deepc and University Hospital Basel Bring One of the World’s Most Widely Used Open-Source Medical AI Tools into Enterprise Healthcare Operations

Munich, Germany / Basel, Switzerland - 22 June, 2026 - deepc and University Hospital Basel today announced that TotalSegmentator has been integrated into deepcOS®, deepc’s enterprise-grade clinical AI infrastructure platform, enabling research institutions to deploy and govern the open-source tool for research use within existing healthcare environments.

TotalSegmentator was developed by Dr. Jakob Wasserthal at University Hospital Basel and published in Radiology: Artificial Intelligence. It has since become foundational to medical imaging research globally, with over 300,000 users processing more than 100,000 CT images daily. Its arrival on deepcOS® does more than simplify deployment. It brings institutions into a shared clinical AI infrastructure where open-source tools, commercial solutions, and institution-developed research models coexist, creating the conditions for a more connected and productive research community.

Infrastructure as the Starting Point

For researchers at academic medical centers, access to an AI tool rarely ends the conversation about how to use it. Local compute requirements, PACS connectivity, data governance frameworks, IT prioritization, and post-deployment monitoring for research workflows all sit between a research team and a running deployment. The challenge is no longer simply whether a model can be developed or downloaded, but whether it can be operationalized responsibly for its intended use: connected to existing imaging and data infrastructure, governed consistently, updated safely, and supported across its lifecycle. deepcOS® provides the infrastructure that helps scientific teams move promising open-source and research-developed AI from experimentation toward governed research deployment and, where appropriate, future operational pathways, so they can focus their energy on the research questions that matter.

TotalSegmentator on deepcOS® extends that further. Institutions that already use deepcOS® can deploy it for research use within their existing deepcOS® environment. It arrives pre-configured to the modules their research requires, connected to existing imaging infrastructure, and can be managed alongside other AI solutions and research workflows already present at their site.

Manual segmentation is time-consuming, subject to inter-reader variability, and in many cases simply not feasible at the scale modern research demands," said Dr. Jakob Wasserthal, developer of TotalSegmentator at University Hospital Basel. "The goal was always to build something the research community could actually use. Having TotalSegmentator integrated into deepcOS® means institutions can get straight to the research without the setup getting in the way.

A Tool Built to Be Built Upon

TotalSegmentator automatically segments over 300 anatomical structures in CT and MR images, spanning organs, bones, muscles, vessels, brain structures and body composition tissues, and outputs both segmentation masks and a JSON file with volumetric measurements and mean attenuation values per structure.

Its value to the research community lies in its breadth and flexibility. A spine researcher measuring vertebral bone density, a body composition team quantifying visceral fat and skeletal muscle mass across a large cohort, a cardiovascular group analyzing aortic diameter: all are drawing on the same tool for fundamentally different purposes. Researchers have also used TotalSegmentator as the first layer of their own AI development, running its organ segmentation as the input stage for a downstream model they are building and validating themselves. On deepcOS®, that kind of workflow becomes concrete: institution-developed models, open-source research tools, and commercial AI solutions can run within a single deployment environment, each governed and used according to its respective purpose, status, and institutional requirements.

deepcOS® users configure only the modules relevant to their research. Each institution selects the segmentation tasks they need, keeping deployments purpose-built for the study at hand.

On Research Use Designations in Medical AI

TotalSegmentator is designated for research use only and has not been submitted for CE marking or FDA clearance as a standalone medical device. Its methodology has been peer-reviewed, its training resources are openly described, and it has been widely adopted by the research community. For academic medical centers, it represents the kind of well-evidenced, openly developed scientific tooling that can support serious medical AI research when deployed within appropriate institutional governance.

A Shared Infrastructure for Clinical AI Research

TotalSegmentator is one of the first academic open-source AI tools in the deepcOS® portfolio, and its arrival reflects deepc's commitment to building the infrastructure where the future of clinical AI research takes shape.

TotalSegmentator is exactly the kind of tool deepcOS® was built to support,” said Dr. Franz Pfister, CEO and co-founder of deepc. “The research community has already validated its value at scale. Our role is to provide the clinical AI infrastructure that makes tools like this usable where research and healthcare operations intersect - connected to real data, existing workflows, and governed enterprise environments. This is what we mean by clinical AI infrastructure: not another isolated model or marketplace, but the operating layer on which health systems can build, deploy, monitor, and govern AI.

deepcOS® is infrastructure for the full spectrum of clinical AI: commercial solutions, open-source research tools, and institution-developed models running within the same enterprise-grade environment. For an academic medical center, that translates into a single environment where researchers can leverage open-source tools like TotalSegmentator alongside the commercial AI solutions already deployed at their site, and use both as inputs for the development, validation, and refinement of their own models. Studies can be built, benchmarked, and iterated on within infrastructure that can also support routine clinical operations for appropriately governed and cleared AI workflows, while reducing duplicated IT work and fragmented governance processes.

When the research community works within a common clinical AI infrastructure, the ability to compare models, replicate findings, benchmark performance across institutions, and build iteratively on shared foundations shifts from aspiration to operational reality. That is what deepcOS® is designed to enable.

For AI researchers and developers who have built tools they want to see used, deepcOS® provides the deployment, integration, monitoring, and governance infrastructure that connects that work with the institutions conducting the research.

About deepc

deepc provides clinical AI infrastructure for health systems. Its platform, deepcOS®, is the neutral operating layer healthcare organizations use to build, deploy, monitor, and govern AI workflows across existing clinical, operational, data, and AI environments. deepcOS® helps health systems move beyond isolated AI pilots and point solutions by providing reusable infrastructure for workflow integration, deployment, monitoring, governance, evidence capture, and lifecycle management. The platform supports commercial AI solutions, open-source tools, institution-developed models, and emerging agentic or foundation-model workflows, making it easier to operationalize AI safely and consistently across sites, departments, vendors, and use cases. With deep experience in medical imaging and clinical workflow integration, deepc helps healthcare organizations connect AI into the systems their teams already use - including imaging, reporting, worklists, identity, data platforms, and operational systems - while preserving flexibility, interoperability, and choice.

About University Hospital Basel

University Hospital Basel (Universitätsspital Basel, USB) is one of Switzerland’s leading university hospitals and a central healthcare provider in Northwestern Switzerland. Located in the heart of Basel’s life sciences ecosystem, USB provides highly specialized medical care across a broad range of disciplines and combines patient care, research, teaching, and innovation in close collaboration with the Faculty of Medicine at the University of Basel. With around 8,000 employees, 52 clinics, 15 medical centers, and approximately 40,000 inpatients treated each year, University Hospital Basel plays an important role in advancing university medicine and translating scientific progress into patient care. Its research and innovation activities are focused on improving outcomes for patients through interdisciplinary collaboration, clinical excellence, and partnerships with academic, clinical, and life sciences organizations.

About TotalSegmentator

The University Hospital of Basel (USB) is the backbone of healthcare in Northwestern Switzerland and is regularly recognised as one of the best hospitals in the world. Our patients are at the center of our work. Around 9000 dedicated employees provide a wide range of services around the clock, 365 days a year, from basic care to highly specialized medicine. The USB is one of Switzerland's leading medical centers with high, internationally recognized standards and treats around 43,000 inpatients every year. In addition, there are more than one million outpatient contacts and excellent research in all areas of medical science.