Innovative Biologics R&D Ensures Long-Term Success, Patient Access
By Life Science Connect Editorial Staff
As budgetary restrictions loom large over drug development programs, strategic planning and investment are critical to finding success. In their respective roles, Nadim Ahmed, president & CEO of Cullinan Oncology, and Robert Stoffel, Ph.D., VP of research at Amgen, are helping powerhouse teams strategically invest in blooming technology like artificial intelligence (AI) and machine learning (ML) to make better decisions throughout drug discovery, development, and commercialization.
In a recent digital event, available on-demand here, Bioprocess Online Chief Editor Matt Pillar sat down with Ahmed and Stoffel to glean greater insight into their open-minded approaches to drug discovery, how AI and ML can be used to yield major R&D efficiency gains, and how these gains, early in the process as they may be, ultimately impact the pace to patient access. Per Ahmed: “Our mission is to create new standards of care. Our progress is not planned to be incremental. It's important to get FDA approval and clinical acceptance, but we need to do better. By continuing to strive for therapies that produce highly durable outcomes and more tolerable experiences for patients, the modality agnostic approach gives our teams the freedom to discover and develop a range of diverse cancer therapies that will hopefully transform outcomes for people with cancer.”
Design A Patient-Centric Strategy
A company’s mindset around drug discovery is everything, which is why Ahmed and Stoffel emphasize theirs. At Cullinan Oncology, a midsize biotech, they leverage the aforementioned modality agnostic approach. Essentially, Ahmed’s team focuses first on identifying high-impact cancer targets. From there, they work to find strong modalities that target the immune system and demonstrate the potential to be first or best in class. Cullinan prefers to also remain agnostic to the source of innovation. Impactful ideas don’t have to come from its internal team; in fact, external collaboration is embedded into its research model. By forging partnerships between internal discovery experts and industry and academic partners, it is maximizing its chances of identifying the best possible innovations for patients.
At Amgen, Stoffel’s team has a similar perspective that he describes as “patient first.” While Amgen has already made significant investments in AI and ML to identify therapeutic pathways and targets, Stoffel is also hopeful that it will be able to leverage AI and ML to understand the patients who respond differently to drugs: “The industry tends to look for the win. For instance, if you're taking a drug into oncology or cardiovascular and hitting a certain subpopulation within that, it's a win. You're benefiting those patients, but there are others left behind. As we grow as an industry, we need to make sure that we're looking at the disease as a whole and making sure that we are not leaving any patients behind.”
By also building greater efficiency into their development approach, drug companies experience a number of benefits, including more cost-effective models that significantly reduce the drug timeline and deliver medicine to patients faster. Per Stoffel, “The thing that excites me the most is the things we don't yet know. We’re going to be able to bring different modalities forward and target pathways in a way we've never been able to before.” The key is taking the right risks.
Identify The Right Risks
Both Stoffel and Ahmed are well acquainted with the importance of taking calculated risks to yield sizeable rewards. According to Ahmed, “Our modality agnostic approach has allowed us to develop a diversified pipeline where we're able to spread risk across programs. We also remain rigorously focused on having a high bar for go/no go decisions across the R&D spectrum. Since we're not dependent on a single platform, it means we're less emotionally invested in continuing to advance programs that shouldn't be advanced. The time, energy, and resources spent on programs that don't meet your defined ‘go’ criteria provide little benefit, but this type of decision-making takes discipline.”
Cullinan achieves this discipline via its thriller or killer strategy. When it gets a “killer” result, it opts to kill a program and reallocate its resources. When it gets a “thriller” result, it develops a strategy to accelerate the program and enhance the investment. Ahmed encourages a bold investment strategy despite capital concerns and tough markets: “Cost saving never leads to prosperity, investment in the business does.” Stoffel agrees, noting that there will always be resource constraints. At Amgen, they choose to invest in therapies that demonstrate sizable impacts.
Prioritize Developability & Manufacturability
Once it is established that a molecule can have a positive influence on disease treatment, sponsors must ensure that said molecule is manufacturable. Working across key departments and stakeholders to assess developability and manufacturability is critical. Per Stoffel: “[Developability is] built into our process.”
Ensuring that a molecule is manufacturable is just as much of a factor toward a go or no/go ruling. Per Stoffel: “It’s a partnership within the organization to make sure you can execute the downstream processes that need to happen. You want to understand the drug-like properties so that it can be delivered in a patient-friendly form.”
Due to the importance of ensuring manufacturability, Stoffel and Ahmed advise that companies start on these assessments as early in the process as possible to help determine the best molecule to move forward. Ahmed cautions, “If you don’t make the spend [on manufacturability and developability assessments] early, it can catch up with you later. If you get a good result, you don’t want to be in a position where your manufacturing becomes the rate limiting step.”
Build AI Into The Equation
How and when AI and ML are leveraged during drug development and manufacturing depends on the company. Biotech leaders debate the most efficient stages to implement AI and ML, whether in molecular discovery, molecular design, or commercialization, but Ahmed and Stoffel emphasize that you don’t have to choose. Amgen leverages AI and ML throughout the early phases of its process: “Design happens through ML and AI. From there, you get into automated testing and then the reiteration of that process. Once you get the virtual components, you must validate it within the biological system,” notes Stoffel. Per Ahmed, “In addition to discovery and development, commercialization is another ripe area for applying artificial intelligence. I see a future where we continue to meet our customers where they want to be met.”
As for the choice of whether to build an in-house technology or to outsource your AI needs, that depends on the size of your company, internal knowledge, and how expansive you want your AI implementation to be. Per Stoffel, “Amgen invested in Decode, a large human genetics database, and there's even more investment now to move into supercomputing types of technology. We’re striving to use AI from the very beginning within patient data sets. ML can be used to access and understand how patients are responding or not responding in clinical studies, allowing for the next generation of therapies to come forward.”
Looking Ahead
While the drug development industry works to navigate resource and capital constraints, AI and ML are emerging as powerful technologies to help companies invest smarter and maximize their drug development strategies. Though every pathway is specific to a company’s size and capital runway, the critical decision is identifying where and how you can build in more efficiency to ensure faster patient access.
To hear the entire conversation with Stoffel and Ahmed, tune in to the Bioprocess Online Live virtual event, Bioprocess R&D: Beginning With The End In Mind, available here on-demand.