Data science and AI accelerate medical development

Two pioneering companies, Agendia and IQVIA, are applying AI and data science to great success. Both make clear it’s a long and bumpy road. But once a foundation is laid, it can result in continual – and occasionally exponential – improvements in terms of patient outcomes.

The long journey from data to impact was the main topic at the most recent edition in the Life Sciences Café event series, organised by the Amsterdam Economic Board and EY. Two presentations inspired the gathered life sciences professionals by showing how artificial intelligence could make drug development and diagnostics faster and more personalised. The drinks and snacks that followed hopefully sparked conversations around future healthcare innovations.

In her opening remarks, Lead Health Gerty Holla explained that the AI and Life Sciences sector continues to boom in the region. Smart Health Amsterdam has now merged with Amsterdam AI, government grant initiative Biotech Booster opened up a second office in Amsterdam, and leading local VC firm Forbion raised €1.35 billion to fund early- and late-stage healthcare-related start-ups.

The co-host, EY’s Frank van de Manakker, also had his eyes on the numbers when he mentioned that the Netherlands now collects vast amounts of healthcare data. “But the challenge here is that we still hardly use it. Data in itself is meaningless. The real journey is turning this data into actionable insights that can provide true value to patients.”

Data science solving a specific use case

IQVIA, a multinational company at the intersection of health information technology and clinical research, is tackling the challenge of using AI for trial recruitment. Martijn Nap , IQVIA’s General Manager for the Netherlands, highlighted these efforts during his presentation, which were summarised in the white paper ‘Finding All the Needles in the Haystack: Technology-Enabled Patient Identification for Clinical Trials‘.

One of the significant issues IQVIA aims to address is the increasing difficulty of finding the right recruits as the number of trials for primary care, specialty, and rare diseases continues to rise. Martijn points out, “In fact, the average likelihood of successfully completing all phases fell to five percent in 2021.” This means that a substantial amount of money is being wasted, and valuable R&D time is being squandered.

To tackle this problem, IQVIA acquired CTcue, a local AI startup located at Amsterdam Science Park. CTcue has developed a tool that applies Natural Language Processing to search through electronic medical records (EMRs) in search of potential trial candidates. The tool allows doctors to extract pseudonymised lists from their EMR systems. CTcue is designed as a tool for doctors, who can operate it independently.

Describing CTcue, Martijn explains, “CTcue is more than just a ‘Google for EMRs’. It can also transform unstructured data into structured tables, essentially turning a hospital database into an actual research database. Additionally, it can harmonise data from various sources for even broader R&D, which is particularly beneficial for larger research institutions with multiple data repositories.”

Currently, CTcue is being implemented in 32 hospitals in the Netherlands, 6 hospitals in Belgium, and the rollout is ongoing in several other European countries.

Finding the right partners

While IQVIA has made progress in deploying CTcue, there are still challenges to overcome. Finding the right partners can be a complex task, particularly in healthcare and academic hospitals where there is often a desire to reinvent the wheel. However, collaboration is crucial, and seeking assistance earlier in the process can lead to more efficient outcomes.

CTcue does not require additional permission to access data because the data already reside within the hospitals, and the doctors already have access to them. However, the installation of CTcue may require assistance and signatures from the IT department of the hospital, which can sometimes cause delays due to their workload.

It’s important to clarify that CTcue does not diagnose patients; that responsibility lies with the doctors. CTcue aids in identifying potential patients, enabling doctors to make accurate diagnoses.

Martijn emphasises the future potential: “I firmly believe that vertical analysis, which combines patient data with other data sources, is the future. While we may never have all the data in a central database, we can work towards making intelligent combinations of data.” By addressing these challenges and harnessing the potential of AI and data science, IQVIA and CTcue are paving the way for more efficient and personalised drug development and diagnostics.

Globe-embracing database

“I must say in terms of challenges, we’ve experienced a tremendous amount of overlap when it comes to accessing data and clinical trials,” says Bastiaan van der Baan, Chief Clinical and BD Officer at Agendia, who has spent over 20 years trying to improve experiences and outcomes for women with breast cancer throughout their treatment journey.

“And I don’t doubt the research on how the Netherlands has the most scattered data – but my last decade in the US led me to assume the US in fact has the most scattered data,” says Bastiaan, with a smile.

Bastiaan often does battle with ingrained institutional thinking. “Cancer diagnosis has become much easier and is done much earlier. This means treatments don’t have to be as aggressive – and patients may be able to avoid the side-effects of chemotherapy. But it’s very difficult to change this mindset. Many doctors still want to throw everything and the kitchen sink at every cancer diagnosis. As a result, you actually need more data to prove the equal effectiveness of less treatment.”

Bastiaan’s company has developed the tests MammaPrint® and BluePrint®, which together provide a holistic view of the biology underlying an individual’s breast cancer, helping their doctors to formulate the best treatment plan. Their platform has now evolved into a worldwide real-world research database that uses AI to enable the discovery of novel genomic profiles to further improve precision in the treatment of breast cancer. And by capturing data from over ten thousand patients (and counting) of all ethnicities, ages, genders, and health statuses, they are ensuring a suitably diverse test group.

Research has already produced actionable insights around the differences between tumours of African and Caucasian patients with HR+ breast cancer, and how age does not seem to affect the gene expression in ER+ cancer. “But when it comes to treatment, there’s always room for improvement,” says Bastiaan.

5 June 2023

More information

Contact us



Want to read more articles like this one?
Follow us daily on LinkedIn, Twitter and sign up for our Board Update.

Read more