Learn how we are Driving Fast, Accurate Diagnosis for All Skin Cancers

Dermalyser is an easy to use, AI-based and clinically validated decision support tool enabling faster and more accurate diagnosis of skin cancers.

Download the article, or read it below.

AI Medical Technology: Digital Assistance for Reliable Skin Cancer Diagnostics

Even skilled, experienced dermatologists make incorrect judgments when examining lesions on the skin to see if they are benign or malignant skin cancer. For patients, these errors can lead to a delay of necessary treatment beyond cure as well as cost time and money, as AI Medical Technology CEO Christoffer Ekstrom knew all too well. Based in Stockholm, Sweden, the company’s intelligent software Dermalyser is able to render a diagnosis of a lesion in an instant with accuracy equal or superior to expert dermatologists.

Despite extensive training and experience, no doctor is infallible. In the field of dermatology, one of the most critical decisions made about a patient is whether a visible skin lesion is benign or malignant skin cancer. The consequences of a false positive or a false negative can range from disconcerting to devastating, and Christoffer Ekstrom, CEO of Swedish start-up AI Medical Technology, is personally familiar with the impact of both outcomes.

A couple of years ago, Ekstrom’s fiancée was concerned about a particular lesion on her body but was told by an experienced dermatologist at a skin clinic in Stockholm that it was benign, not cancer. Reassured but not completely convinced, they sought a second opinion and got a biopsy. Several weeks later when results were returned, the couple’s fear was confirmed, unfortunately the suspicious mole was malignant melanoma skin cancer which may have been ignored indefinitely if the couple had accepted the original diagnosis. On the other hand, not rarely suspicious moles are in primary care mistaken for cancer, causing a so-called overdiagnosis of malignant melanoma. This happened to one of Ekstrom’s close friends, who had suffered several painful excisions for various skin lesions or moles, though all of them had turned out to be benign.

After an immense amount of time spent consulting with colleagues and investigating the market, Ekstrom discovered a great need for augmented diagnostic precision for skin cancer, the answer to which was Artificial Intelligence (AI). AI utilizes massive data to train an algorithm for a designated goal. In this case, by uploading over 100,000 annotated images of benign and malignant lesions into a machine learning database the algorithm is trained to recognize differences between benign and malignant lesions. By using this kind of technology, AI Medical Technology aims to make diagnosis faster and more accurate for any practice. As CTO Ben Clarkson explains, cycling through the data set many times allows the AI to practice millions of decisions, improving its accuracy to surpass the average performance of a general practitioner or dermatologist.

The company’s debut product is Dermalyser, a mobile app that comes bundled with an off-the-shelf phone-compatible dermatoscope. Though the software is the engine behind AI Medical Technology’s solution, the cut of the lens and polarized light in the dermatoscope enables higher optical quality and a deeper view into the skin than a mobile phone camera can achieve alone, giving Dermalyser the best possible image for analysis.

As confident in the AI’s results as he is, Ekstrom is clear about characterizing Dermalyser as a decision support tool, as primary care physicians using the app will have the final say. The app returns a diagnosis of a given lesion in just a few seconds, saving the patient weeks, sometimes months, of dealing with the anxiety of waiting for a diagnosis, not to mention saving the patient from often unnecessary painful biopsies. Ekstrom and Clarkson both believe Dermalyser can help not only patients avoid these unnecessary biopsies but also dermatologists by avoiding referrals and reducing workload in an already hectic and overloaded clinical environment. Ekstrom emphasizes also that Europe is facing and tackling a shortage of dermatologists.

Although Ekstrom initially expected primary care physicians to show skepticism toward AI, particularly concerning diagnosing diseases, the company’s 2021 ten-site feasibility study revealed the opposite. Participants in the study unanimously expressed a need for the kind of aid Dermalyser can provide for their clinical workflow. The ongoing prospective clinical trial dubbed AI-DSMM, currently running at 30 primary care facilities, will serve as clinical evidence validating Dermalyser’s safety and performance and serve as the basis for AI Medical Technology’s CE Mark application in Europe. Preliminary results confirm the excellent accuracy of the AI in distinguishing malignant from benign moles. The diagnostic results are in line with or are better than those of experienced dermatologists. FDA submission and approval in the US are currently under planning and investigation.

Reimbursement for healthcare services varies regionally in the EU, but Ekstrom explains there is already some precedent of recognizing additional value for an AI-based solution. He anticipates that providers in the US will follow a similar approach. As of November 2022, the company has raised nearly two million Euros, primarily from angel investors, with a minority share held by Northern Capsek Ventures AB, a listed tech investment company. In a press release announcing the trial, Ekstrom says, “Our solution shows an outstanding performance, exceeding that of a trained dermatologist, and we now look forward to validating these numbers in a real-world setting by conducting clinical studies.“

Sam Polesie
Specialist Physician and Associate Professor at the Department of Dermatology & Venereology, Sahlgrenska University Hospital, Gothenburg, Sweden.
“The rapid development of AI within dermatology is exciting. However, to date, the promising results observed in the retrospective investigations, have not been matched by prospective clinical trials. We need to remember that it is important to reconcile physicians’ considerations with algorithmic output. In the end, physicians will be in charge of determining what will be best for the patient and what treatment to pursue. Clinical trials that involve primary health care physicians as well as dermatologists are prerequisites for clinical implementation.

For more information, please contact:

Christoffer Ekström, CEO AI Medical Technology
Email: christoffer@aimedtech.org
Cell phone: +46721748339

About malignant melanoma

According to Cancer Today, the incidence of new cases of malignant melanoma in 2020 was over 320,000 worldwide, https://gco.iarc.fr/today/home and every 4 minutes, a person dies of skin cancer. With a growing trend, patient cases are expected to reach almost 500,000 by 2040. Statistics further show that as many as 99% of the cases are curable if they are diagnosed and treated early enough, thus underscoring the importance of continuous self-examination and visiting your doctor in the case of a suspected skin lesion. Hence, early and accurate diagnosis is vital for saving more lives. 

About AI Medical Technology

AI Medical Technology is a company operating in the interdisciplinary fields of data science, software development, and medicine. The company is dedicated to developing AI-powered diagnostic solutions that enable frontline healthcare practitioners to make easier, faster and more reliable diagnoses for their patients. AI Medical Technology was founded by a diverse team of driven, passionate tech entrepreneurs, software developers and clinicians. Following the initial funding round in 2021, the team is now focused on bringing the first product, Dermalyser, a clinically validated decision support tool for the diagnosis of all skin cancers through clinical trials and to the market. https://aimedtech.org/.