Donate to Science & Enterprise

S&E on Mastodon

S&E on LinkedIn

S&E on Flipboard

Please share Science & Enterprise

AI-Aided Mobile App Shown to Screen for Melanoma

Detect cancer scrabble

(Marco Verch, Flickr)

7 Feb. 2023. The developer of an app using algorithms to analyze images of skin lesions says clinical trial findings show the system detects melanoma more reliably than some physicians. The company AI Medical Technology in Stockholm, Sweden reported today on its web site initial top-line results of the study, which are not yet peer-reviewed.

AI Medical Technology is a three year-old enterprise developing medical diagnostics using artificial intelligence, starting with a mobile system to identify melanoma from images of the skin. Melanoma is an aggressive form of skin cancer, which while not as common as other skin cancers, is more likely to become malignant and spread to other parts of the body. People with lighter skin shades are more susceptible to the disease. A study in the journal JAMA Dermatology in March 2020 estimates 325,000 new cases of melanoma to occur worldwide that year, leading to 57,000 deaths, with numbers increasing to 510,000 annual cases and 96,000 deaths by 2040.

AI Medical Technology offers a system called Dermalyser that it calls a diagnostic decision support system, designed to assist physicians in detecting and classifying skin lesions as cancerous. The Dermalyser attaches to a smartphone a dermatoscope, a usually hand-held device that illuminates and magnifies small areas of the skin or scalp to identify abnormalities that would not be visible to the naked eye. A physician first enters a patient’s medical details into the app, then points the attached dermatoscope at a lesion and takes a photo. Software in the app, says AI Medical, uses algorithms trained by more than 100,000 images to analyze the submitted image and medical data to return a malignancy risk score in seconds.

First a conventional examination, then the app

The observational clinical trial is still underway, enrolling 500 participants with suspicious skin lesions at 37 primary care sites in Sweden. The study has no control or comparison group. Once enrolled, participants are examined by a physician, who conducts a conventional examination with a dermatoscope, and takes further action, such as a referral to a specialist or making an excision sample for analysis. The primary care physician then uses the Dermalyser to take an image of the same lesion and return results of its analysis.

Findings from the diagnoses by primary care or referral dermatologists and Dermalyser results are then compared to follow-up pathology lab analysis of excisions from the skin lesions made by primary care physicians or dermatologists. AI Medical Technology reports on data from 240 participants showing the Dermalyser app returns results with 95 percent true-positive sensitivity and 86 percent true-negative specificity. The company does not provide comparable data from conventional diagnostics, but says the Dermalyser sensitivity and specificity results outperform primary care physicians and dermatologists.

“The remarkably high sensitivity and specificity levels demonstrate the clinical performance and benefit of Dermalyser,” says AI Medical Tech CEO Christoffer Ekström in a company statement, “particularly since the study was conducted in a real world, primary care setting representing different demographics, personnel, and geographical location.” Magnus Falk, professor of medicine at Linköping University and principal investigator of the study notes, “Results from the study will be reported in a clinical publication, but already the clear impression is a high acceptance of Dermalyser as a diagnostic support tool among the medical personnel involved.”

Ekström adds that AI Medical Tech plans to complete its application for a CE mark or European authorization of the Dermalyser system this year, and at the same time prepare for clinical trials in the U.S.

More from Science & Enterprise:

We designed Science & Enterprise for busy readers including investors, researchers, entrepreneurs, and students. Except for a narrow cookies and privacy strip for first-time visitors, we have no pop-ups blocking the entire page, nor distracting animated GIF graphics. If you want to subscribe for daily email alerts, you can do that here, or find the link in the upper left-hand corner of the desktop page. The site is free, with no paywall. But, of course, donations are gratefully accepted.

*     *     *


Comments are closed.