We've updated our Privacy Policy to make it clearer how we use your personal data. We use cookies to provide you with a better experience. You can read our Cookie Policy here.

Advertisement

Getting New Drug Modalities to Market: How Can Your Data Strategy Help?

RNA Polymerase II transcribing DNA into RNA.
Credit: iStock.
Listen with
Speechify
0:00
Register for free to listen to this article
Thank you. Listen to this article using the player above.

Want to listen to this article for FREE?

Complete the form below to unlock access to ALL audio articles.

Read time: 11 minutes

From CAR T to radiopharmaceuticals, we are generating new drug discovery modalities faster than you can say “deoxyribonucleic acid”. At this pace of innovation, even relatively recent advances, such as monoclonal and bispecific antibodies, could be seen as traditional modalities.


In this whirlwind of innovation, we are also generating and capturing huge quantities of data, which can be mined for multiple research projects, powered by artificial intelligence (AI) in one form or another. There is clearly an emerging digital modality, which we will be exploring more deeply.


Together, new drug modalities combined with the digital revolution are promising to transform the pharma and biotech landscape. However, like any technological leap, bringing new drug modalities to market comes with challenges. Internal challenges such as staff alignment and data accessibility, coupled with external ones like investment hesitancy, outdated approval processes and regulatory demands, can all delay the delivery of promising drugs to the patients who need them.


Alongside the physical processes of design, production, testing, selection and analysis, the adoption of a unified data strategy is a key element in successfully getting new drug modalities to market.

 

The novel modality revolution and its successes

Over the past two decades, there has been a surge of technological advances and novel drug modalities that promise to overcome the limitations of traditional approaches like monoclonal antibodies (mAbs) and small molecules.1 As of March 2022, there were 12 advanced therapy medicinal products (ATMPs) marketed in the EU and UK, and 482 were in clinical development.2


The boundaries between modalities are increasingly blurring. Antibody drug conjugates (ADCs) and, more recently, radiopharmaceuticals (RDPs) represent novel cross-modality therapeutics. The concept of attaching a small molecule to a larger antibody complex to produce ADCs has evolved into inserting radionuclides into established delivery mechanisms to create RDPs. Unlike traditional external radiotherapy, RDPs use biologic techniques to deliver the radiation payload systemically or locoregionally. Early studies have shown promising results, demonstrating very high efficacy and highlighting the significant potential of these therapies.3,4 As a platform or modality, RDPs open up a vast range of possibilities for treating solid tumors such as prostate, breast and colon cancers.

 

How to create an effective data strategy to support AI efforts 

Research strategies

Drug discovery approaches can be target-based or target-agnostic. Target-based drug discovery (TDD) typically addresses a specific disease by screening candidate molecules based on a chosen modality. In this context, success will lead to profitable results in a known, addressable market, and the timeline is generally set by existing regulatory processes. On the other hand, target-agnostic approaches, such as phenotypic drug discovery (PDD), offer the potential to identify novel therapeutics that work through atypical drug mechanisms.5,6 Examples include groundbreaking medicines for cystic fibrosis, hepatitis C and spinal muscular atrophy.5


Naturally, using an approach like TDD carries the risk of uncertainty about whether the identified molecule will produce the desired, safe clinical outcome. On the other hand, PDD may lead to the development of novel drugs with poorly defined mechanisms of action and targets. Both approaches raise challenges for drug development.5


Some research programs aim to create platforms or methodologies, usually with a specific drug target in mind. A validated platform can deliver multiple therapies using the same standardized techniques.7,8 For example, the highly accelerated global rollout of vaccines during the SARS-CoV-2 pandemic proved the validity of mRNA as a platform, which is now addressing a wide range of new areas including malaria, influenza, Ebola and HIV.9


To put this into perspective, by the end of 2019, no mRNA vaccines had achieved regulatory approval. However, by December 2020, the FDA granted an Emergency Use Authorization for the Pfizer-BioNTech vaccine, accelerating a process that typically takes five years to less than one.9,10


Bringing a new therapeutic modality to market often requires a considerable length of time. The development and commercialization of mAbs, antisense oligonucleotides, small interfering RNAs, gene therapies and gene-editing products take far longer than the 10–12+ years needed for a typical small-molecule drug. Centocor, for example, was founded in 1979 and took nearly 20 years to translate hybridoma technology into market success with mAb infliximab (Remicade).11


Accelerating influence of data

At the heart of the industrialization of drug discovery’s design, production, testing and analysis processes is a data revolution. Labs are shifting from paper notebooks and manual stores of PDFs and documents to electronic lab notebooks (ELNs), connected systems, and integrated project workflows. The goal is to fully digitize these processes, enabling researchers to take advantage of the latest data-to-decision tools, now enhanced by AI.


For example, Google AlphaFold is transforming the analysis of complex protein structures, solving problems that were previously almost intractable. Similarly, AI-driven interpretation of medical imaging is starting to outperform human diagnosticians, based on the extensive library of annotated reference images.12


To advance digital exploration, AI relies on curated datasets such as public databases, academic publications, national health statistics and experimental data captured by ELNs. In practice, the industry is following a proven path pioneered by many manufacturing sectors, using AI’s capacity for pattern recognition and automation to accelerate processes.


AI is already driving two primary discovery channels, both reviewing existing drugs for new therapeutic applications and screening candidate molecules for potential therapies. For example, AI can predict drug efficacy and side effects with a high degree of accuracy, while efficiently managing vast amounts of documents and data that support pharmaceutical development.13


It is well known that bioactive compounds often follow a complex therapeutic journey. For example, thalidomide was introduced as a sedative, withdrawn when it was identified as causing fetal malformations, and is now used to treat cancer (myeloma) and bacterial infections (Mycobacterium leprae).14 Similarly, AI can screen tens of thousands of new candidate molecules, generating filtered lists for investigation and narrowing research programs to more manageable proportions.


Effective data strategies

The FAIR Guiding Principles for scientific data management and stewardship were developed in 2016 by a consortium of scientists and organizations.15 FAIR defines simple guideposts to facilitate good data management, and recommends that all given data points should be:


Findable: Assigned a unique identifier, described with rich metadata, and registered or indexed in a searchable resource

Accessible: Readable by both humans and machines

Interoperable: Stored in broadly applicable format or language

Reusable: Associated with relevant attributes and a detailed provenance, and released with a data usage license


In the drug development industry, unifying and contextualizing data is crucial to support scientific research and innovation. By providing each data point with rich metadata, scientists can convert unmanageable volumes of information into organized, navigable resources suitable for complex analysis by AI tools.


In addition to following FAIR guidelines, flexibility in data management is also paramount to enable AI and ML analysis. Methodologies such as late binding of schema permit the refinement of a data structure right up to the point of analysis, ensuring that the data remains agile and adaptable. This approach can save researchers time and preserve data integrity, leading to accelerated R&D processes. This flexibility is particularly important in drug development, where drug modalities are quickly changing.

 

What are the future challenges to overcome?

While AI can process huge data volumes in mere moments, the make-test-decide cycle remains exclusively human. For biopharmaceutical and biotech companies, investors and other stakeholders, novel modalities require careful consideration of R&D risks and rewards, in addition to new operational and manufacturing models, commercialization strategies, pricing and reimbursement. All of these challenges will continue to demand significant human ingenuity.

 

Will it take another pandemic to bring novel drugs to market?

The development and approval of mRNA-based COVID-19 vaccines were accelerated by the pandemic crisis. With potentially millions of lives at risk, a process that might have taken around five years was completed in less than 12 months. This unusual case proved the possibility of rapid approval, but it has not shifted business-as-usual expectations.


Putting to one side the specter of disease X, the pathogenic pandemic yet to come, there are good reasons for regulatory caution. In the case of a pandemic, the global health risk significantly outweighs the approvals risk. Conversely, for elective therapy, the equation is inverted: sadly, thalidomide was approved for use as a sedative without identifying the tragic fetal risk when prescribed to pregnant mothers.


Drugs developed using traditional modalities, many of which were established more than 30 years ago, are now generating large numbers of new therapies.16 As proof of concept is achieved and clinical trials near completion, a growing array of therapeutics are awaiting commercial launch – and, of course, tens of thousands of patients are awaiting transformative outcomes. At the time of writing, more than two dozen AI-assisted drugs are currently in or entering clinical trials.17


Faster regulatory approvals – or not

Perhaps due to their inherent unfamiliarity, achieving regulatory marketing approval for novel drug modalities can be challenging. The approval success rates for biologics and small molecules, for example, are around 13–15%, whereas the rate for novel modalities is just 4.3%.18


In a series published by Maxwell Tabarrok titled We Need Major, But Not Radical, FDA Reform, the approvals regime is clearly under the microscope, too.19 Will it be possible to create novel, effective and safe drugs in which the FDA will approve and still be successful within the market?


A careful balance is needed in regulatory approvals, to allow new, effective and safe medicines to reach market while retaining a robust system that protects practitioners and ultimately patients from the effects of untested medicines.20


The HSBC Global Investment Summit, looking at the promise of AI, concluded that companies and regulators are moving ahead to ensure that technological developments are safe for patients.21 Similarly, in a paper titled Artificial Intelligence in Pharmaceutical Technology and Drug Delivery Design, Lalitkumar K Vora et al note that AI may also help reduce the need for extensive and costly animal testing—a proposition that will attract significant attention as a means to meet legislation to reduce animal use in research.22 

Summary

What will be the next breakthrough in this whirlwind of drug discovery innovation? Will it be the non-viral CRISPR-Cas9 delivery nanosystem for PCSK9 base-editing in atherosclerosis;23 the novel anti-psychotic for schizophrenia – muscarinic agonist KarXT;24  the BCL-XL-targeted proteolysis targeting chimera (PROTAC) DT2216 for relapsed or refractory malignancies;25 or the use of deubiquitinase-targeting chimeras (DUBTACs) for targeted protein stabilization?26


The astonishing rise of AI means that today’s research is fueling tomorrow's analytics. Keeping up to date with the rapid pace of innovation in drug discovery is a significant challenge for research organizations, and developing a unified data strategy is crucial to meeting these challenges.

 

Revvity Signals Perspective 

At Revvity Signals, we are committed to empowering scientists with reliable software solutions that integrate AI and ML thoughtfully, enhancing rather than replacing human ingenuity. In drug discovery, our tools, like Signals Notebook, Signals Research Suite™ and Signals Synergy, support the generation of AI-ready data, secure data management, and collaborative ecosystems. This approach ensures that AI augments research efforts, fostering more effective and efficient drug discovery processes while remaining mindful of the balance between human expertise and AI capabilities.