In February the US Food and Drug Administration (FDA) granted its first orphan drug designation (specific to drugs for rare conditions) to a drug discovered and designed using artificial intelligence (AI), and a global phase-two trial for the drug is now under way.
Gerard Quinn, vice-president, innovation and informatics, at Icon, says this news is a perfect example of AI exponentially speeding up progress.
“The discovery process for the new drug began in 2020. The preclinical phase of drug discovery can take six years, and it entails many hours of exhaustive laboratory work, starting with as many as 100,000 potential molecules. It’s truly groundbreaking that AI can accelerate that process and arrive quickly at the strongest candidate molecules for onward clinical trials in people,” says Quinn.
The drug discovery and development journey is lengthy, complex and expensive, with many potential points of failure between the initial concept, development, testing and clinical trials until a new drug eventually hits the market. This is true across all therapeutic categories and, to give one example, 97 per cent of cancer drugs fail at the clinical trials stage.
‘A gas emergency would quickly turn into an electricity emergency. It is low-risk, but high-consequence’
The secret to cooking a delicious, fuss free Christmas turkey? You just need a little help
How LEO Digital for Business is helping to boost small business competitiveness
‘I have to believe that this situation is not forever’: stress mounts in homeless parents and children living in claustrophobic one-room accommodation
In an industry where hundreds of millions to billions of dollars are routinely invested in the preclinical stages, the scope for using AI is phenomenal; Morgan Stanley predict a $50 billion opportunity over the next decade.
AI has the potential to create time and cost savings that can revolutionise processes and crucially, results. With the digitisation of medical records, there is scope for the health and pharma sector to benefit from data-driven approaches. Analysis of huge swathes of medical data by AI cannot only save time but also prevent many of the points of failure along the way.
AI can also facilitate drug design by modelling how various molecules will react and interplay with each other, which leads to focusing precious research funding and time on drugs with a higher chance of success.
“AI has enabled huge forward strides in drug discovery by accelerating processes like target identification, compound screening and molecular modelling,” says Liam Doyle, key account manager, life sciences at Schneider Electric.
“Machine-learning algorithms can analyse vast data sets to identify potential drug candidates, reducing the huge time and resources that traditional methods require. It truly is an exciting time to be involved in the pharmaceutical industry – despite the many challenges the sector faces in terms of its rightly complex regulatory and compliance frameworks.”
Quinn says that matching potential trial participants with clinical trials and recruiting them is one area where AI and machine learning (ML) are proving extremely useful.
“AI/ML moves us to a whole new level by analysing billions of data connections in a multidimensional space inaccessible to human thought. When it comes to trial design, the use of AI/ML to predict and stratify patient outcomes is leading to more targeted trials, reduced placebo effect, reduced trial sample sizes and ultimately faster trials with a better chance of success.”
Along with the acceleration of newer treatments moving from clinical trial to launch, Paul Pierotti, data and analytics partner, EY Ireland, has noticed another trend emerging where many new drugs have an outcome-based contract, whereby the life sciences company is remunerated based on the continued health of the patient. This has an impact on the life sciences supply chain, he says, requiring it to be nimbler and more integrated to health services.
“All of this requires thousands of insights and interventions,” says Pierotti. “This is where artificial intelligence can help; bringing together the wealth of data across the supply chain and patient services to deliver unique drug and treatment plans.”
Pierotti has also seen AI at work to create tangible impact with one of EY’s global biopharma clients based in Ireland.
“We leveraged AI to identify yield improvements across their plant network, including contract manufacturer outsourcers. This resulted in a series of potential improvement initiatives and uncovered the possibility of increasing manufacturing output,” he says.
Collaborating with production engineers, the outcome in real terms was a production yield increase of an extra batch for every 20 made.
“For patients in desperate need of these medications this can help to increase the supply available and, potentially, reduce the cost,” says Pierotti.
The biggest challenge to the use of AI in the life sciences sector is the same one that occupies the minds of many, regardless of sector, when it comes to AI – how to use it safely and ethically.
“Perhaps the biggest challenge we see on a day-to-day basis concerns regulatory compliance. Life sciences is a space which is highly regulated – and for very good reasons,” says Doyle. “The sheer pace and scale of AI-enabled development in healthcare will make it very challenging for the regulators to allow for safe progress, transparency, and model validation while safeguarding patients’ health and securing sensitive data.”
Quinn notes that it is an encouraging sign that the US FDA and the European Medicines Agency are engaged in a consultation process across industry, academia and other interest groups to help them develop guidelines on the use of AI/ML in drug development.
“There is growing recognition that we need to create well-curated, managed repositories and build more collaborative partnerships around data and content with the aim of developing the game-changing AI solutions that we all believe can transform the delivery of new treatments to more people, faster,” he says. “The keys to success are focusing on outcomes-led, patient-centric solutions and retaining strong human oversight.”