Home > News > Transforming cancer care with Artificial Intelligence

Transforming cancer care with Artificial Intelligence

Published

17 September, 2025

A new national initiative aims to develop a secure, Swiss-hosted infrastructure to enable AI-enhanced precision oncology across Switzerland while ensuring sensitive and personal health data remains protected.

According to the Global Cancer Observatory, cancer remains one of the country’s most pressing public health challenges, with nearly 58,000 new cases and close to 20,000 deaths recorded in 2022. Achieving fully personalized care remains challenging due to fragmented data and limited integration across institutions. A closer coordination across the national healthcare network will result in more effective and equitable treatments for patients.

NAIPO (National AI Initiative for Precision Oncology) responds to this need with an integrated, AI-powered precision oncology platform to transform cancer care delivery. By applying advanced AI models at every stage of the patient journey, it aims to optimize diagnostics, personalize treatments, and support data-driven clinical decision-making. “Building on lessons from previous efforts in precision oncology in Switzerland, our initiative targets the development of novel, clinically informed AI tools by seamlessly integrating a common data platform, continuously adapting robust models, and designing effective clinical interfaces and patient apps.” says Dorina Thanou, lead of the initiative at the EPFL AI Center.

Selected as a Flagship Initiative by Innosuisse, the Swiss Innovation Agency, NAIPO will unfold over four years under the leadership of the EPFL AI Center and ETH AI Center, uniting a large transdisciplinary team from a wide array of institutions including the Swiss Data Science Center (SDSC), the Swiss National Supercomputing Centre (CSCS), the Universities of Applied Sciences and Arts of Northwestern Switzerland, the Bern University of Applied Sciences, the Universities and University Hospitals of Basel, Bern, Geneva, and Zurich, Debiopharm, Roche, SOPHIA GENETICS, Switch, Tune Insight, as well as the regional hospitals of Aarau, Baden, Ticino, Luzern and Winterthur and the private clinics of Hirslanden and Swiss Medical Network. With an expected total cost of CHF 18.9 million, the project will receive approximately CHF 8.25 million in public funding from Innosuisse with the remaining amount coming from the implementation partners.

Overview of the NAIPO partners, Lukas Klein© 2025 EPFL

Transforming cancer research

NAIPO pioneers new AI approaches in cancer research and care, from clinical decision-support agents and large language models for records mining, to foundation models for treatment response prediction and privacy-preserving approaches. “Combined with high-throughput experimental models and patient avatars, these technologies will allow us to capture and model each patient’s uniqueness.The program will redefine AI’s role in medicine and strengthen Switzerland’s position as a leader in medical AI innovation” said Elisa Oricchio, director of the Swiss Institute of Experimental Cancer Research (ISREC) at EPFL

“Tailoring predictions and recommendations to individual patients is one of the most exciting aspects of NAIPO,” said Charlotte Bunne, professor at EPFL working on model development. “Our models will continuously learn from curated biomedical literature, as well as from individual biological and clinical data to identify potential new targets, biomarkers, and investigational drugs. Novel AI-driven insights will be integrated with clinically validated models and translated into decision-support systems.” Placing patients’ specific needs at the center of the initiative, dedicated solutions will be developed, such as a mobile app, to enhance communication and ensure patients remain actively informed and engaged throughout their care.

Deployment and long-term vision

The program’s roadmap foresees clinical pilots at university and cantonal hospitals and private clinics, leading to an initial rollout at participating hospitals nationwide within four years. In addition to advancing cancer care, the infrastructure is intended to serve as a model for future applications in other disease domains.

“This initiative marks a transition toward a proactive model for precision oncology,” said Olivier Michielin, Head of Precision Oncology at Geneva University Hospitals (HUG) and Clinical Co-Coordinator of the project. “It reflects a commitment to ensuring that all patients, regardless of where they are treated within this network, benefit from the latest advances in AI-supported medicine.”

Secure, privacy-conscious collaboration is central to the initiative. Using modern data governance, the infrastructure will enable collective intelligence without centralizing sensitive health data. “We’re creating a secure and federated system that allows collaboration across institutions without compromising privacy,” said Nora Toussaint, Lead Health & Biomedical at the Swiss Data Science Center (SDSC). “Trust and transparency will be built into the design.”

“NAIPO is exactly what clinical oncology needs today. We are able to produce much more data than a couple of years ago, but we often don’t know how to integrate this in actual patient care. NAIPO is instrumental to close this gap.” Says Andreas Wicki, oncology professor at the University of Zurich and Clinical Co-Coordinator of the project.

NAIPO’s long-term vision includes reducing disparities in access, accelerating the discovery of new biomarkers and treatments, and supporting sustainable innovation across the Swiss healthcare system. Milestones and key results will be shared as the project progresses.

Author: Mélissa Anchisi
Source: EPFL

EPFL researchers create an AI model that thinks like we do

An EPFL team has created a new Large Language Model that is structured similarly to a human brain, allowing users more control and moving away from “black box” AI.

When a standard Large Language Model (LLM) is confronted with a problem, it tries to solve it by matching it to similar information it has seen before, and then give an answer based on those past patterns. But how it decides which information to use and what value it gives to different pieces of information can be somewhat inscrutable from the outside.

The LLM MiCRo (Mixture of Cognitive Reasoners) is architecturally divided into four specialized areas that act like different parts of the human brain, allowing users to have more control over how it approaches a question, and to better understand how it comes to its answers. The model, which was presented at the International Conference on Learning Representations, comes from the NLP Lab, part of the School of Computer and Communication Sciences (IC), and the NeuroAI Lab, part of IC and the School of Life Sciences at EPFL.

The four experts

To create MiCRo, researchers identified four regions of the brains specializing in different functions, which they call ‘experts’: language, logic, social reasoning, and world knowledge.

“The brain is organized into specialized regions, each tuned to handle a specific function. So far, we don’t see this division of labor as clearly in current language models,” says Badr AlKhamissi, a PhD candidate leading this research. “We picked four brain regions that neuroscientists know well and gave the model its own specialized modules, each one trained to be analogous to one of those brain regions.”

An LLM usually functions as a stack of layers that a problem or question can be processed through. In the case of MiCRo, each layer is divided into the four different experts. You give a sentence to the model starting at layer one, for example “The cat is asleep”. Then within this layer, the router can choose one expert for the first word “the”, but a different epxert for second word “cat” and so on, making it modular and highly adaptable.

“Each word of a sentence can go to different experts,” AlKhamissi explains. “So one sentence can actually be processed by multiple experts at each layer.”

Consider a prompt like: “Emma wants to split a CHF 60 dinner bill among three friends, but she knows that Jake lost his job last week and is too proud to say he’s struggling.” A purely mathematical module handles the arithmetic: CHF 60 divided by three is CHF 20 each. But the social reasoning module picks up on something subtler: Emma’s awareness of Jake’s situation, his unspoken pride, and the implicit suggestion that she might quietly cover his share. Both kinds of reasoning are needed to fully understand what’s going on, and in MiCRo, each aspect of the prompt is routed to the expert best equipped to handle it.

“When we see how the model works, we can see that it routes the words that relate to the social aspects to the social expert, and when it does the mathematical part, it routes those numbers to the logic expert.”

This separation makes it easier to see how the model is ‘thinking’ and why it makes certain decisions. It also means decisions can be steered – for example, you can decide to increase the impact of the social expert, or suppress the logic expert, depending on what kind of model you want to use in a certain situation.

“In traditional LLMs, you can do this via prompting by telling the model to make the output more social or make it more related to emotions,” AlKhamissi says. “But here, this is done by intervening in the architecture itself without doing any prompting.”

“A virtuous circle”

To create MiCRo, the EPFL team worked with Greta Tuckute, a neuroscientist from Harvard and MIT, to understand which parts of the human brain are activated by different problems, and then applied that learning to the model.

To identify the region analogous to the ‘logic’ expert in the brain, neuroscientists give humans demanding tasks, such as hard mathematical equations, and less demanding tasks, like easy mathematical equations, and then recorded their brain activity to find which brain regions are the most active for the demanding tasks versus non-demanding tasks. AlKhamissi’s team then did the same for the model, giving it demanding mathematical equations to see which experts would be most activated.

“The cool thing is we just used exactly what they do in neuroscience, but in the model. And the model was able to identify those experts on its own.”

While neuroscience informs the model, the model also informs the understanding of the brain, potentially allowing neuroscientists to discover the contributions of different areas for a given problem or question; for example that a certain sentence activates the language areas 20%, the mathematical areas 50%, and the social reasoning areas 40%.

“For my PhD work, I have been interested in this virtuous circle between neuroscience and AI. In one direction, we use findings and insights from neuroscience about the brain and integrate them into language models,” AlKhamissi says, “and now, with models like MiCRo, we can explore the other direction and ask how we can use AI models to help us understand the brain in a better way.”

Author: Stephanie Parker
Source: EPFL

Smarter waste sorting with AI

EPFL startup WasteFlow has developed an AI-powered copilot that identifies and measures waste streams, helping sorting facilities work more efficiently. Support from several EPFL entrepreneurship programs helped the company accelerate the development of its technology.

(more…)

EPFL launches the world’s first fully open medical LLMs

MeditronFO is the first fully open framework for building medical large language models, to make AI in healthcare more transparent and accountable.