TU Munich, a Digital Case Study
Technical University of Munich Bridges the Gap Between Research and Clinical Care
Rob Monroe | | 7 min read | Discussion
This is part two of a three-part series showcasing how pathology laboratories in New Zealand, Germany, and Portugal have successfully tackled the transition to digital. Read part one here.
When Katja Steiger started working in the Collaborative Research Center at the Technical University of Munich (TU Munich) more than 10 years ago, she was a rare species of pathologist: a translational pathologist.
“The Collaborative Research Center was searching for pathologists with animal models experience,” Steiger recalls. “I applied for the job because I'm trained as a veterinary pathologist but have always worked on both animal and human medicine. This experience made me a slightly exotic person at the time.”
Now, a growing number of pathologists worldwide specialize in translational medicine – a relatively new field that connects basic medical research with clinical treatment. Like Steiger, they are “translational pathologists,” focusing on the translation of laboratory findings into clinical practice. At TU Munich, five veterinary pathologists and seven technical assistants have joined Steiger at the Collaborative Research Center in the last decade. This team complements the 20 pathologists at TU Munich’s Institute of Pathology, a department responsible for diagnostic pathology that receives more than 30,000 clinical cases annually.
“The Technical University of Munich built a large facility for comparative experimental pathology that supports animal pathology for translational research as well as comparative animal–human studies,” notes Steiger. “We work up tissues and read the slides for the researchers that are working in cooperation with us. In this setup, digital is fundamental to our work.”
The beginning of the journey – assisting education
In 2012, TU Munich began its digital journey with a foundational project for the European Council. The aim of the project was to teach biology PhD candidates and postdocs about pathology in animal models – an initiative that necessitated digital slides. “We scanned the slides in Munich and sent them by hard drive to Finland,” Steiger recollects with a smile. “It was basic, but it worked. We continued with this protocol for a few years.”
Steiger emphasizes that slide digitization is key in the research setting. “We’ve learned how important it is to have slides available to researchers; digital images make this possible. When images are digital, we don’t have to sit at the microscope together to work together. We can be in different locations at TU Munich – or even working remotely – and access the same slide at the same time, or even sequentially, to facilitate dialogue, exchange information, and annotate the images digitally.”
From that initial project, the TU Munich team developed an affinity for digital. “People got used to it,” Steiger continues. “In our research environment, they started to ask, ‘Can we digitize the slides from our mice? From our animals? Or can we get H&E sections from human patients to compare them?’ Over the next four or five years, we were able to increasingly respond to these requests, which resulted in expanding our capabilities in comparative experimental pathology with a variety of models, such as xenografts, genetically engineered mouse models, and genetically engineered pig models. Using digital pathology, we could do a standard pathological workup and easily compare the findings, especially from the genetically engineered models.”
The halfway point – impact on cancer research
By 2016, the volume of requests for digital images compelled the research team at TU Munich to purchase a high-throughput scanner. In 2019, the organization added another as demand for digital continued to grow. “In concert with the expansion of digital slides, we implemented a database so people could access their slides. This was used quite often,” Steiger adds. “We are situated in the pathology department so we have full access to the archive, making it easy to compare these findings to cancer cases of real human patients. We can pick up the slides and the blocks, and we can provide, for example, biomarker validation.”
Concurrent to the research pathology teams’ digital work, TU Munich’s clinical pathology team had begun to integrate the digital process into its workflow by scanning slides for archiving and retrospective analysis. “I find the quality of digital images is excellent, even for very small structures, such as viruses and tuberculosis,” notes Atsuko Kasajima, a TU Munich clinical pathologist. “In addition to image quality, there is a big benefit in the ability to share and review images remotely. For example, I work with a pathology colleague who often has challenging pancreatic cancer cases. These cases often necessitate accessing slides from external institutions to confirm our diagnosis. With our Aperio GT 450, we can scan other institutions’ slides then return the slides to them, retaining access to the images to retrieve in the future while not cutting into the samples themselves, which are often quite small.”
The COVID-19 global pandemic outbreak in 2020 further accelerated TU Munich’s adoption of digital on a broader scale for both clinical and translational needs. “We had the digital foundation in place and an established collaboration with our technology vendor so we could act fast to support pathologists’ needs during the pandemic,” Steiger recollects.
Steiger’s unit now supports a range of translational cancer researchers working closely with the clinical pathology team. Approximately 30,000 slides per year are digitized for research purposes with storage provided by the organizations’ lab administration center.
“Digital work enables day-to-day exchange between veterinary pathologists and clinical pathologists,” she comments. “For example, if I have a new phenotype in a genetically engineered model, I can just call the human surgical pathologist and ask, ‘Can you please have a look and tell me what you think about this?’ This unique setup enables us to efficiently connect across Europe to enhance our understanding of cancer and support patient care.”
On the horizon – better patient outcomes
TU Munich continues to learn how digital pathology can benefit the institution and its patients, including exploring how it serves as a gateway to computational pathology. The organization recently added a dedicated computational pathology professor for both research and clinical pathology.
“Our computational pathologist frequently interacts with the deputy director of the institute, who is responsible for the lab. Together they focus on performance. What are the network requirements? What are the errors? What can we change to make the process faster and better? To ensure all pathology needs are met, the computational pathologist talks to the pathologists to ensure the approach is working and to gain their feedback and ideas,” Steiger notes.
Computational pathology, including augmented (or artificial) intelligence, has potential benefits across clinical and research, such as enhancing biomarker discovery.
“AI can play a role in biomarker research through computer-aided morphologic assessment,” notes Vivian Tan, a computational pathologist with Leica Biosystems. “What fascinates me about AI is the ability of deep learning to extract a significant amount of information from an H&E image that can increase our understanding of tumor biology, which can translate into the clinical setting to facilitate patient stratification and treatment planning.”
Even as Steiger looks ahead to future applications of digital pathology, she reflects on what she has learned over the past decade. Her top advice? Think early and often about data management: a catch-all term for a set of topics including data throughput. “For example, if you want to scan on a server and provide slides on a server, you need the network environment. At TU Munich, having two high throughput scanners for the research team allowed our pathologists to have the first connection,” she explains.
Digital storage is another important component of data management. “TU Munich’s data management approach recognizes Europe’s regulatory requirements are varied and evolving. For example, the European Union requires storage of research data for at least 10 years while here in Germany there is a local regulation that data must be stored locally, not in a cloud-based system,” Steiger notes. “As a result, we require a storage room and personnel to manage our data, and we must keep revisiting our needs to ensure we plan for appropriate storage capacity and have an ability to easily store and access images.”
Steiger also notes that engaging with colleagues “top-down and bottom-up” is vital to ease adoption and expansion of digital technology across a distributed organization. “You need the support of people in leadership positions as well as the technical people involved day to day. If both those groups are supportive, it’s easier.”
In the end, she concludes, “All our efforts are in service of increasing knowledge on all sorts of conditions and improving public health.”
Pathologist currently serving as Chief Medical Officer for Leica Biosystems and Chief Scientific Officer, Oncology, for Danaher Diagnostics. Monroe is board certified in cytopathology, anatomic pathology, and clinical pathology and holds a PhD in genetics. He has years of experience in the digital pathology space and frequently consults with pathologists around the world.
