The international regulatory environment for drug approvals is diverse, but generally considers the use of computational simulations and predictions to be of value to drug makers throughout the discovery and development lifecycle .
In the United States, the FDA has acknowledged that model-informed drug development (MIDD) approaches can be used to support decisions on whether, when, and how to conduct certain pharmacology studies, and to support dosing recommendations in product labeling, among other applications. The use of Artificial Intelligence and Machine Learning techniques has also been recognized as increasingly playing a key role in drug discovery and development. Data and analysis derived from these approaches are acceptable within submissions for new drug and biologics applications. The presence of this data has supported hundreds of applications, as well as clinical trial design and specific regulatory approval decisions.
This resource informs the structure and standardization of computational analysis. The FDA writes, “This guidance outlines the recommended format and content for a sponsor or applicant to submit physiologically based pharmacokinetic (PBPK) analyses to the FDA to support applications including, but not limited to, investigational new drug applications (INDs), new drug applications (NDAs), biologics license applications (BLAs), or abbreviated new drug applications (ANDAs).” Learn more about how BIOiSIM’s reporting was built to be in compliance with this guideline.
This document, co-authored by VeriSIM Life CSO Dr. Szczepan Baran and Dr. Weida Tong, Director of Division of Bioinformatics and Biostatistics for FDA, emphasizes the need for tailored regulatory measures to accommodate AI’s diverse roles, ensuring AI enhances rather than complicates regulatory processes.
This document, referred to as the “goal letter,” “represents the product of FDA’s discussions with the regulated industry and public stakeholders, as mandated by Congress. The performance and procedural goals and other commitments specified in this letter apply to aspects of the human drug review program that are important for facilitating timely access to safe, effective, and innovative new medicines for patients.”
This paper was released by the FDA in order to address the developing landscape of AI in MIDD and to request input from industry experts. VeriSIM Life responded to this paper with guidance that we provided based on the principles we used to create our computational drug development platform, BIOiSIM.
Research from VeriSIM Life founder & CEO Dr. Jo Varshney, and other VeriSIM Life scientists is cited throughout this article from the AAPS Journal.
This review details the potential of the use of AI in model-informed drug development and encourages organizations to embrace/adopt AI as a way of evolving traditional methods.
This legislation was critical in acknowledging the benefit of, and approving the usage of AI in model-informed drug development. Passed at the end of 2022, it authorizes the FDA to accept safety and effecacy research derived from technologies such as in silico modeling as a suitable alternative to animal testing. Read our response to this legislation here.
This is a list of examples of model-informed drug development approaches that support recent regulatory action.
FDA will build on the success of the “model-informed drug development” (MIDD) approaches by continuing to advance and integrate the development and application of exposure-based, biological, and statistical models derived from preclinical and clinical data sources in drug development and regulatory review.
AtlasGEN Novel Drug Designer efficiently identifies new molecules based on favorable chemical properties and simultaneously evaluates them for clinical safety and effectiveness based on biology using VeriSIM Life’s groundbreaking Translational Index™ technology.
Iterative ranking of target engaging hits by their Translational Index values reduces the time and cost associated with other computational hit discovery approaches, while compressing the hit-to-lead refinement and optimization process by pre-validating hits for translatability. This unique integration dramatically reduces the number of compounds to evaluate in subsequent experimental research.
The ability to generate target engaging compound hits depends largely on the size of the molecular search library. The AtlasGEN search space is supported by more than 1 trillion compounds. Generative AI and deep learning techniques deliver unprecedented discovery capability.
AtlasGEN accurately predicts protein-lygand binding affinity by combining geometric conformer analysis with machine learning based on experimental data 72% more efficiently than molecular docking-based approaches.
.png)
