ZHAW Involved in DIZH-Funded Project RADICAL: Radiology AI-Driven Clinical Decision-making with Multi-modal Exploration
The ZHAW Institute of Computer Science (InIT) will participate in RADICAL (Radiology AI-Driven Clinical Decision-making with Multi-modal Exploration), a project funded by the Digitalization Initiative of the Zurich Higher Education Institutions (DIZH) through its 2024 innovation program.
RADICAL aims to address a pressing challenge in clinical radiology: the overwhelming complexity and fragmentation of medical data. Radiologists often face a flood of unstructured and disconnected information—ranging from clinical notes and lab results to imaging and medication plans—that makes accurate, timely diagnoses more difficult.
To tackle this, RADICAL is developing a novel AI-based assistant platform capable of interpreting and integrating multi-modal clinical data using natural language queries. This intelligent system will allow radiologists and clinicians to search, interpret, and act on diverse data sources more efficiently, helping ensure critical findings are not overlooked and improving patient care.
AI that Thinks and Acts
The solution builds on the concept of agent-based AI systems—LLM-powered tools that plan, act, and adapt autonomously. These systems break down complex queries, retrieve relevant information across images, text, and structured data, and present verifiable, explainable responses. The project leverages cutting-edge frameworks to enable secure, privacy-conscious integration with real clinical environments.
A Collaborative Effort Across Institutions
RADICAL is an interdisciplinary collaboration between the following participants lead by Farhad Nooralzadeh, Principal Investigator at ZHAW InIT.
The Intelligent Information Systems (IIS) Group at ZHAW InIT
The Krauthammer Lab at the Department of Quantitative Biomedicine, University of Zurich (UZH)
The Institute for Diagnostic and Interventional Radiology, University Hospital Zurich (USZ)
Spanning 2025 to 2028, the project brings together experts in machine learning, natural language processing, radiology, and clinical informatics to co-develop an impactful solution grounded in real-world clinical needs.
Towards Better Outcomes and Training
Beyond supporting clinical workflows, RADICAL also promises benefits for medical education and retrospective research. The platform will enable more accessible case review and cohort identification, making it a valuable tool for students, junior doctors, and medical researchers alike.
Learn more about the project here: RADICAL project details