Mix and match of present and future in healthcare

The 18th Medical Data + Pizza took place on April 19, 2022 in Amsterdam UMC. Although the organizer, the Amsterdam Medical Data Science (AMDS) network – with over 1900 members – has organized several online events in recent years, this was only the second ‘live’ event since the COVID-19 outbreak. So it was only natural that those in attendance were hungry for both conversation and pizza.

Building bridges for innovation in healthcare

The first presentation, ‘Startup Village AI in Health hub | The Hub for AI and quantum’ described how one organization has bounced back from the fallout of the pandemic. Ferry Tillekens is strategy and partnerships manager at Startup Village Amsterdam. This now iconic location in the heart of Amsterdam Science Park is where shipping containers form the basis for an ecosystem of tech startups, many of which focus on developing AI and quantum-based products and services.

Since its inception in 2016, the Startup Village has already delivered several success stories. For example, Lalaland is now an established player in sustainable fashion, providing AI-generated fashion models for e-commerce brands. Meanwhile, Dyme continues to grow as a personal finance organization app that also ensures all your subscriptions still offer the best deal.

As a result of COVID-19, the Startup Village was forced to up its game and move beyond its mere landlord position. “The occupancy rate of the village dropped from 95% to 60%. It was then decided to become a network organization that could play a more active role in connecting science, talent and business – for our tenants and beyond,” says Tillekens.

The village’s location in the heart of the Science Park, with its heady mix of academia and more established technology companies – plus its proximity to a range of research hospitals – certainly streamlines the hunt for new partners. “You have a mix of the real world, rooted in the now,” Tillekens notes. “But there’s also CWI, the Dutch national research institute for mathematics and computer science, which conducts research that is 20 years ahead of the market. There’s a lot to be inspired by.”

“It’s all about building bridges,” he continues. With this in mind, one sea container has been transformed into a new ‘AI in Health Hub‘: a meeting place to exchange ideas. Partners include Smart Health Amsterdam, the MedtechPartners Foundation and Nvidia. Tillekens: “We really want to bring partners, students, financiers, startups, health organizations and hospitals together. After all, mixing and matching is really the core of innovation – especially in the Life Sciences and Health sector.”

Ferry is meanwhile happy to report that the occupancy rate is again above 75%. He also invites everyone to attend one of the hub’s upcoming events.

Lighten the workload of doctors

Tom van Sonsbeek, PhD student at the University of Amsterdam, is not yet moving into a sea container at the Startup Village. He works at the AIM Lab (AI for Medical Imaging) – another Science Park research organization that is more concerned with long-term societal impact than short-term market share.

“We’ve seen how much data can change our daily lives, based on what the big tech companies are doing,” says Van Sonsbeek. “The impact that data can have in healthcare can also be immense, but it remains largely unexplored. In this lab, we want to transfer some of this knowledge from the general domain to the medical domain.”

The research that Van Sonsbeek presented on Multi-modal learning from electronic health records and medical images is still in its infancy. It aims to reduce the workload of doctors by taking over some of their more mundane and time-consuming tasks. This results in an ultimate win-win scenario, where doctors can spend more time with their patients.

At present, Van Sonsbeek’s work focuses on one such task: making a diagnosis on the basis of medical images.

In general, AI methods have already proven to be faster and more consistent in interpreting images than doctors. “But there’s one important thing that doctors do better: they can take into account other information about the patient, such as medical history or demographics, for their diagnosis,” says Van Sonsbeek. “This additional information could be essential and help make a more reliable and potentially life-saving diagnosis.”

“That’s why we took on the challenge of creating methods that can process both images and textual information and extract the relevant information from each of them. In this way, we create improved automated diagnostic methods that can process and assess information just like a doctor does.” – while the process becomes more reliable and transparent.”

Back to mixing and matching

Van Sonsbeek’s research results are promising. However, challenges remain, from the limited access to medical data due to privacy concerns, to the messy and inconsistent nature of the data once it is accessible. And then there are the myriad of problems associated with scaling an academic research project into a usable product that would work in a hospital. “And a lot of these things, including coming up with a good interface, are not part of my job description,” says Van Sonsbeek with a smile.

But during the question time, and later at the pizza, many participants of the event appeared eager to help: they offered possible solutions and even their hands-on expertise. With all that mixing and matching, Van Sonsbeek may end up at the Startup Lab sooner than expected.

Check the reports from previous Medical Data + Pizza Meet-up.

The Amsterdam Medical Data Science Group meetings are supported by The Right Data Right Now consortium, which consists of Amsterdam UMC, OLVG, VU University, Pacmed, Amsterdam Economic Board and Smart Health Amsterdam.