Medical Technology: Detecting Diseases with Algorithms (German Health News)

​Four students of the Landshut University, Biomedical Engineering Department, have succeeded in their first project to show that in the future there will be algorithms that will assist radiologists in examining medical images. The field of machine learning in medical technology "is still in its infancy", but is on the rise. The four students Jakob Dexl, Lisa-Marie Kirchner, Maximilian Reiser and Michael Uhl had support from the Praxis Radiologie Mühleninsel from Landshut, who provided the real MRI data of his patients anonymously. The radiologist Prof. dr. Andreas Lienemann and the technical department of the health IT company Cerner Deutschland GmbH supported the project. Radiologists need to use x-ray, ultrasound, or MRI to classify them as healthy and ill by distinguishing tiny details. In the future, this classification and presorting could deprive them of an algorithm that can correctly classify the images based on existing and "trained" data. The students defined image details and features by shapes and sizes of brain structures, so that the program could then differentiate between healthy and sick. But the images of healthy people can vary greatly, so that there were difficulties in the recognition. In a first test, however, the study project was already successfully launched, which indicates a major step forward for medical technology.