Vara's explosive new study on AI's power in combating breast cancer

Plus some exciting fundraising news from the Merantix Capital venture.

Vara Founder and CTO Stefan Bunk

Here’s an impressive statistic: 40% of breast cancer screenings in Germany are managed on Vara’s AI platform.

That figure is a testament to the rapid success that Vara has had since being founded in 2018 by CEO Jonas Muff and CTO Stefan Bunk here at Merantix Capital.

But things are just getting started. In an explosive new study with The Lancet, Vara once again has demonstrated the power of artificial intelligence when it comes to catching cancer. “Different AI strategies can indeed increase cancer detection rates by up to 8.6% and dramatically reduce the workload for radiologists—by as much as 84% in some cases,” Stefan told me in a recent conversation.

Furthermore, on Tuesday, Vara announced that it has raised an additional funding round of $8.9 million USD (€8M EUR), which it will use to further its expansion into emerging healthcare markets, with India as the primary focus. The company last year launched a partnership with NM Medical, a leading provider of imaging services in Mumbai, Bangalore, and Pune.

I sat down with Stefan to chat about the study, Vara’s fundraising news, and AI’s impact on cancer treatment, to find out what’s next for this auspicious company. Our edited conversation is below:

Adrian: You just published an explosive new study in The Lancet Digital Health about breast cancer screening. Can you tell me a little bit about how this study came together and what it found?

Stefan: A few years ago, we published our first big study in Lancet Digital Health about combining the strengths of AI and radiologists. In that study, we analysed how AI can improve radiologists and how that affects the radiologists main metrics (sensitivity and specificity). However, in practice it’s not one radiologist alone who reads mammograms (x-rays of the breast). In breast cancer screening, every case is read by two radiologists independently and if at least one radiologist finds suspicious lesions, there is follow up diagnosis via for example ultrasound or MRI. Therefore, just helping one radiologist actually does not directly help women get their breast cancer detected earlier. What if the radiologist with the help of our AI finds a cancer – but the other radiologist anyways would have found it? Then the AI did not actually detect cancers earlier.

Therefore, in this study, we took a full screening program perspective and analysed the metrics that truly matter: breast cancer detection rate and low false positive rate for women participating in screening. In this screening program perspective, an important open research question also is how AI and radiologists best work together. Should AI take over the work of one radiologist? Or should the AI focus on the easy cases, leaving the harder ones for radiologists? What is the best strategy to combine AI and radiologists? Partnering with the University of Cambridge and The Royal London Hospital, we used data from over 1.9 million women in the UK, Sweden, and Germany to analyse various AI integration strategies within existing mammography programs. What we found was promising: different AI strategies can indeed increase cancer detection rates by up to 8.6% and dramatically reduce the workload for radiologists—by as much as 84% in some cases. This research not only demonstrates AI’s standalone potential but also underscores how hybrid approaches that combine AI and radiologist expertise yield the best results, enhancing screening accuracy and efficiency for global healthcare systems.

What does the study mean in practice for your business?

Our goal has always been to leverage AI to make breast cancer screening more accessible and effective, and the study’s findings validate our strategy. The research provides strong evidence that AI can play a practical, impactful role in mammography screening already today, and can create win-win situations for everyone involved: higher detection rates and lower false positive rates for women, lower workload for healthcare providers. We’re already integrating these findings to expand our technology’s reach and effectiveness, focusing on settings where access to skilled radiologists is limited, such as emerging markets. This study equips us to scale confidently and underscores Vara’s role as a leader in AI-driven breast cancer screening.

What do you think AI’s impact will ultimately be when it comes to cancer screening in terms of impact? What do you think cancer treatment will look like in 5-10 years?

AI has transformative potential in cancer screening, especially as it scales to support early detection on a global level. We envision AI not just as a supportive tool but as a critical component of standard care that enables screenings to be quicker, more precise, and accessible even in regions with limited resources. Looking five to ten years ahead, we foresee AI models being applied in a highly personalised fashion, where screenings and treatments are tailored to individual risk factors and health profiles. This shift could bring earlier and more targeted interventions, drastically improving patient outcomes and reducing the burden on healthcare providers.

Longevity is a big topic right now. Do you think AI will expand the human lifespan?

Absolutely, AI has the potential to extend human lifespan, though it’s likely to be through improving quality of life and early disease detection rather than adding years outright. With its ability to detect diseases like cancer at early, more treatable stages, AI can significantly reduce mortality rates and enhance patient outcomes. Additionally, AI can optimise treatment plans, making healthcare more proactive and personalised. While it’s early to predict the full extent of AI’s role in longevity, its preventive and precision-based capabilities certainly pave the way for healthier, longer lives.

You’re a German company originally but have recently launched in India. Can you tell me about why you’ve targeted that market and the traction you’ve seen there thus far?

India faces a pressing need for accessible breast cancer screening due to a shortage of radiologists and limited early detection options, particularly in rural areas. Expanding into this market allows us to directly address these challenges by providing AI-driven screening that supports radiologists and improves early diagnosis. Since launching, we’ve seen positive traction, with healthcare providers increasingly adopting our technology. This early adoption underscores AI’s potential to improve accessibility and outcomes in India’s healthcare system, helping make critical screening services available to more people.

You recently announced new fundraising. Can you tell me a little bit about what that means for Vara and how you’re going to use the money?

This new round of funding is a major step forward for Vara, allowing us to deepen our R&D efforts and accelerate our expansion into emerging markets where our technology can make the greatest impact. We’ll be investing in refining our AI models, making them more adaptable and effective across diverse healthcare environments. Additionally, these funds will support the growth of our screening infrastructure and partnerships worldwide, bringing us closer to our goal of universal, data-driven breast cancer screening. Ultimately, this funding fuels our mission to improve early cancer detection and patient outcomes on a global scale.