Early diagnosis of breast cancer

Breast cancer is one of the most common malignant cancers in women. In Germany, almost 70,000 new cases are recorded every year. At the Carl Gustav Carus University Hospital in Dresden, an imaging technology developed by ScreenPoint Medical is used to diagnose breast cancer.  The technology has been trained with data from more than five million mammograms taken in the USA and Europe. The AI-based software matches the smallest nodules and calcifications, which can be precursors of cancer, with this data and thus enables even earlier diagnosis. This increases the chance of recovery when cancer is diagnosed and thus the number of women who survive breast cancer. „The examination using artificial intelligence provides additional safety and will become the standard in diagnostics in the future. Dresden University Medicine is thus once again setting new standards in terms of modern, forward-looking therapies and diagnostics. Now it is our task to further evaluate the application in practice in the coming years“, says Prof Michael Albrecht, Medical Director at the University Hospital.

The clinically tested deep learning algorithms are to be continuously improved in the future. In addition to faster and earlier breast cancer detection, the aim is also to minimise false findings and take the different breast densities of women into account even more intensively.

„So far, one radiologist with software has achieved results similar to the four-eyes principle of two radiologists“, says Senior Physician Dr Sophia Blum, Head of Mammography. She has been working with the new technology since November. Blum checks the patient's mammography images on a screen, while Transpara simultaneously analyses the images and then classifies them into one of three categories. „L“ as in Low means a low risk, „I“ as in Intermediate means a medium risk and „E“ as in Elevated means an increased risk of breast cancer. The corresponding areas are precisely marked by the software in the image. „In 90 per cent of cases, the detected focus is actually a carcinoma“, says Blum.

Source: https://transkript.de/medtech-zwo-artikel/2024/fruehe-diagnose-von-brustkrebs/