TL;DR:
- Leading pharmaceutical companies like Amgen, Bayer, and Novartis are embracing AI to expedite clinical trials and cut costs.
- AI is streamlining patient recruitment, reducing trial enrollment times by up to half, and saving millions of dollars.
- The FDA reports a surge in AI applications for drug development in recent years.
- Amgen’s AI tool, ATOMIC, identifies suitable trial patients, cutting enrollment times and aiming to reduce drug development timelines by two years.
- Bayer successfully used AI to link mid-stage trial results with real-world data, saving time and resources.
- AI’s ability to analyze vast amounts of real-world patient data quickly is a game-changer for drug development.
- Bayer is in discussions with regulators to use AI-generated external control arms for pediatric trials.
- Despite potential concerns, regulatory bodies emphasize the need to maintain rigorous evidence standards in drug development.
Main AI News:
In the world of pharmaceuticals, innovation is the key to success. Big Pharma’s quest to develop groundbreaking drugs has now embraced the power of artificial intelligence (AI) to revolutionize clinical trials. This strategic move is not only accelerating drug development but also potentially saving vast sums of money.
The costliest and most time-consuming phase of drug development has traditionally been human studies. The arduous process of patient recruitment and the exhaustive testing of new medicines can take years, consuming over a billion dollars from drug discovery to market readiness.
Pharmaceutical giants like Amgen, Bayer, and Novartis have been dabbling in AI experimentation for several years. Their ambition? To have machines unearth the next blockbuster drug. Although it will take some time for these bets to fully pay off, interviews with industry experts, drug regulators, and AI firms reveal that AI is now playing a substantial and growing role in human drug trials.
These companies are training AI algorithms to sift through vast volumes of public health records, prescription data, medical insurance claims, and their own internal data to identify suitable trial patients. This cutting-edge technology is slashing the time required to enroll participants, in some cases halving it.
Jeffrey Morgan, Managing Director at Deloitte, a trusted advisor to the life sciences industry, comments, “I don’t think it’s pervasive yet, but I think we’re past the experimentation stage.”
The U.S. Food and Drug Administration (FDA) reports receiving approximately 300 applications involving AI or machine learning in drug development between 2016 and 2022, with over 90% of them arriving in the past two years. Most of these applications focus on implementing AI at various stages of clinical development.
Before AI, Amgen engaged in a laborious process of sending surveys to doctors worldwide to identify clinics or hospitals with patients matching specific clinical and demographic criteria for trial participation. However, approximately 80% of studies fell short of recruitment targets due to overestimations, high dropout rates, or patient non-compliance.
Amgen’s AI tool, ATOMIC, now harnesses the power of data to identify and rank clinics and doctors based on their historical performance in recruiting trial patients. What previously took up to 18 months for mid-stage trial enrollment can now be achieved in half the time, Amgen reports.
Amgen plans to employ ATOMIC in the majority of its studies by 2024, with the expectation that AI will help reduce drug development timelines by two years by 2030.
Novartis, too, has streamlined patient enrollment using AI, making it faster, more cost-effective, and more efficient. However, the success of AI in this context heavily depends on the quality and quantity of data available. Less than 25% of health data is publicly accessible for research, according to Sameer Pujari, an AI expert at the World Health Organization.
German pharmaceutical company Bayer successfully utilized AI to reduce the number of participants required for a late-stage trial of its experimental drug, asundexian, designed to mitigate long-term stroke risks in adults. AI linked mid-stage trial results with real-world data from millions of patients, allowing for more accurate predictions and fewer participants in the late-stage trial. This approach saved Bayer millions of dollars and up to nine months in recruitment time.
Bayer now aims to take this innovation further by leveraging real-world patient data to create an external control arm for a pediatric trial of asundexian. This move could eliminate the need for pediatric patients to receive placebos, given the rarity of the condition and ethical considerations.
While the use of external control arms is uncommon, Amgen’s drug Blincyto, designed for a rare form of leukemia, received U.S. approval using this approach, albeit with a follow-up study to confirm its benefits.
AI’s true advantage lies in its ability to rapidly analyze real-world patient data at scale, as Blythe Adamson, Senior Principal Scientist at Roche subsidiary Flatiron Health, explains. Traditional methods could take months to analyze data from 5,000 patients, whereas AI can accomplish the same task for millions of patients in a matter of days.
Bayer is currently in discussions with regulators, including the FDA, regarding the use of AI to create an external control arm for its pediatric trial. Although some experts express concerns about AI-generated external arms, regulatory bodies emphasize that safety and effectiveness standards for drugs remain unchanged.
Conclusion:
The integration of AI in pharmaceutical clinical trials is ushering in a new era of efficiency and cost savings for major players in the industry. This technological advancement has the potential to significantly impact drug development timelines and resource allocation, making it a strategic advantage in the competitive pharmaceutical market. Companies that embrace AI in their clinical trial processes are poised for greater success and faster time-to-market for their innovative drugs.
