MultivisionDx’s diagnostic tests match cancer patients with therapies that actually work
MultivisionDx is developing a highly scalable, AI-powered tool to support treatment decisions in solid cancers. The company is among 9 startups that were selected to Health Incubator Helsinki in the spring of 2025.
In 2022 alone, more than 20 million people received a cancer diagnosis. At the same time, the number of cancer drugs entering the market is higher than ever before. As the number of available treatments continues to rise, so does the complexity of choosing the right one: especially since most novel therapies are effective for only a small fraction of patients.
Helsinki-based startup MultivisionDx is working on an AI-based tool to help clinicians pick the most effective treatment from the sea of options. If everything goes according to plan, the technology will be available on the market already in 2026.
“We are using cancer biology to match patients with the treatments that will benefit them most. Our diagnostic tests reveal patients’ risk profiles and offer oncologists the data for better treatment decisions, while minimizing unnecessary costs and harmful side effects”, Karolina Punovuori, CEO of MultivisionDx, explains the science behind the technology.
From research lab to spin-out company
MultivisionDx traces its origins to Professor Sara Wickström’s lab at the University of Helsinki, where the research team worked with large cohorts of cancer patient tissue images.
Punovuori joined the project shortly after finishing her PhD at the University of Edinburgh.
“After my doctoral studies in stem cell biology, I wanted to do something with a more direct link to human health. I looked for projects in cancer biology, which significantly overlaps with stem cell research, and found this one”, Punovuori recalls.
As the team worked with the tissue images, they quickly realized that no existing tools could handle the level of analysis required. So, they built their own – and that led them to identify a previously unrecognized combination of markers that accurately predicted how patients would respond to different treatments.
That discovery marked the beginning of a new phase: turning a scientific discovery into a clinical solution. The team began preparing for commercialization through Business Finland’s Research to Business program, and in April 2025, the project officially became a company. Alongside Punovuori, the founding team includes Head of Data Science Fabien Bertillot, Head of Engineering Janos Lengyel, Science Advisor Sara Wickström and Business Advisor Michael Wittinger.
Today, MultivisionDx’s technology has progressed to the point where it can analyze complex protein patterns at single-cell resolution.
“Our tool offers a highly detailed view of what’s happening inside tumor tissue. It can analyze biological processes within the tumor itself and also in the surrounding microenvironment”, Punovuori says.
Validated in head and neck cancer
MultivisionDx aims to launch its diagnostic tool next year. Before entering the market, the company is preparing for analytical clinical validation to ensure the technology works just as well in clinical settings and new patient samples as it does in academic research.
The team has already validated the technology using a cohort of 650 patients with head and neck cancer.
“We’ve focused on head and neck cancers first because they’re highly heterogeneous, and at the moment, there are no diagnostic tests available to guide treatment selection. As a result, patients often get standard therapies that only work for about half of them. The side effects of these treatments, such as difficulty in breathing, can be severe”, Punovuori explains.
Validation results suggest that MultivisionDx’s technology could significantly reshape treatment strategies, particularly for aggressive and late-stage cancers.
“We discovered that roughly one in five head and neck cancer patients with early-stage disease responded very poorly to current therapies. This indicates that their cancers were more aggressive than initially thought. These patients should be identified early and directed to more intensive treatment, but that’s not possible with current diagnostic tools.”
The MultivisionDx’s technology also makes it possible to recognize late-stage cancer patients that are unlikely to benefit from commonly used cancer therapies, such as chemotherapy. This allows clinicians to shift course earlier – and spares patients the burden of ineffective treatments.
Now, the company is already applying its technology to a 2,000-patient colorectal cancer cohort.
“Our platform scales across various types of solid tumors, and even to other diseases where tissue biopsies are part of the diagnostic process. Our goal is to build a portfolio of diagnostic tests for a wide range of disease indications,” Punovuori says.
Peer recommendation led the way to Health Incubator Helsinki
MultivisionDx joined the Health Incubator Helsinki program in spring 2025, following a strong recommendation from another spin-out company.
“We heard that the program had provided them with real value and excellent networking opportunities,” Punovuori says.
“So far, that’s been true for us as well, and we’re very excited to see what lies ahead.”
Interested in collaborating with MultivisionDx? Please contact the team for details.
Contact details:
Karolina Punovuori, CEO
karolina.punovuori(at)multivision.ai
Startup Fast Facts
Name: MultivisionDx
Product: The MultivisionDx computer vision workflow analyzes images of tumor biopsies, generating patient-specific tumor profiles describing the unique biology of each patient’s disease.
Company founded: 2025
Funding Stage: Pre-seed
Text: Saana Lehtinen, Uoma Media Oy
Photo: MultivisionDx