In dermatology, advances in digital cameras and dermatoscopy devices have contributed to a significant increase in the amount of digital data available, although not as structured and standardized as in radiology. There have been attempts to create standardized dermatology image datasets for AI applications, such as the recent effort by The International Skin Imaging Collaboration. AI outperforms dermatologists in the diagnosis of skin cancers, and combining the performances of the two groups results in even better performance without a decrease in specificity. (9, 10) These studies show that AI has a promising potential to improve the diagnosis and management of skin cancers. It is important to note, however, that the results are based on artificially controlled study settings: AI prediction products have not been tested in the real-world setting. It is unknown, for example, whether the image or training datasets used to make predictions in AI studies have general applicability. Parenthetically, a recent review of six AI-based smartphone apps — none of which is FDA approved — found that the performance of the apps was variable and unreliable in detecting melanoma. (11)

Point to Remember: Transformative AI-augmented practice of medicine is here to stay. Dermatologists should not underestimate its eventual impact on the specialty. As Zakhem and co-workers recommend, dermatologists should embrace the technology and contribute to its development and implementation, setting the future stage of the specialty that is more efficient and accessible with improved patient outcomes.

Our expert's viewpoint

Kiran Motaparthi, MD

Deep learning in dermatology is now expanding beyond dermatoscopic recognition of skin cancers; recently, an algorithm was trained to recognize nonneoplastic disorders with an accuracy (diagnosis captured within top 3 possibilities) on par with dermatology residents and faculty. (12) While these developments can be intimidating, understanding the terminology and methodology of studies in deep learning can reduce fears, allowing clinicians to critically evaluate the potential impact of new deep learning algorithms. Additionally, these tools may soon help us improve triage, reduce wait times for appointments, and treat urgent skin disorders earlier. Importantly, in addition to contributing data to the development of accurate deep learning algorithms, dermatologists should also formulate policy and guidelines for their use to assist — rather than replace — us.

  1. Ackerman, A. Dermatology awake—dermatopathology is in peril. J Am Acad Dermatol, 1991; 25:128–130

  2. Zakhem Ga, Fakhoury JW, Motosko CC, Ho RS. Characterizing the role of dermatologists in developing artificial intelligence or assessment of skin cancer: a systemic review. J Am Acad Dermatol. Article in press. https://doi.org/10.1016/j.jaad...

  3. Turing AM. Computing machinery and intelligence. Mind, 1950. 49: 433-460.

  4. Shortliffe E. “Artificial Intelligence in Medicine: Weighing the Accomplishments, Hype, and Promise.” Yearbook of medical informatics 28.1 (2019): 257–262.

  5. Lee, Kai-Fu. AI Superpowers: China, Silicon Valley, and the New World Order. New York. Houghton Mifflin Hartcourt 2018.

  6. Walsh, Fergus. “AI ‘outperforms’ doctors diagnosing breast cancer.” BBC New, Jan 2, 2020. Accessed 11/22/2020.

  7. Grady, Denise. “A.I. Took a Test to Detect Lung Cancer. It Got an A.” New York Times, May 20, 2019. Accessed 11/22.2020.

  8. Zebra Medical Vision Secures its First FDA Clearance in Oncology, Boosting Early Detection of Breast Cancer in Mammograms. Business Wire, Jul 27, 2020. Accessed 11/22/2020.

  9. Han P et al. Augmented intelligence dermatology: deep neural networks empower medical professionals in diagnosing skin cancer and predicting treatment options for 134 skin disorders.” J Inv Dermatol 140.9 (2020): 1753–1761.

  10. Hekler U et al. Superior skin cancer classification by the combination of human and artificial intelligence.” Eur J Can 120 (2019): 114–121.

  11. Freeman, Karoline et al. “Algorithm based smartphone apps to assess risk of skin cancer in adults: systematic review of diagnostic accuracy studies.” BMJ (Clinical research ed.) vol. 368 m127. 10 Feb. 2020, doi:10.1136/bmj.m127

All content found on Dermatology World Insights and Inquiries, including: text, images, video, audio, or other formats, were created for informational purposes only. The content represents the opinions of the authors and should not be interpreted as the official AAD position on any topic addressed. It is not intended to be a substitute for professional medical advice, diagnosis, or treatment.

DW Insights and Inquiries archive

Explore hundreds of Dermatology World Insights and Inquiries articles by clinical area, specific condition, or medical journal source.

Access archive

All content solely developed by the American Academy of Dermatology

The American Academy of Dermatology gratefully acknowledges the support from Bristol Myers Squibb.

Bristol Myers Squibb Logo