Skip to main content

The Rise of AI in Pharmaceutical R&D

I'm an accredited journalist working at the intersections of science, food and public health. I am also a certified nutritionist.

Research Triangle Park in Durham, North Carolina, is home to the world's top CROs.

Research Triangle Park in Durham, North Carolina, is home to the world's top CROs.

Using AI to Discover Novel Drugs

Robust opportunities afforded by the medical applications of artificial intelligence (AI) abound. The use of AI for drug discovery by biotechnology companies has grown thirtyfold since 2017.

AI applications are far and wide, ranging from the discovery of more reliable targets and better trial design and planning to more efficient execution from recruitment through monitoring.

This wave of innovation accelerated with the convergence of technological trends and a significant increase in the number of co-developed assets, rising from 32% in 2020 to 46% in 2021, according to Deloitte’s 12th annual Pharmaceutical Innovation report.

Almost half of revenues from late-stage trials are being generated through partnerships between AI-biotech start-up companies, pharmaceutical manufacturers and contract research organizations (CROs). Tech companies are also entering the healthcare space with their AI-ML offerings.

AI Advances for Biotech and CRO Companies

AI technology itself has matured in the last decade. So has computational resources, graphics processing units (GPUs), big data infrastructures and big data availability—with the refinement of data providers, lab automation and robotization to gather high throughput data.

Proof-of-concept cases have been flourishing in the last five years, especially in trials of pre-clinical candidates using deep learning-based software or advanced analytical systems. The results suggest that it takes 18 months to discover clinical drug candidates at a fraction of old drug discovery costs.

The benefits of AI platforms in pharmaceutical research and CROs include:

  1. Increasing quality management and data capture
  2. Helping to discover and prioritize novel targets in pre-clinical research
  3. helping to generate novel molecules and make drug repurposing predictions
  4. Helping to design and predict clinical trials
AI is revolutionizing biotech and pharmaceutical companies.

AI is revolutionizing biotech and pharmaceutical companies.

Unlocking AI's Untapped Potential

The design aspect is especially important to allow faster site identification and faster recruitment rates, which has an enormous cost-saving implication for sponsors. Clinical trial site management costs sponsors and CROs millions of dollars each day clinical trials run.

AI R&D platforms will be applied to many different use cases on a global scale as data access widens and inter-operability increases.

The AI frameworks themselves can be customized within specific business processes and decision points to create value for emerging biopharma companies (EBPs) and pharmaceutical manufacturers who can then layer their specific domain expertise.

The biggest AI area of opportunity is in leveraging extensive patient historic records or real-world data to get a better sense as to which specific patient populations to target for a protocol and their likely positive response or the potential number of adverse events.

AI-Augmented Data Auditing

Data validation in bioresearch monitoring is a cornerstone of clinical research safety and quality assurance. Regulatory oversight issues and poor study validation are two major focus areas of pharmaceutical firms' FDA inspections against The International Conference on Harmonisation GCP (ICH GCP) guideline.

AI can augment source data reviews to reduce human error in matching clinical trial documentation, like the case report form (CRF), patient information, toxicity data, etc. So the technology can help a clinical trial monitor or clinical research associate inspect research progress reports.

Some industry professionals, like former FDA bioresearch monitoring investigator, Patrick Stone, believe that while this application of AI is promising, the limiting factor will be its programming.

"There's a rigid logic-based aspect to AI, whereas a human can extrapolate extraneous variables. So I would hope we use AI to augment data integrity and add value to the struggle of connecting A to B. But I'm not sure we're at a point where AI can extrapolate some of the variables we need to review."

The value-added services of risk-based monitoring that experienced auditors like Stone provide will be hard for AI to replicate. Having the first pass through AI and then bringing in senior auditors to connect the dots is an attractive proposition and probably a good starting point.

This AI capability is being trialed in various research areas in certain parts of the world, like Japan and Europe, that have a robust AI developing rapidly. The market of AI-enabled trials is expected to consolidate further in the near future as the appetite for AI and venture capital funding keeps increasing.

This content is accurate and true to the best of the author’s knowledge and is not meant to substitute for formal and individualized advice from a qualified professional.

© 2022 Camille Bienvenu