By Demetris Iacovides, Principal, Big Pi Ventures
Our portfolio company Intelligencia recently closed a Series A round of $12M and a strategic partnership agreement with a major solutions provider to the pharma industry. In this post we discuss why the tools it has developed are set to accelerate and improve drug discovery.
The pharmaceutical industry has traditionally been suffering from low R&D productivity: the R&D investment needed for getting a new molecular entity (NME) from Phase I to FDA approval surpasses $1.8B on a risk-adjusted basis. This is due to the extremely low approval rates, long development timeframes, and huge costs involved in clinical development.
How biotech has transformed drug discovery
Prior to the emergence of biotechnology, drugs were all chemical-based with no connection between the molecule itself and the disease target. Thousands of chemical compounds were being screened through a trial-and-error process to find a potential drug target. This relatively random process was risky thus contributing to the low success rates in clinical development.
Biotechnology has completely transformed the drug discovery process. A biological connection between a disease and a potential drug can now be identified using modern techniques such as gene sequencing. Over the past two decades, big pharma has found its way into biotech through numerous acquisitions and licensing deals with smaller and more innovative biotech companies. That seemed the only way for big pharma to improve R&D productivity, boost their pipeline and tap into new diseases and therapeutic areas.
Some argue that big pharma missed the boat in the 1980s/1990s and that they were late in the game. Regardless, by looking at big pharmas’ pipeline now it is obvious that their primary focus is the development of large molecules.
Biotech has admittedly delivered its promise to bring effective therapies for patients against many diseases, especially so-called “orphan” ones. It has also slightly boosted R&D productivity by rationalizing drug discovery process and consequently improving success rates in clinical development. The 8–10-year development timeframe, however, still holds for biologics.
The emergence of new, disruptive technologies such as AI, big data, and data analytics create hopes that drug discovery and development can be improved and accelerated. But is it too early?
Applying AI to drug discovery
As the drug discovery process has been improved through biotech, pharma is looking to tackle the next big challenge: improving R&D productivity by reducing attrition rates and shortening development timelines. To achieve these goals the industry is betting on AI.
Although advances in AI have enabled computational biologists to uncover new ways to discover and develop drugs, uncertainties related to the process remain. The extent to which AI can improve R&D productivity through identification of new molecules is speculative. Although there has been an increase in deal activity between large pharmaceutical companies and AI startups in the past few years, we are yet to observe a meaningful impact on R&D productivity.
Only increased adoption of AI in drug discovery on the long run would reveal the concrete benefit of AI on R&D productivity through faster identification of safer and more efficacious drugs.
Utilizing AI in clinical development can have immediate benefits
According to a study from Nature the average cost per approved New Molecular Entity (“NME” i.e. a novel drug that has not been previously approved by the FDA) is $1.8B (adjusted for the cost of capital over a period of 10+ years). A major portion of these costs is attributed to the low probability of success (“PoS”) of drugs in clinical trials. Based on the sensitivity analysis performed in the study, PoS in Phase II is the most critical parameter of the total cost per approved NME. Indicatively, increasing the Phase II PoS from 34% to 50% would reduce cost per approved NME by 12%, or approximately $200M.
Therefore, if AI could help gauge the PoS of clinical drug candidates and help companies prioritize those that are more likely to succeed, potential savings are significant.
Assume that a pharmaceutical company has a portfolio of 5 Phase II and 2 Phase I products across different diseases. Which of these products have the highest probability of approval and should be prioritized? AI can be used to predict the success rate of these drugs based on historical clinical trial results, molecular specificities, academic literature, and other data sets.
One of the leading companies that applies AI to de-risk drug development is Intelligencia (www.intelligencia.ai). Intelligencia’s platform focuses on estimating the risk of clinical trials and interpreting the multitude of factors that contribute to that risk. It aims at bridging the gap between innovation and risk reduction, with the ultimate goal of bringing novel therapies to patients faster.
The round will also include an exclusive partnership with ZS Associates, where ZS will utilize Intelligencia’s AI expertise to optimize R&D portfolio strategy of its pharma and biotech clientele.
Intelligencia’s technology is expected to transform the industry by prioritizing pharma’s R&D portfolio and determining the most promising drug candidates. It will reduce the investment needed to bring a novel drug in the market as well as improving patients’ lives through safe and effective pharmacological treatments.