Artificial Intelligence and the Pharmaceutical Industry

Reproduced from Drug Discovery Today, Volume 25, Number 9, September 2020

People worldwide are embracing this technology in the form of smartwatches, smartphones, smart home systems, and self-driving cars. Even the ads for products and services you see online are AI-generated, in that they are selected based on your browsing history. Several innovations are still in their infancy, and others have advanced to become very reliable tools. According to an article published in Statista in 2019, the number of AI-related patents increased from about 23,000 to over 78,000 in the 10-year span from 2008 to 2018.

‍

The pharmaceutical Industry is also beginning to embrace this new wave of technology.

‍

According to a recent review in Drug Discovery Today  leading pharmaceutical companies are in an ‘early mature’ phase of using AI in research & development (R&D). The authors studied 21 major pharmaceutical companies and found that between 2014 and 2019 there were 417 AI-related publications, of which 398 were for R&D projects. The figure below summarizes the AI-related activities for the major pharmaceutical companies in the period from 2014 to 2018, categorized in terms of internal research projects, co-operations, alliances/consortia, acquisitions and investments, and joint ventures.  

Challenges for implementing AI in the pharmaceutical industry are multifaceted, but the major hurdles are the dynamic global regulatory environment, existing inflexible processes, incomplete and/or non-homogenous data records that result in misinterpretations, the lack of specialized talent, and the high costs currently associated with such initiatives.

‍

In spite of the challenges, the pharmaceutical industry is still making advances.

‍

Two examples of FDA‑approved products that utilize AI technology are (1) the 3D-printed levetiracetam tablets SPRITAM® (treatment for epilepsy) enabling high-dose medication in a rapidly disintegrating form, which is a huge step towards developing personalized drug therapies, and (2) the first digital tablet, Abilify MyCite® (treatment for mental health disorders), which has an embedded ingestible sensor that records if the patient took the medication. These are significant steps towards embracing this exciting new technology.

In drug discovery, AI is being used to improve the speed and focus of the drug discovery process. Examples include designing small molecules, predicting structure-activity relationships, performing 3D protein structure simulation, finding biomarkers, and predicting blood-brain barrier permeability.

AI can also streamline the clinical trial management process by offering the analytical power, flexibility, and speed required for drug development. In a recent perspective from Deloitte Insights, the authors discuss intelligent clinical trials of the future and suggest that AI-enabled technology can aid clinical trial design by collecting, organizing, and analyzing large sets of data to extract meaningful patterns; improve patient selection; increase clinical trial effectiveness through mining, analysis, and interpretation of multiple data sources; help identify and locate qualified investigators; and help monitor and manage patients by automating data capture, digitalizing standard clinical assessments, and sharing data across systems.

One noteworthy example of a company that is using AI for clinical trials is Deep 6 AI. This US-based company matches patients for clinical trials in minutes, rather than months, with software that analyses clinical data using machine learning. Their product can work with both structured and unstructured data to find more, better-matching patients for clinical trials at lower costs and in less time. Given that subject recruitment and screening is typically a rate-limiting hurdle, this may present a significant advancement for clinical trials.

One more area where AI can be very beneficial to the pharmaceutical industry is pharmacovigilance (PV), which is the science involving the collection, detection, assessment, monitoring, and prevention of adverse drug reactions. For example, a partnership between Genpact and Bayer is accelerating patient safety data monitoring using AI. This program is enabling potential drug-related adverse effects to be detected and reported to regulatory agencies with improved speed and efficiency.

To conclude, the pharmaceutical industry, a highly regulated industry, is moving towards adopting this new technology with the support of regulatory bodies as the technology itself evolves. We will have to wait and see what the AI revolution is going to bring, but not for too long – I hope!

‍