Clinical trials are an essential part of developing new treatments for disease, but they’re highly complex, costly and prone to failure. It costs an estimated $2.3bn to develop a new drug, with around 90% of trials failing at the clinical stage. In oncology, it’s even higher, with a failure rate of 97%. Yet the process is also ripe for innovation — something that startups across Europe have in their sights.
Dr Irina Babina, CEO of biotech company Concr, which focuses on cancer, says part of the challenge with trial inefficiency is the shift towards precision medicine. “It’s getting more difficult to carry out clinical trials because you’re going after increasingly narrow populations to develop more novel treatments,” she says. “There’s loads of research ongoing but it’s only translating into marginal benefits for patients.”
Concr is just one of the European startups using AI to revolutionise the clinical trial. Its digital twin technology has been developed using data from more than 40k patients with a diversity of cancers, and can be used to model and predict treatment response. “The individual participant is effectively their own control and you can simulate a number of really different scenarios,” Babina says. “What we want to do is enable smaller, efficacious, more informed and ethical clinical trials that are based on the understanding of biology, especially in the earlier stages.”
In 2023, the $150m Debiopharm Innovation Fund invested in Concr, as part of the startup’s £1.94m seed round. Debiopharm’s fund invests in startups looking to revolutionise pharmaceutical R&D, an approach that exposes its in-house clinical team to some of the newest technologies on the market and creates a blueprint for what the future of clinical trials might look like.
Dr Heidi Nauwelaerts, Debiopharm’s head of clinical development, says one of the most pressing challenges for her team is complying with all of the different initiatives from regulators, such as the Food and Drug Administration’s (FDA) Project Optimus and In Vitro Diagnostic Regulation in EU. As costs and time scales for clinical trials start to spiral, the team needs to find new ways of working to boost their efficiency. “This results in larger sample sizes and bigger trials, which increases costs or delays trial initiation and thus delays effective treatments going to the market,” Nauwelaerts says. “We need to streamline somehow to be efficient in our clinical development. Addressing these and other challenges requires innovative solutions.”
Her team is currently working with Protocol AI software, developed by the Swiss biotech firm Risklick, which Debiopharm Innovation Fund also invested in, to develop high quality trial protocols. By analysing historical protocol data, the tool is able to recommend the best guidelines for a clinical trial, and digitise all of the relevant research information. Without this sort of system, protocols typically take months to write. One pharma company working with Risklick has been able to reduce the time for protocol development by 50% and estimates it will be able to bring a new drug to market up to a year quicker. Another biotech company was able to develop protocols in just a week.
The Covid-19 pandemic really showed the whole industry how inefficient we are
“It’s the basis for the success or failure of a trial,” Risklick founder Poorya Amini says about the importance of protocols. “Everybody wants to learn from historical data but you can’t do it manually, it’s too time consuming. With Protocol AI, you can do it with one click. Everything has a label, so you can identify an inconsistency in a 600-page document, and you don’t need to input all that information manually into an operational system.”
For startups such as Risklick, it can be difficult to get a foot in the door to work with pharmaceutical companies, Amini adds, particularly as it’s always been a risk-averse sector. But that’s starting to change. “The Covid-19 pandemic really showed the whole industry how inefficient we are,” he says. “And thanks to the advancements we’ve had in AI – with ChatGPT for example – people understand what AI is capable of.” In October, pharma giant Eli Lilly appointed its first Chief AI Officer to support AI initiatives across the company, including clinical trials and manufacturing. “Pharmaceutical companies know they have a problem and they are looking for solutions,” Amini says.
A case in point: Interpreting medical imaging in clinical trials. The process is notoriously subjective and prone to disagreement, and there have been a series of advances in AI being applied to the problem. German entrepreneur Felix Baldauf-Lenschen says it was this that inspired him to co-found Altis Labs, a computational imaging company based in Toronto. “What really irked me was how subjective the radiological interpretation was,” he says. “Radiologists only measure a tiny sample of the tumours and there are high discordance rates. To me, it was just appalling that this was how drugs were being evaluated and prioritised for development.”
Ultimately, we’re all doing this to help patients get access to the most effective drugs, faster
Altis Labs’s tool IPRO, which has been trained on the industry’s largest real-world imaging database (composed of over 180m images), allows research teams to cut through the noise and understand if their drugs are working or not. Debiopharm co-led the startup’s $6m seed funding round in 2023, and the company is also working with pharmaceutical companies Bayer and AstraZeneca.
At a recent European Society for Medical Oncology conference, Altis Labs presented a lung cancer study that showed a 40% increase in the model’s ability to predict survival, compared to traditional tumour size measurements. Clients are now asking if the model can be used in other cancer types, such as colorectal, prostate and breast cancer, as well as other disease areas. “This is where AI-based tools can have a big impact on increasing the statistical power to facilitate smaller, faster trials and earlier readouts,” Baldauf-Lenschen says. “Ultimately, we’re all doing this to help patients get access to the most effective drugs, faster.”
Clinical trials have traditionally operated in isolation within healthcare systems, but they are increasingly enriched by real-world data (RWD) from medical records, providing insights into larger and more diverse populations. The use of RWD to enhance trial design, improve efficiency and support evidence for regulatory approval and reimbursement continues to grow. In Europe, accessing standardised, interoperable data across fragmented healthcare systems remains a challenge. BC Platforms, an early investment of the Debiopharm Innovation Fund, addresses this with its Trusted Research Environment (TRE) platforms. These platforms securely connect researchers with patient data while ensuring compliance with strict privacy regulations such as GDPR.
Key applications include the remote identification and screening of eligible trial participants and the creation of external control arms. For instance, NHS England leverages the platform as part of the Cancer Vaccine Launch Pad to match patients with cancer vaccine clinical trials. External control arms, which use existing patient data as comparators instead of traditional control groups, enable more ethical trials by allocating more patients to the intervention arm, reducing recruitment needs and accelerating timelines.
“It’s hard to cover all aspects of a clinical trial, but using RWD can help mitigate risk and identify the optimal sequence of interventions,” says interim CEO Mikaela Bruhammar.
What’s needed is a wider adoption of these sorts of technologies by the industry. But at the moment, Debiopharm’s Nauwelaerts acknowledges this means extra effort for the clinical teams. “These AI tools are often in exploratory stage so if we want to test [a tool] in a trial, you may need to run parallel tracks, which is not always feasible or affordable.” Using new AI algorithm software to select patients in a clinical trial, for example, also increases regulatory scrutiny and can delay the start of a trial by several months in the EU because of all the regulatory bodies involved. The European Medicines Agency recently clarified guidance on the use of AI in the drug lifecycle, with more work ongoing.
The European Parliament (MEPs) approved the creation of the European Health Data Space (EHDS) is also drafting the implementation of European Health Data Space regulation (EHDS), which aims to provide a framework for health data access and sharing. BC Platforms recently launched an EHDS-ready Trusted Research Environment to help healthcare institutions comply with EHDS. Bruhammar sees other commercial advantages to making these changes. “One of the drivers for the EU to put this new legislation in place is to catch up with the US and become a more competitive continent for clinical and RWD research.”
Babina from Concr says the FDA in the US has been very open to working with startups on this and recently established an AI Council to provide guidance and oversight around AI use. “I’m buoyed by how open the FDA is, and how willing they are to talk about novel technologies. In a startup, time is of the essence. You cannot wait months for a response,” she adds. Concr has validated its technology in 10 clinical trials, and is in the middle of a two-year study into triple-negative breast cancer with the Institute of Cancer Research and the Royal Marsden NHS Foundation Trust.
Nauwelaerts is hopeful for the future but says all sides of the sector must collaborate together to improve trial efficiency. “Digital innovation will revolutionise our clinical development, with more efficient trials, more patient-centric [approaches], and with a higher rate of success. We are very interested and are investing a lot in discovery research. But one company cannot do it on their own,” she says. “We need to work together to do this in the right way.”
Read the orginal article: https://sifted.eu/articles/startups-cvc-clinical-development-brnd/
