Dresden - The work of pathologists will be supported by artificial intelligence (AI) in the medium term. In many areas of cancer diagnostics, internationally active companies are now committed to developing AI-based analyses for the examination of cancer tissue. A Dresden-based startup is now also involved in this business field, working closely with the Institute of Pathology at the Carl Gustav Carus University Hospital Dresden (UKD) and other associated scientific institutions at the clinical campus in Dresden.
As its first product, Katana Labs has developed a solution for an important breast cancer diagnostic method. The so-called "HER2 gene amplification diagnostics" for breast cancer can be decisively accelerated by Katana Labs' AI-supported analysis process. In this process, the AI names PAIKON - consisting of specially trained neuronal networks - analyzes the cancer cells in the biopsies and resected specimens. This enables pathologists to evaluate samples more accurately and quickly.
Meanwhile, PAIKON is undergoing clinical trials at various centers, such as the EMPAIA platform initiated by Charité - an ecosystem for AI-powered applications in pathology - alongside products from other companies. Meanwhile, Katana Labs is working on the development of further AI modules for PAIKON for the analysis of additional cancer entities. Here, AI networks trained on specific detection tasks can improve cancer diagnostics not only by saving time, but also by supporting particularly complex diagnostic methods that enable better patient care all the way to personalized medicine.
At a time when cancer cases are rising tremendously worldwide, medical methods are evolving, and at the same time there is a shortage of well-trained pathologists, technologies like Katana Labs' PAIKON will play a crucial role. Dr. Falk Zakrzewski of the UKD for Pathology sees AI not only as a way to increase efficiency and quality in cancer diagnostics, but also to keep Germany on an equal footing in international competition. Countries such as Israel, the USA, Sweden and Holland have already made progress in this area. The task here is for Germany to catch up.
However, there are also major challenges: Many processes in the field of pathology are still organized in analog form. This makes it difficult to integrate AI systems, for which a minimum level of digitization is required. To achieve this transition, Dresden University Hospital is already actively working to digitize such processes. This includes investments in scanners, automated workflows and other technologies. Dr. Falk Zakrzewski estimates the cost of building this complete final digital infrastructure to be several million euros.
However, these investments are absolutely necessary. Because only with digitized processes and AI-supported assistance systems will it be possible to cope with the emerging "tsunami" of cancer diagnoses and treatments. In particular, the aforementioned general conditions with rising case numbers, increasingly complex diagnostic methods and the shortage of specialists are currently pushing pathology to its limits.
It is inevitable that the future of cancer diagnostics will be digital and AI-assisted. Pathologists will then no longer work at microscopes, but in a digital cockpit with highly specialized measuring instruments. With companies like Katana Labs, Germany could also take its place in this new era of medicine.