Artificial intelligence for the medicine of the future

Artificial intelligence (AI) offers many opportunities for clinical practice. For example, analyzing data using algorithms can help to better predict disease risks. In addition, new diagnostic tools can be developed based on data. New approaches were also the focus of the two-day AI and entrepreneurship hackathon organized by the Center for Medical Data Integration at the Bosch Health Campus in collaboration with Bosch Corporate Research. The approximately 35 participants worked on two questions using real patient data from the Robert Bosch Hospital, which had been completely anonymized and standardized in compliance with all regulatory requirements. In addition, a business team considered how the findings could be used to develop digital applications or other business ideas.

Focus: intensive care medicine, sleep laboratory, and cardiac MRI

The hackathon focused on two use cases: the evaluation of highly complex data from cardiac MRI for clinical diagnostics, and the interdisciplinary use of vital parameters from the intensive care unit and the sleep laboratory. The cardiology use case focused on whether the 3,000 available data sets reveal deviations from normal blood flow values and can therefore be used to improve the diagnosis of cardiac diseases. “Experienced physicians can recognize pathological changes, but for junior doctors or in time-critical situations, AI-supported evaluation can be enormously valuable,” said Nico Schmid, head of the Center for Medical Data Integration at the Bosch Health Campus.

In the second use case, the development teams focused on the extent to which AI solutions can be transferred between clinical disciplines. The application scenario for this so-called transfer learning was the use of physiological data from the sleep laboratory and the intensive care unit. Both areas continuously record vital parameters such as heart rate, respiratory rate, and oxygen saturation – but in different clinical contexts and with sometimes different objectives. Among other things, this approach offers the opportunity to discover previously unknown connections between different clinical disciplines. Specifically, the goal was to find out whether 36,000 time series from the sleep lab could ultimately help identify critical developments in the intensive care unit so that faster responses could be made in the future. In an earlier study using data from the intensive care unit, researchers at the Center for Medical Data Integration had already shown that AI can be used to predict acute kidney failure after heart surgery at an earlier stage.

Putting AI solutions into practice

In addition to the development teams, there were business teams that conducted targeted research, interviewed clinical practitioners, and developed models for how new products or digital applications could be developed for the two use cases—for example, apps for diagnostics. “Our task in business development at the Bosch Health Campus is to identify specific needs in everyday clinical practice and find economically viable solutions and suitable partners. The hackathon is a good opportunity to test this in relevant use cases,” said Prof. Dr. Ralf Kindervater, Chief Business Development Officer at the Bosch Health Campus.

Win-win situation for research

The collaboration between Bosch Health Campus and Bosch Research is a win-win situation: "Our employees rarely have the opportunity to experience a hospital from the inside. Here, they recognize the complexity of medical data and the importance of clinical expertise. Bosch Research contributes valuable AI methods and application expertise to uncover hidden clues about patients' health risks that would not be detectable without AI. Through this collaboration, we are creating real added value in medical research and patient care," said Dr. Jochen Feichtinger, head of the strategic research portfolio Healthcare Solutions at Bosch Research.

At the same time, researchers at the Bosch Health Campus benefited from the AI and business expertise of Bosch employees. The combination of medical, technical, and entrepreneurial knowledge provides a good starting point for further shaping the digital transformation of medicine for the benefit of patients.