Accelerating Drug Discovery With AI for More Effective Treatments

Artificial intelligence can revolutionize drug discovery by expediting development, reducing costs, and improving treatment options, but addressing its limitations is crucial for future success. | Image Credit: AddMeshCube - stock.adobe.com

The implementation of artificial intelligence (AI) throughout the drug discovery process can assist with faster drug developments, cost savings reductions, and more effective treatment options for patients but further research must consider the knowledge gaps of AI to improve them for future use.¹

In 1950, Alan Turing used computers to simulate intelligent behavior and critical thinking, marking the earliest use of AI in health care.² By the early 2000s, AI allowed health care workers to screen patients with great accuracy for diseases ranging from diabetic retinopathy to skin cancer.³ Over several decades, AI research has evolved within the health care landscape.² Currently, the FDA has more than 900 AI and machine learning-enabled medical devices approved for use.³

Pros and Cons

The financial implications of AI integration within health care could be significant due to faster development times, improved success rates in clinical trials, and more efficient use of resources throughout the research and development (R&D) process.⁴ For every 1 drug that is a success, there are 9 failed drug candidates.⁵ The integration of AI would be beneficial to help manufacture medicines quicker and reduce the number of failed candidates.

The accelerated rate at which AI operates allows machine learning algorithms to quickly analyze large amounts of biomedical data.⁴ This will result in faster clinical trials, along with automation and error reductions. The outcomes of AI integration within the health care sector are intertwined because cost savings on drug development will lead to cheaper prices for consumers, increasing overall accessibility.

Despite the undoubtable benefits of AI integration within the medical field, the tool comes with flaws as well. One of the most popular concerns regarding AI surrounds the lack of diverse training that leads to poor model performances and shortcut learning.¹ Additionally, AI systems tend to calculate their predictions very opaquely which conflicts with the regulatory approval frameworks required for rationale decision-making.

AI tools neglect high-quality data inclusion during the subsequent stages of drug discovery and outcomes are uncertain based on regulatory expectations that concern late-stage AI applications.¹ The potential for AI exists but the performance is only as useful as the data it was trained on.⁵ Data must be in the proper form to be considered useful for generating leads on potential drug candidates using AI.

The FDA and other watchdog agencies are struggling to decide how to evaluate AI-assisted drug development processes.⁴ These systems are learning how to ensure they will be able to meet the same rigorous safety and efficacy standards compared with traditional methods.

Paving the Way for AI-Driven Drug Discovery

The average total expense of designing a drug is $1 billion over an estimated 10 to 15 years.⁶ The future goal of AI integration within drug development is to utilize technology to improve the overall odds of drug candidates while reducing costs and accelerating production.

Market predictions are expected to increase over 1000% between 2022 to 2029 from $13.8 billion to $164.1 billion.¹ Researchers have presented substantial promises of AI within drug development, proving benefits from advances in the use of AI can predict protein folding, molecular interactions, and cellular disease processes.

Johnson & Johnson has integrated AI into its drug discovery and development processes as well.³ The company employs AI to accelerate the identification of new drug targets, optimize molecule discovery, and streamline patient recruitment. These AI-driven strategies have contributed to more efficient and personalized patient care.

The AbbVie R&D Convergence Hub (ARCH) is an AI-powered drug development platform that integrates data from diverse sources to accelerate research.⁷ By analyzing data and predicting outcomes, ARCH aids in identifying new drug targets. Additionally, AbbVie employs large language models to design drugs computationally, and precision medicine tools help identify patient specific biomarkers for targeted treatments.

Other companies integrating AI include Eli Lilly and Insitro’s new paradigm of collaborations between pharmaceutical companies and smaller biotechs to produce metabolic drugs.⁸ The agreements of the 2 companies will allow further support of research that could result in human trials for the first set of programs.

Recently, Pfizer partnered with the Ignition AI Accelerator to enhance AI expansion across the health care sector.⁹ Pfizer expects the collaboration to lead to faster, more effective forms of communication with stakeholders to provide a more efficient patient drafting system while improving the manufacturing process.

AI systems are being leveraged to incorporate drug discovery, data extraction, and 3-dimensional computerized tomography image organization, according to the technology company Nvidia.¹⁰ The company utilizes a Nvidia Inference Microservice (NIM) cloud-native microservice designed to decrease time-to-market and simplify deployment of AI. Nvidia applies 3 NIM microservices to help researchers optimize libraries of small molecules to seek promising candidates that can bind to a target protein.

The integration of AI into the drug discovery process offers immense potential for accelerating drug development, reducing costs, and improving patient outcomes. However, the successful implementation of AI requires addressing knowledge gaps, ensuring data quality, and navigating regulatory challenges. By overcoming these hurdles, the health care industry can harness the power of AI to revolutionize drug discovery and bring innovative treatments to patients more efficiently.

References

1. Druedahl LC, Price WN, Minssen T, Sarpatwari A. Use of artificial intelligence in drug development. JAMA Netw Open. 2024;7(5):e2414139. doi:10.1001/jamanetworkopen.2024.14139

2. Kaul V, Enslin S, Gross SA. History of artificial intelligence in medicine. Gastrointest Endosc. 2020;92(4):807-812. doi:10.1016/j.gie.2020.06.040

3. Welch A. Artificial intelligence is helping revolutionize healthcare as we know it. Content Lab U.S. September 14, 2023. Accessed October 15, 2024. https://www.jnj.com/innovation/artificial-intelligence-in-healthcare

4. AI in pharma shows promise with faster development, lower costs. PYMNTS.com. October 11, 2024. Accessed October 15, 2024. https://www.pymnts.com/artificial-intelligence-2/2024/ai-in-pharma-shows-promise-with-faster-development-lower-costs/

5. AI powers a new era of drug discovery and development. Nature.com. 2024. Accessed October 15, 2024. https://www.nature.com/articles/d42473-024-00250-9

6. Lohr S, Lowell S. How A.I. is revolutionizing drug development. The New York Times. June 17, 2024. Accessed October 16, 2024 https://www.nytimes.com/2024/06/17/business/ai-drugs-development-terray.html?ogrp=dpl&unlocked_article_code=1.SE4.B97k.YzgBs26ftSka&smid=url-share

7. Three ways AI is changing drug discovery at AbbVie. AbbVie. September 25, 2024. Accessed October 16, 2024. https://www.abbvie.com/who-we-are/our-stories/three-ways-ai-is-changing-drug-discovery-at-abbvie.html

8. Jensen K. Lilly partners with AI specialist Insitro to develop metabolic medicines. BioPharma Dive. October 9, 2024. Accessed October 16, 2024. https://www.biopharmadive.com/news/lilly-insitro-metabolic-drug-discovery-ai/729347/

9. Salmon K. Pfizer partners with Ignition AI to enhance drug discovery. IT Brief Asia. October 14, 2024. Accessed October 16, 2024. https://itbrief.asia/story/pfizer-partners-with-ignition-ai-to-enhance-drug-discovery

10. Vella H. AI deployed in health care for drug discovery, data and imaging. Aibusiness.com. October 14, 2024. Accessed October 16, 2024. https://aibusiness.com/generative-ai/ai-deployed-in-health-care-for-drug-discovery-data-and-imaging#close-modal