The concept of artificial intelligence (AI) in drug discovery and manufacturing has been in progress for more than 10 years; however, with the current advancements in technology, medicine, and research, AI is closer than ever to becoming a realistic, treatment planning option.
In February 2023, MIT Technology Review published an article on the recent work and success of Exscientia, the first company in the intersect between pharma and AI. The UK-based company had been working with the Medical University of Vienna in Austria to develop new “matchmaking technology” that “pairs individual patients with the precise drugs they need, taking into account the subtle biological differences between people.”
The article focused on an 82-year-old male, identified as “Paul,” with an aggressive form of blood cancer. Paul underwent six long courses of chemotherapy that were unsuccessful at treating his cancer, did not change his prognosis, and left Paul experiencing the adverse effects of treatment. After several months of trying different chemotherapeutic agents and while feeling there was nothing to lose, Paul was enrolled in a collaboration trial between Exscientia and the Medical University in Vienna, Austria where Paul lived. A small sample of tissue was retrieved from Paul that contained normal cells, as well as cancer cells. The sample was divided into more than one hundred pieces and, through AI, were exposed to different drug variations with different chemical properties.
AI was able to accurately trial different medications while seeing the real-time response of the healthy cells, as well as the cancer cells. With the ability to trial different medications while seeing the real-time effectiveness, it was discovered that Paul’s cancer responded effectively to a drug that had been unsuccessful in the treatment of similar patients in the past. AI allowed testing without adverse effects to Paul and was less time consuming versus traditional trial and error processes. Within two years, Paul was in complete remission and cancer-free.
Does AI Replace the Human Component of Drug Development?
AI presents an opportunity for collaboration between machine learning (ML) and individuals with significant training and experience. In 2018, Pfizer voiced its support of AI by stating, “[t]he software isn’t doing the thinking on its own…[h]uman professionals train it and monitor its results for legitimacy. The software then runs far, far, far faster and more accurately than an army of humans ever could.” In November 2022, Harvard Medical School published an article noting the urgent need for a more efficient and effective process for drug discovery, development, and implementation given “the world’s aging population, the growing burden of chronic and infectious diseases, and the emergence of novel pathogens.”
With the implementation of AI in drug discovery, there is the hope of better outcomes for patients in a more efficient, effective, and cost-conscious manner. AI also presents more investment opportunities to support pharmaceutical development and discovery as well as the unique occasion for intense collaboration between trained and specialized professionals and machines.
The Financial Aspects of AI in Drug Discovery and Manufacturing
The traditional drug discovery and development process consists of four steps that can span over 12 – 15 years and cost in excess of one billion dollars per drug, whether it is successful making it to market or not. In 2019, the pharmaceutical industry spent $83 billion dollars on research and development, which is about 10 times more than what the industry spent per year in the 1980s. Morgan Stanley, an investment company, remarked that “the median investment required to bring a new drug to market is estimated to be nearly $1 billion, while the true cost of research and development may be as high as $2.5 billion per marketed therapy, when factoring in abandoned trials and clinical failures.” AI presents an opportunity for researchers to capture and store reams of historical data, while introducing patient-specific medical history, radiological imaging, and pathological analysis. The AI platform can utilize all the data to create a recommended drug treatment in a more timely and efficient manner while increasing the chances of successful management of the disease. Morgan Stanley biotechnology analysts forecast that, with the use of AI, approved drug therapies may increase by as much as a 15% compared to 2021, leading to the potential creation of dozens of new medicines while generating a $50 billion market over the next decade meaning better outcomes for patients and more investment opportunities for investors.
Some medicolegal implications in the use of AI in the treatment of patients were addressed in one of our earlier articles. However, AI offers promise not only in the development of drugs but also in how they are made. How will the FDA approach the use of AI in drug manufacturing? Earlier this year, the FDA Center for Drug Evaluation and Research (CDER) published a discussion paper on the topic of Artificial Intelligence in Drug Manufacturing. CDER noted five areas of concern:
Cloud applications may affect oversight of pharmaceutical manufacturing data and records.
The IOT [internet of things] may increase the amount of data generated during pharmaceutical manufacturing, affecting existing data management practices.
Applicants may need clarity about whether and how the application of AI in pharmaceutical manufacturing is subject to regulatory oversight.
Applicants may need standards for developing and validating AI models used for process control and to support release testing.
Continuously learning AI systems that adapt to real-time data may challenge regulatory assessment and oversight.
In its invitation for public comments, CDER solicited feedback on topics, such as (1) how AI might be used in pharmaceutical manufacturing, (2) whether there are features of the current FDA regulatory framework that could impact the use of AI on drug manufacturing, and (3) whether and what kind of guidance regarding the use of AI would be useful to manufacturers. Thirty-six comments were received. All acknowledged the benefits of AI in drug manufacturing, acknowledging areas of use that include process development, quality assurance, equipment monitoring, document processing, inventory management, regulatory compliance, and the list goes on.
To further explore how the FDA may approach the regulation of AI in drug manufacturing, the FDA and the Product Quality Research Institute (PQRI) have scheduled a two-day workshop for September 26-27, 2023 - FDA/PQRI Workshop on the Regulatory Framework for the Utilization of Artificial Intelligence in Pharmaceutical Manufacturing. Both FDA, industry and academic officials are anticipated to speak at that workshop. One of the sessions will address comments received on the CDER discussion paper. This workshop will hopefully provide further insight into how the FDA will approach the regulation of AI and ML in drug manufacturing. Stay tuned for a future post summarizing the workshop and its highlights.