Pfizer's AI Drug Discovery Cuts Years Off Development Time

Who is the leader in AI drug discovery? Drug discovery and formulation have been plagued by significant challenges, leading to lengthy and expensive development processes.

Traditional methods often rely on serendipity, resulting in a hit-or-miss approach that hampers efficiency. As a result, only 12% of drugs successfully progress from phase-1 clinical trials to FDA approval.

But that’s about to change.

In this era of rapid technological advancement, artificial intelligence (AI) is poised to transform the drug discovery landscape—as it’s already revitalizing the medical field at large. Pfizer is one of the companies that is leading the change.

Pfizer developed the life-saving COVID-19 vaccine in record time, leveraging AI drug discovery’s game-changing potential. This is just one of the company's AI initiatives.

In this article, let’s delve into Pfizer’s AI drug discovery and personalized medicine strategies and what they mean for the future of global healthcare.

Pfizer's Early Adoption and Use of AI in Pharmacovigilance 

When Google’s AI program, AlphaGo, defeated the human champion in the board game Go in 2016, it was seen as a decisive moment in AI and ML. However, Pfizer recognized the importance and necessity of AI even before that. The biopharma giant has been actively using AI in pharmacovigilance since 2014. 

“Artificial intelligence and machine learning enable us to use data to gain insights into disease and increase our understanding of how different patient populations respond differently to disease and therapies.” 

—Anastasia Christianson, Head of AI, ML, Data, and Advanced Analytics at Pfizer

These technologies have been making greater inroads into the company. More recently, Pfizer has doubled down on AI drug discovery and more and is working on bringing transformative medicine to patients. 

Let’s look at two use cases to understand the significance of AI in Pfizer.

AI Drug Discovery

AI drug discovery has always been touted as a disruptive force in what is otherwise a notoriously slow and expensive practice. As an early adopter of the technology, Pfizer is accelerating the process.

It’s important to note that AI is a blanket term encompassing various advanced computing techniques. Machine learning and NLP or natural language processing are the techniques that matter to a pharma company like Pfizer.

Machine learning is how machines learn and carry out functions faster than humans. A single drug testing project can create petabytes of data. On top of that, companies can mine public and proprietary data sets, which further increases the volume of data to deal with. This involves reviewing a large volume of data and recognizing patterns at Pfizer. But even that is an oversimplification.

The link between biology and symptoms of a disease is complicated. Just because a specific drug targets biomechanisms for arthritis, it’s not a given that it’ll have the same results for joint inflammation.

In its R&D department, Pfizer uses a subset of machine learning called “deep learning” to better understand the relationship between disease symptoms and biology.

Austin Huang, Associate Director and the Biomedical Data Science lead at Pfizer, says, “To enable AI to reach breakthroughs, it’s important to teach computers how to “think” abstractly in discovering patterns in large datasets. This is a research field known as deep learning”.

When combined with a large data dataset, the AI software finds a relationship between the data. This narrows down the search area for the “holy grail” and speeds up the drug development process.

‍

Pfizer’s increased investments in ML will focus on developing predictive models and tools, aiming to establish itself as a premier ‘ML Research Hub.’ 

Completing the drug development is only half the battle. The other half involves obtaining regulatory approval to ensure the drug reaches patients. In this regard, Pfizer is also at the forefront of the AI race.

The company uses AI in clinical development to generate documents, tables, and reports for drug approval. This happens in conjunction with the drug discovery process, shortening the time to market.

“In the future, we believe that AI may help us predict what queries regulators are likely to come back with. We may then be able to improve our submissions by predicting in advance what regulators are likely to ask, and coming prepared with those answers ahead of time.”

Boris Braylyan, Vice President and Head of Information Management at Pfizer

‍

To accelerate drug development, Pfizer employs other state-of-the-art computing techniques, like Modeling and simulation (M&S). M&S is a feedback loop-driven computing technique where data links the bench to the bedside. The company successfully used this technique to develop Neurontin (gabapentin).

Recently, it leveraged modeling and simulation to screen over a million protease inhibitor compounds to develop the COVID-19 vaccine. The results are now available.

This end-to-end usage of AI tools grants Pfizer an edge in the overtly competitive pharma marketplace.

AI-Powered Precision Medicine 

The medical community has long longed for personalized medicine. Since every human has a different genetic makeup, the existing one-size-fits-all drug delivery paradigm could be more efficient. With the help of AI drug discovery and more, Pfizer is leading the change again.

The quest to deliver personalized medicine starts in the early clinical development process. Subha Madhava, the Head of Clinical AI/ML and Data Sciences, who joined Pfizer because of its ‘lightspeed thinking,’ shared, 

“Within clinical AI/ML, we’re really driving a paradigm shift in precision medicine. Our focus is on using multimodal data to inform trial design, first-in-human studies [and] our sign of clinical activity studies.”

‍

‍

At its ML Research Hub, the scientists define the core requirements for drug programs. For this, they gather historical trial data, biomarker data, and real-world evidence, which includes EMR or electronic medical records data.

This helps the team define the patient population and tailor the study design.

Once the molecular data sets are prepared, the scientists apply classical and deep learning techniques, deriving valuable insight for each patient subpopulation. This insight allows the team to identify subpopulations that may better respond to specific treatment options. This way, Pfizer personalizes drug delivery, which reduces the chances of side effects.

Last year, Pfizer partnered with Tempus, an AI-powered precision medicine company, to enhance its personalization efforts.

While the approach is a game changer, the company acknowledges the problems that come with it.

One of the problems is privacy. Collecting sensitive patient data risks their privacy, especially during data breaches.

As a solution, Pfizer's privacy policy is embedded into its practices. The company abides by three principles for AI drug discovery and healthcare, the second of which is respecting individuals’ privacy.

It requires the teams to inform patients and their families where the data is coming from and where it will be used. The principle also forces teams to make their systems ‘explainable’ to all stakeholders, including patients. This improves transparency and trust, which are necessary for advancing personalized medicine.

AI Drug Discovery: The Takeaway

Despite such marvelous advancements in AI drug discovery and other technologies, we may be at the tip of the iceberg. With disruptive technologies like generative AI, there’s more radical innovation to come, especially in the healthcare sector.

Pfizer is among the companies that correctly identify and leverage AI’s potential to reap rewards. Pfizer’s success also sends a strong message to other healthcare organizations to take note and get started with AI sooner rather than later. As in the case of Kodak, Blockbuster, and Blackberry, disruptive forces seldom clement companies that fail to adapt.

‍