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

Integrating Complex In Vitro Models Into the Regulatory Pipeline

Blue pharmaceutical capsule with microbeads floating among neuron-like structures, illustrating microphysiological systems.
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: 4 minutes

Complex in vitro models (CIVMs), such as microphysiological systems (MPS), organoids, spheroids and organ/tissue-on-a-chip, present an opportunity to enhance the efficiency and accuracy of drug discovery and development in an ethical manner.


The Critical Path Institute (C-Path), a nonprofit organization created to improve drug development, has been working to facilitate the adoption of MPS as drug development tools (DDT) in regulatory science.


At this year’s Society for Laboratory Automation and Screening (SLAS) annual meeting, Dr. Graham Marsh, scientific director at C-Path, presented the organization’s framework to support a US Food and Drug Administration (FDA) qualification of CIVMs for use in regulatory submissions.


Technology Networks had the pleasure of speaking with Marsh to discuss how the framework was developed and how automation and AI can increase the confidence in CIVMs. 

Molly Coddington (MC):

What is C-Path, and what are its core aims?


Graham Marsh, PhD (GM):

C-Path's mission is to lead collaborations that advance better treatments for people worldwide. Globally recognized as a pioneer in accelerating drug development, C-Path has established numerous international consortia, programs and initiatives that currently include more than 1,600 scientists and representatives from government and regulatory agencies, academia, patient organizations, disease foundations and pharmaceutical and biotech companies.


Uniquely different from a contract research organization or service provider, our neutral collaborations offer a unique front-row seat to the dialogue on the latest in regulatory science from FDA and the European Medicines Agency.


These collaborations allow for information and data sharing, which serves as the foundation for C-Path to spearhead the transformation of such information and data into actionable solutions that address specific unmet needs in the drug development process. Such solutions can include data resources, biomarkers, clinical outcome assessment tools, clinical trial simulators and other quantitative tools.

These tools and solutions help de-risk decision making in the development and regulatory review process of novel medical products.


MC:

Can you discuss how CIVMs could supplement or even replace existing models? 


GM:

Regulatory agencies rely on models for demonstrating drug efficacy and safety. These models have gone through rigorous characterization to ensure confidence in the data that they produce. The stance of regulatory bodies – and I think rightly so – is to be conservative and cautious to choose the in vitro and in vivo data from models that has been qualified and routinely used when evaluating new drugs.


At the same time, scientists are rapidly developing new methods that promise to better represent human biology. MPS bring multiple cell types together with relevant cell–cell interactions and shear forces, and more faithfully reproduce the biology observed in vivo.


We have been working to bridge these two camps and enable the robust characterization of new models so that they can be integrated into the regulatory pipeline.


There is a huge opportunity for these systems to supplement the data that are currently being generated in animals with more human-relevant data. This is particularly important for biologics and gene therapies where the modalities are incredibly human-specific.


We are looking for specific contexts of use for these tools where they have a significant value add/or demonstrated improvement over the current standard assay. We believe that’s where we should focus to qualify potential drug development tools.



MC:

What is the stance of regulatory bodies on the use of MPS in drug discovery?


GM:

The FDA is willing to accept any data that the applicant wants to include in their submission, and, to date, some drugs have moved into clinical trials on the weight of evidence provided by MPS. A nice example is the Sanofi/Hesperos case, where the MPS data from their myelination chip were used to show efficacy and advance a therapeutic.


The FDA has also formalized the Innovative Science and Technology Approaches for New Drugs (ISTAND) pathway for qualifying MPS models as drug development tools.


There are a number of reasons why a developer might choose to qualify their model through this pathway, but I hope that the work that we are doing will make it easier for developers to create validation packages for models that are being included in submissions, and find appropriate contexts of use and drafting qualification plans if they want to move their model into the drug development tool pipeline.


The impression that I get from talking to some colleagues in regulatory agencies is that they are excited by the possibility that these tools will improve patient safety, but we need to make sure they are properly characterized to ensure their appropriate use.

What is the ISTAND Pilot Program?

ISTAND, which launched in 2020, is a program that supports the development of new DDTs to be used in regulatory applications for novel medical products. Examples of submissions that might be considered for ISTAND include the use of MPS, AI-based algorithms or the use of novel digital health technologies, such as wearables, for patient assessment.



MC:

At SLAS 2025, your talk discussed a novel framework to support FDA qualification of CIVMs for use in regulatory submissions. Can you tell us about this framework, how it was developed and its current status of use?


GM:

Our group, the Predictive Safety Testing Consortium (PSTC), was the recipient of a Broad Agency Announcement project contract from the FDA. The project contract was to fund work with the team in the Division of Applied Regulatory Science to perform a landscape analysis of the regulatory readiness of commercially available MPS models and to host a series of public workshops to bring together stakeholders from regulatory agencies, 3D tissue model developers, academics and pharmaceutical industry scientists.


The goal of these public meetings has been to open lines of dialogue between the groups, identify areas of unmet regulatory need and highlight gaps in existing models where CIVMs would improve regulatory evaluation of drugs, whether that’s by improving the ability of the models to predict drug efficacy in a disease model or demonstrating drug safety for moving a new molecule into the clinic.


We’ve taken the output from the first meeting and written a whitepaper on the pathway to use CIVM for regulatory applications. The output from the second workshop is being drafted as a manuscript and updated to the existing whitepaper document.



MC:

What role can automation and AI play in advancing the role of MPS in drug discovery and development? 


GM:

The benefit that automation can provide is increasing the reproducibility of MPS assays and thus increasing the confidence in the models. A key component of validating MPS models for regulatory assessment is building confidence that the data they produce is robust and reproducible. The tools being developed for laboratory automation and liquid handling can eliminate a key source of variability in these systems.


Combining AI/ML models with MPS models has great potential to produce datasets with increased depth and predictivity, as the algorithms promise to find higher-order trends in the data. Computer vision systems and automated analysis tools will enable faster and more thorough processing of large datasets to enable higher throughput screens and the generation of larger and more reproducible and relevant datasets.