It takes between 10 and 15 years and more than €2 billion to bring a drug candidate from discovery through clinical trials and, finally, to the patient. In those circumstances, anything that can speed up the process and reduce costs would be very helpful both for the companies involved and the people who benefit from the new treatments.
Enter AI. Generative AI (GenAI) can, for example, search vast databases of published scientific literature to identify potential new drug candidates in a fraction of the time it would take humans and with greater accuracy. The same technology can also be used to engineer a new drug in silico to determine what its attributes will be in the real world and establish if it might have the desired therapeutic effect.
“Artificial intelligence is transforming drug development across a range of specific use cases, enabling optimised processes and enhanced efficiency, from discovery to clinical trials and post-marketing,” says Tom O’Leary, chief information officer with clinical research organisation ICON. “By utilising predictive models that analyse vast data sets, AI can identify promising drug candidates in record time, predict efficacy and identify optimal patient populations – all crucial steps that streamline the drug discovery and development pipeline.”
The technology has already had life-saving applications: when scientists at Pfizer were developing a Covid-19 vaccine during the pandemic they used a newly available AI tool to speed up the process.
READ MORE
Normally, at the end of a clinical trial, it can take 30 days or more for the patient data to be cleaned up to enable scientists to analyse the results. That cleansing process generally involves data scientists manually inspecting the data sets to check for coding errors and other inconsistencies in tens of millions of data points. By using a tool known as Smart Data Query (SDQ), the vaccine clinical trial data was ready to be reviewed within 22 hours. According to Pfizer, that took a full month off the development and approval process for the vaccine.
The technology also has applications for the trials themselves, says O’Leary: “In clinical trials, AI currently aids in trial design, automates data capture and analysis, and improves patient recruitment by matching patient profiles with trial requirements more precisely, which is improving diverse representation in trial populations.
“AI plays a key role in patient monitoring and trial management, enabling real-time tracking and anomaly detection, which helps ensure high data quality and compliance.”
He adds: “With automation of routine tasks and faster insights from complex data, AI allows researchers to focus on strategic decisions and analysis.”
ICON has developed its own suite of AI-driven solutions for use at various stages of clinical development and commercialisation. One of those solutions, OneSearch, is designed to improve site selection and enrolment.
“Enrolment can account for up to one-third of the trial’s timeline, and delays here can have significant impacts on the rest of the trial,” O’Leary explains. “The OneSearch platform analyses a massive network of data points, from protocol details to extensive real-world data, to identify the best trial and most connected sites to improve site activation timelines, patient recruitment and cost efficiency. By accelerating and improving the site selection process and identifying sites with higher accessibility to target patient populations, OneSearch reduces the risk of non-enrolment and enhances performance metrics.”
In addition, the company’s ICONex platform accelerates the identification of key opinion leaders (KOLs) in specialised therapeutic areas, such as rare diseases.
“ICONex consolidates millions of scientific publications into a visual, interactive map of expert relationships, reducing the time to identify relevant KOLs from months to seconds,” says O’Leary. “Similarly, SmartQuery, a clinical data science solution using natural language processing, greatly reduces human effort in data review processes for reconciliation of adverse events and concomitant medications.
“As drug development progresses, ICON’s Cassandra platform predicts post-marketing regulatory requirements with over 96 per cent accuracy, enabling early planning for regulatory submissions and post-marketing studies. Each of these AI-powered tools enhances the development process by enabling more informed, data-driven decisions and accelerating key activities.”
The benefits are clear and there is more to come, O’Leary believes. “ICON’s AI-enabled digital ecosystem is enabling more efficient, adaptive and patient-centric trials, and the real-time exchange of insights that drive trial efficiency and reduce risk,” he says.
“It’s an exciting time for innovation. Exploring how we can leverage the full potential of AI will take time and investment, but with the right approach, the benefits will be transformative.”
