News

Video

Ashfaq Marghoob, MD: Artificial Intelligence, Smartphone Use for Pigmented Lesion Classification

Author(s):

This segment of Marghoob’s interview featured a discussion of pigmented lesion identification and use of technology such as AI.

During the Maui Derm NP+PA Fall 2024 conference in Nashville, Ashfaq Marghoob, MD, was included as 1 of 3 experts in dermatology who presented a talk on the diagnosis of pigmented lesions and the differentiation of such lesions as congenital nevi and metastatic melanoma.

Marghoob, director of the Memorial Sloan Kettering regional skin cancer clinic on Long Island, had highlighted some of the most important diagnostic challenges in his prior interview segment.1 Here, Marghoob was asked first about important elements of change that have made it easier for clinicians to diagnose and differentiate pigmented lesions.

“When I first started, there was no photography even,” Marghoob said. “I mean, we just had the clinical eye and then you basically jotted down in your notes what a lesion looked like. But in individuals who had 100 plus lesions, it was impossible to write down the morphology of every lesion…Kodachromes that were projected on a rear view projector were used.”

He noted that when computers became more widely used, images were digitized to compare the appearance of lesions. This somewhat recent evolution occurred rapidly and was quite significant in the field, and Marghoob noted that computer algorithms and artificial intelligence are now being utilized to evaluate such images.

Marghoob later added, however, that AI is viewed by some as a kind of holy grail in which lesion images can be submitted, evaluated, and answers will emerge without much thought from health care providers.

“I don't think that that's realistic, nor do I think that's going to happen in the near future,” Marghoob explained. “I think if people were to rely on that, what will occur is you will find the cancers, but in the process you will also be removing a lot of benign lesions.”

Marghoob noted that while AI sensitivity can be quite strong in its use for melanoma detection on a mass scale, the specificity will prove to be poorer than that of the clinician.
“But where I do think that AI will make our lives a lot easier is, for example, if I have a patient that has 500 lesions on the skin, I can have total body photography done,” Marghoob said. “Then AI can extract all of the lesions off of the skin, evaluate them, and then create ‘risk bins.’ Let's say there'll be 20 lesions at highest risk for melanoma…So instead of having to look at 500 lesions, perhaps you only have to look at 20.”

To find out more about this topic, view the full interview posted above. For related topics from Maui Derm, view our latest coverage of the conference here.

The quotes included in this summary were edited for the purposes of clarity.

Marghoob is supported in part through the Memorial Sloan Kettering Cancer Center institutional National Institutes of Health/National Cancer Institute Cancer Center Support Grant P30CA008748.

References

  1. Smith T. Major Diagnostic Challenges for Pigmented Lesions, with Ashfaq Marghoob, MD. HCPLive. October 1, 2024. https://www.hcplive.com/view/major-diagnostic-challenges-pigmented-lesions-with-ashfaq-marghoob-md. Date accessed: October 2, 2024.
Related Videos
Linda Gillam, MD, MPH | Credit: Atlantic Health System
Jonathan Meyer, MD: Cognitive Gains, Dopamine-Free Schizophrenia Treatment with Xanomeline Trospium Chloride
Chelsie Monroe: Challenges Clinicians Should Consider When Prescribing Muscarinic Modulators for Schizophrenia
Allysa Saggese, NP | Credit: Weill Cornell Medicine
Zobair Younossi, MD, MPH | Credit: American College of Gastroenterology
Thumbnail for schizophrenia special report around approval of Cobenfy.
Thumbnail for schizophrenia special report around approval of Cobenfy.
Thumbnail for schizophrenia special report around approval of Cobenfy.
Thumbnail for schizophrenia special report around approval of Cobenfy.
© 2024 MJH Life Sciences

All rights reserved.