Klinikum rechts der Isar Technische Universität München
Neuro-Kopf-Zentrum
Abteilung für Neuroradiologie
Forschungsgruppen und Projekte
Human-Centered AI in Medicine (HAIM)
Summary
The rapid evolution of artificial intelligence (AI) is reshaping healthcare, fundamentally changing how clinicians diagnose and treat patients. Our work is dedicated to advancing the understanding of how clinicians and AI systems can collaborate to transform clinical decision-making. Our research focuses on uncovering the key factors that enable effective human-AI interaction, with a particular emphasis on large language models (LLMs), a cutting-edge technology capable of processing natural language to assist in medical tasks. We are driven by the following core research questions:
- How can clinicians and AI systems work together to improve decision-making accuracy and efficiency?
- What potential risks and challenges arise from human-AI collaboration, and how can these be mitigated?
We are particularly interested in how cognitive biases shape a clinician's openness to AI suggestions and how trust in AI systems is built and maintained in clinical practice. In addition, our work investigates the educational frameworks necessary to help healthcare professionals adopt AI tools safely and effectively. Through our work, we aim to unlock the full potential of AI in medicine, ultimately contributing to a more precise, efficient, and patient-centered healthcare system.
Projects
- Automation Bias in AI-assisted detection of cerebral aneurysms on Time-of-Flight MR-angiography
- Human-AI collaboration in LLM-assisted brain MRI differential diagnosis
- Evaluation of the impact of an LLM tutorial on radiology resident’s performance in brain MRI differential diagnosis
- Benchmark study on the performance of open-source LLMs in challenging radiological cases
- Impact of multimodal prompt elements on diagnostic performance of GPT-4(V) in challenging brain MRI cases
Lead
Team
PD Dr. med. Dennis Hedderich;
Prof. Dr. med. Benedikt Wiestler (AI for Image-Guided Diagnosis and Therapy);
Dr. med. Severin Schramm;
cand. med. Jonas Wihl
Partners
- PD Dr. med. Lisa Adams, Prof. Dr. med. Rickmer Braren (Radiology, TUM)
- PD Dr. med. Keno Bressem (German Heart Center Munich)
- Friederike Jungmann (AI in Medicine, TUM)
- Matthias Keicher, Dipl.-Ing. (Computer-Assisted Medical Procedures, TUM)
- Dr. med. Georg Prokop (Neurology, TUM)
- Smart Reporting GmbH, Munich
Funding
- Clinician Scientist Program, TUM University Hospital
Selected Publications
Schramm S, Preis S, Metz MC, Jung K, Schmitz-Koep B, Zimmer C, Wiestler B, Hedderich DM, Kim SH. Impact of Multimodal Prompt Elements on Diagnostic Performance of GPT-4V in Challenging Brain MRI Cases. Radiology. 2025 Jan 21;314(1):e240689.
Kim SH, Schramm S, Adams LC, Braren R, Bressem KK, Keicher M, Platzek PS, Paprottka KJ, Zimmer C, Hedderich DM, Wiestler B. Benchmarking the diagnostic performance of open source LLMs in 1933 Eurorad case reports. npj Digital Medicine. 2025 Feb 12;8(1):97.
Kim SH, Schramm S, Riedel EO, Schmitzer L, Rosenkranz E, Kertels O, Bodden J, Paprottka K, Sepp D, Renz M, Kirschke J, …, Wiestler B, Hedderich DM. Automation bias in AI-assisted detection of cerebral aneurysms on time-of-flight MR angiography. La radiologia medica. 2025 Feb 12:1-2.
Kim SH, Wihl J, Schramm S, Berberich C, Rosenkranz E, Schmitzer L, Serguen K, Klenk C, Lenhart N, Zimmer C, Wiestler B, Hedderich DM. Human-AI collaboration in large language model-assisted brain MRI differential diagnosis: a usability study. European Radiology. 2025 Mar 7:1-2.