Dr. Kimeswenger (Department of Dermatology and Venereology) received the award in recognition of outstanding work focusing on automated tumor image recognition on skin sections using artificial intelligence and neural networks. ADF is the leading society in dermatological research in German-speaking countries (Austria, Switzerland, and Germany).

Could you briefly outline your project?
Dr. Susanne Kimeswenger: White skin cancer is a common type of tumor found in humans. For each individually suspected case we encounter, we surgically remove a section to then be examined and diagnosed by pathologists. This can take a lot of time. In order to attain a diagnosis, we stain very thin sections of specimens so they can be analyzed under a microscope. Many clinics, including the Kepler University Hospital, also digitize these stained sections.

As part of our research, we have developed a computer program - an artificial neural network – to use these high-resolution images and help diagnose white skin cancer. With a high degree of accuracy, the AI can predict whether or not a tumor is present in a pathological specimen. In addition, the program indicates which areas on the slice are important to predict the neural network and what they look like. This ultimately helps pathologists reach a diagnosis at an initial glance and spot specimens that are most likely to be malignant. We were also able to show that a neuronal network partially uses new histological structures to classify a section.

How does your work impact patients?
Dr. Susanne Kimeswenger: In the long run, our program will not only reduce the pathologists’ workload, but also help to prevent errors as well as accelerate and improve diagnoses.

What does winning this award mean to you?
Dr. Susanne Kimeswenger: I am extremely pleased to be receiving the ADF’s Translational Research Award. This award is a personal confirmation of our working group’s research focus. In my opinion, using AI methods in medicine is one of the important developments today.