Artificial intelligence (AI) is one of the biggest buzzwords of 2023. With new generative AI platforms hitting the internet for free use, it's no wonder that the technology has quickly become a mainstream phenomenon. It's also picked up steam in the life sciences industry as a way to expedite the drug discovery and development process.
Current processes for drug discovery and development are time-intensive and expensive, taking between 10 to 15 years and $1.5 to $2 billion to bring a new drug to market. But with the power of AI and machine learning (ML) to analyze and manage large bodies of data, it has been estimated that life sciences companies can save nearly $54 billion in research and development costs each year by using the technology.
AI can sift through data from clinical trials, electronic health records and medical publications to find patterns and even predict outcomes at much faster and more efficient rates than humans. It can help R&D teams understand biological mechanisms and underlying diseases of populations and discover novel targets to counteract those diseases. It also has the potential to identify the right therapeutic dose, optimize the selection of suitable subjects for clinical trials, find new avenues for repurposing existing therapies, and can enable life sciences companies to explore de novo drug design.
Advancements in AI-enabled Drug Discovery
Since 2020, investment in intelligent drug discovery and development has grown exponentially. Third-party investment in AI-enabled drug discovery has more than doubled annually for the last five years, jumping from $2.4 billion in 2020 to $5.2 billion in 2021. Life sciences companies are also increasingly investing internally to build their own AI infrastructure to pursue intelligent drug research and development. According to the Food and Drug Administration (FDA), more than 100 drug and biologic application submissions in 2021 had AI/ML components.
Indeed, while AI-enabled drug R&D is still relatively new, there have been a number of key milestones in recent years that indicate intelligent drug development is real and is gaining momentum.
- 2020: First-ever AI-designed drug molecule created by start-up Exscientia entered human clinical trials to treat patients with obsessive-compulsive disorder.
- 2021: AI system called AlphaFold from DeepMind predicted the protein structures for 330,000 proteins, including all proteins in the human genome.
- 2022: Insilico Medicine started Phase I clinical trials for the first-ever AI-discovered molecule based on an AI-discovered novel target.
- 2023: The FDA granted the first-ever Orphan Drug Designation to a drug discovered and designed using AI from Insilico Medicine.
Notably, this past May the FDA published an Initial Discussion Paper seeking feedback from stakeholders on the use of AI and ML in drug discovery and development to inform the regulatory landscape. Recognizing the increased use of AI and ML throughout the drug development life cycle across a range of therapeutic areas and acknowledging that this technology is sure to play a critical role in drug development moving forward. The FDA has stated that it "plans to develop and adopt a flexible risk-based regulatory framework that promotes innovation and protects patient safety."
Criteria for AI/ML in Drug Development
While the FDA appears to be cautiously optimistic about AI/ML drug development, it has outlined key regulatory areas of interest including human-led governance, accountability and transparency; quality, reliability and representativeness of data; model development, performance, monitoring and validation. With this framework in mind, there are a few considerations for AI/ML-led drug discovery and development from FDA's discussion paper that life sciences companies interested in this area should keep in mind.
- Human involvement – The FDA believes that human-led governance, accountability and transparency will be critical to AI-enabled drug discovery to ensure adherence to legal and ethical values. The FDA suggests that risk management plans be used as a form of governance to guide the level of documentation, transparency and explainability of AI/ML models being used and their outputs. Essentially, life sciences companies will need to be able to provide critical insight into how algorithms function and be able to explain and interpret outputs.
- High-quality, reliable, and representative data – A common criticism of AI today is that it can amplify existing bias in data. Life sciences companies using AI/ML models will want to ensure the data they are using is strong and representative of the entire population that the intended therapy is targeting. The FDA is also focused on data privacy and security, ensuring that appropriate measures are in place.
- Criteria for development and assessing models – While the FDA acknowledges the potential of AI/ML to accelerate drug development and make clinical trials more efficient, it is wary of the technology introducing specific risks and harms. Life sciences companies interested in using AI/ML should look to establish criteria for developing AI/ML models that are trustworthy and for assessing models on risk, credibility and complexity.
- Evaluation of model over time – The FDA emphasizes the importance of regularly monitoring AI/ML models and documenting results to ensure the models are reliable, relevant and consistent over time.
- External validation – Finally, the FDA recommends that AI/ML models and algorithms used for drug discovery and development be externally validated using independent data.
Your Roadmap to Success
As AI and ML models continue to advance and mature, their use in drug discovery and development will only continue to grow. According to Boston Consulting Group, biotech companies using an "AI-first" approach had more than 150 small-molecule drugs in discovery and more than 15 in clinical trials as of March 2022. For life sciences companies interested in tapping into the world of AI/ML-enabled drug discovery and development, here are a few steps you can follow to get started.
- Identify the business goals you hope to achieve by integrating AI. Whether you want to use AI/ML to reduce research costs or are interested in encouraging the market, your business goals will impact how you integrate AI/ML into your R&D processes.
- Align AI use cases with business objectives. As outlined previously, AI and ML have many use cases in drug discovery and development such as clinical trial participant selection or de novo drug design. Determine what use case aligns best with the needs of your organization.
- Identify project requirements and assess whether you have the team and technology to meet them. If you're investing in building an AI/ML model, you want to ensure you have the right people and resources to maintain and monitor it. For example, do you have the in-house capabilities to build an algorithm or will you need to engage a third-party partner? Knowing the answers to these questions will help you develop the proper strategy and determine what your investment needs to be.
- Determine how "success" will be objectively measured. Finally, it's always important to outline key performance indicators to guide your AI initiatives. For example, is there a specific ROI you want to achieve? What is the breaking point of an AI investment?
Intelligent drug discovery is a burgeoning area of the market that will only continue to see increased investment. Life sciences companies looking to expedite drug discovery and development while managing costs should consider the benefits – and the drawbacks – of AI and machine learning. Buchanan's life sciences attorneys are attuned to advances in AI-led drug discovery and are equipped to help life sciences companies navigate the changing requirements and regulations of this new market.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
