The Use of Artificial Intelligence in Drug Discovery and Development

By: Aastha Vaidhya and Keiran Pace

Demonstrated by the mobilization of pharmaceutical companies and researchers in the recent pandemic, getting safe and reliable drugs and vaccines to the market faster is more important than ever. This requires the use of cutting edge technologies such as artificial intelligence (AI) and machine learning (ML) that have the potential to allow the pharmaceutical industry to keep up with global disease patterns. AI is defined as the science and engineering of creating intelligent machines and computer programs, of which machine learning is a subset striving to build analytical models by building from existing data.¹,² An increase in high quality data from genomics, imaging, and digital wearable device use has allowed rapid advancements in AI methods to expedite processes, formulating a pattern of reliance on digital technology in the pharmaceutical industry.³ Accordingly, the number of patent applications in the industry has risen from 66 in the three months ending March to 77 in 2022, a 16.6% increase.4 Johnson & Johnson has recently expanded research and development (R&D) in machine learning and is the top innovator in the pharmaceutical sector.⁴ Hence, the ability of AI to promote innovation and enhance productivity gives it an important future in the pharmaceutical science field.

From early-stage studies to drug synthesis and distribution, drug discovery requires heavy expenditure and time consumption with a low success rate. This is illustrated by the average $1.3 billion R&D investment per drug and 5.9 to 7.2 median development time with only a proportion of 13.8% amounting to an approval.³ However, the capability for machine learning to accumulate and analyze mass amounts of data can allow scientists to work efficiently to simplify this process, reducing cost and increasing the probability of success (POS).3 ML algorithms like natural language processing (NLP) can extract scientific insights from literature to help identify novel targets, after which predictive modeling can help predict protein structure and promote molecular compound design.³ Recent developments include applications in clinical trial research. An estimated 80% of trials do not meet enrollment timelines, leading to premature termination of phase 3 trials in 30% of trials.³ AI methods can be used to enhance patient enrollment by collecting patient data digitally and decentralize trials to reduce the need for in person visits to site. Instead, real-time automated monitoring of clinical trial sites that can mitigate this labor-intensive, time-consuming, and costly control step.³ Transitioning from traditional site monitoring to a risk-based monitoring model that predicts risk with thresholds for signal detection is viable with AI. These methods can conduct “smart monitoring”—accumulating trial data and continuous learning from it to improve data quality checks.³

Personalized medicine, or precision medicine combines an individual’s genetic profile, symptoms, medical history and integrates this data with predictive analytics to adopt a personalized diagnostic and therapeutic procedure that ensures effective treatment.⁵ Accumulating data from various sources and analyzing them to identify groups based on genetic composition is effective for drug candidate selection.⁶ Precision medicine with AI can be deployed in clinical development to match patients with the right therapy to maximize patient benefit. Among 54% of clinical development failures of which 57% are due to inadequate efficacy, a major contributor involves a failure to identify the appropriate target patient population with the right dose regimen.³ Systematic models that integrate ML can be used to build a probabilistic model to forecast the POS and pinpoint patient subgroups with a higher probability of therapeutic advantage.³

Providing the capability to monitor and conduct predictive forecasting can help plan pharmaceutical supply chains to prepare inventory at appropriate time and quantity. Demand knowledge is the basis of designing future markets to aid in the robust development of supply chains. With AI, changes in demand can be predicted several months ahead of time within 95% accuracy to ensure manufacturers and suppliers manage pharmaceutical production and distribution accordingly, or develop new products.⁷ This can be achieved through deep learning techniques that use a multi-layered architecture to depict the relationships between inputs and outputs.⁷ In other words, teaching computers to learn by example; performing classification tasks from images, texts, and sound.⁷

Figure 1: Classification of AI³

This technology has the potential to predict new treatments and drugs to limit the spread of contagious diseases and reduce the societal burden of other diseases. Further, predictive forecasting of epidemics can help estimate disease proliferation and can guide pharmaceutical public health measures.⁸ The response to the COVID-19 pandemic has highlighted a need for digital AI-based public health solutions. Using AI text and data mining to analyze large volumes of data have provided numerous applications for responding to this crisis, as illustrated in Figure 2 below.⁹

Figure 2: Applications of AI in COVID-19 response⁹

Leveraging intelligent AI solutions to speed up discovery and drug development, increase their success, and recognize the need for new therapies is essential in the face of global trends such as population growth, urbanization, and climate change. However, the use of artificial intelligence in the pharmaceutical industry is still in its infancy and will likely take years before becoming fully integrated into R&D processes.¹⁰Recent research suggests that 50% of global healthcare companies plan to incorporate AI strategies by 2025, especially in the investment of new drugs for chronic and oncology diseases.¹¹ These tools will become more accessible over the years, and eventually a natural process in pharmaceutical science.

Reference List:

1. Angehrn Z, Haldna L, Zandvliet AS, et al. Artificial Intelligence and Machine Learning Applied at the Point of Care. Frontiers in Pharmacology. 2020;11. doi:10.3389/fphar.2020.00759

2. Peng J, Jury EC, Dönnes P, Ciurtin C. Machine Learning Techniques for Personalised Medicine Approaches in Immune-Mediated Chronic Inflammatory Diseases: Applications and Challenges. Frontiers in Pharmacology. 2021;12. doi:10.3389/fphar.2021.720694

3. ‌Kolluri, Sheela, et al. “Machine Learning and Artificial Intelligence in Pharmaceutical Research and Development: A Review.” The AAPS Journal, vol. 24, no. 1, Jan. 2022, www.ncbi.nlm.nih.gov/pmc/articles/PMC8726514/, 10.1208/s12248-021-00644-3. Accessed 25 Oct. 2022.

4. Data Journalism Team. “Pharmaceutical Industry Companies Are Increasingly Innovating in Machine Learning.” Pharmaceutical Technology, Pharmaceutical Technology, 16 May 2022, www.pharmaceutical-technology.com/analysis/pharmaceutical-industry-companies-are-increasingly-innovating-in-machine-learning/. Accessed 25 Oct. 2022.

5. Bresnick, Jennifer. “What Are Precision Medicine and Personalized Medicine?” HealthITAnalytics, HealthITAnalytics, 11 Jan. 2018, healthitanalytics.com/features/what-are-precision-medicine-and-personalized-medicine. Accessed 25 Oct. 2022.

6. “Machine Learning: The next Generation Manufacturing for Pharmaceutical Industry.” Pharmafocusasia.com, 2022, www.pharmafocusasia.com/articles/machine-learning-the-next-generation-manufacturing. Accessed 25 Oct. 2022.

7. Mohaideen S, Keerthana V, Vigneshwaran, L, Senthil Kumar M. View of Artificial Intelligence in Predictive Forecasting for Healthcare. International Journal of Recent Advances in Multidisciplinary Topics. 2022;3(3). https://journals.resaim.com/ijramt/article/view/1864/1804

8. Zeng, Daniel, et al. “Artificial Intelligence–Enabled Public Health Surveillance—from Local Detection to Global Epidemic Monitoring and Control.” Artificial Intelligence in Medicine, 2021, pp. 437–453, www.ncbi.nlm.nih.gov/pmc/articles/PMC7484813/, 10.1016/b978-0-12-821259-2.00022-3. Accessed 25 Oct. 2022.

9. “Using Artificial Intelligence to Help Combat COVID-19.” OECD Policy Responses to Coronavirus (COVID-19), 23 Apr. 2020, www.oecd.org/coronavirus/policy-responses/using-artificial-intelligence-to-help-combat-covid-19-ae4c5c21/, 10.1787/ae4c5c21-en. Accessed 25 Oct. 2022.

10. GlobalData Healthcare. “It Will Take Years for AI Use to Peak in Drug Discovery and Development Process.” Pharmaceutical Technology, Pharmaceutical Technology, 13 June 2022, www.pharmaceutical-technology.com/comment/ai-peak-drug-discovery-development/. Accessed 25 Oct. 2022.

11. PharmaNewsIntelligence. “AI in the Pharma Industry: Current Uses, Best Cases, Digital Future.” PharmaNewsIntelligence, 30 Apr. 2021, pharmanewsintel.com/news/ai-in-the-pharma-industry-current-uses-best-cases-digital-future. Accessed 25 Oct. 2022.