Using AI to fight diseases

The medical world is set to be transformed. Artificial intelligence (AI) allows huge quantities of data to be analysed - more quickly and more comprehensively than ever before. This makes earlier diagnoses, better therapies and more efficient drug development possible. Pharma companies, research institutions and hospitals in Germany are therefore investing specifically in AI-based solutions. The following examples illustrate how AI is revolutionising healthcare.

Better understanding cancer

One key objective of many AI applications is to identify patterns in complex sets of data – when it comes to the spread of cancer cells, for example. Funded by Germany’s research ministry, the DECIPHER-M project combines data from tissue analyses, imaging and genetic information to better predict the patient’s individual risk of metastasis.

“Despite the advances that have been made, metastasis is one of the biggest challenges,” explains Project Leader Jakob Kather from the Else Kröner Fresenius Center for Digital Health (EKFZ) in Dresden. According to Kather, the goal is to develop a course of therapy tailored to the patient’s individual needs - before the cancer starts spreading in the body.

Screening for breast cancer while preserving data privacy

In many cases, medical data is not allowed to be stored centrally - yet AI needs large volumes of data. The ODELIA project at the EKFZ is therefore focusing on swarm learning, whereby data remains where it is generated but AI models can nonetheless learn collectively. The idea is to enable breast screening results to be reliably analysed while preserving data privacy. An international study even found the collectively trained model to be more accurate than locally developed systems.

Quicker diagnoses via the cloud

Industry is also advancing developments. Bayer for example has developed a platform for radiologists that allows AI-based applications to be incorporated directly into their daily work processes. This helps to automatically identify abnormalities in X-ray or MRI images - in patients who have suffered a stroke or contracted pneumonia, for instance.

“We want to relieve the burden on doctors so that they can focus on their patients,” says Bayer. The platform is already being used in several European countries.

Accelerating drug development

It often takes more than ten years to develop new drugs and place them on the market. Darmstadt-based pharma company Merck wants to use an AI platform to shorten this process to under four years. The technology aims to identify potential drugs more quickly and make it possible to more efficiently plan chemical processes. “AI hugely increases productivity. It can help us make new and more effective pharmaceuticals available to our patients more quickly,” says Walid Mehanna, Chief Data & AI Officer at Merck. He talks of great opportunities - both in medical and business terms.

Quantum AI - a technology of the future

The approach being followed at the Fraunhofer Institute for Cognitive Systems (IKS) in Munich is even more visionary. Together with Ludwig Maximilian University (LMU) Hospital, the IKS is researching hybrid systems that combinequantum computing and AI. The idea is that relatively small amounts of data will enable reliable diagnoses, such as in the case of rare diseases. What are known as Quantum Bayesian Neural Networks will not only deliver results but also take diagnostic uncertainty into account - an important aspect in clinical application practice. The researchers believe they offer great potential, especially when it comes to the early detection and ongoing monitoring of brain tumours.