Approximately 50% of late-stage clinical trials fail due to ineffective drug targets, meaning only 15% of drugs advance from Phase II trials to approval. Artificial intelligence (AI) can help to enhance drug development by making more accurate predictions in novel areas of biology and chemistry.
AI can easily extract the relevant information from scientific literature and create connections between biomedical entities with relatively little information. Therapeutic areas, such as amyotrophic lateral sclerosis (ALS), with only a few treatments currently approved, would benefit greatly from the use of artificial intelligence.
The United States (US) Food and Drug Administration (FDA) is contemplating the wide variety of technologies including AI, that can help to curate data. AI-algorithms are software that can learn from and act on data. They are already being used to aid screening for diseases and recommendations for treatments. For example, the FDA has authorised an AI-based device for which can alert providers of a potential stroke in patients.
They recently released a regulatory framework for AI-based medical devices. The FDA have stated that the goal of the framework is to ensure the safety and effectiveness of AI software by using a validation process that guarantees improvement of performance.
Across the Atlantic, the United Kingdom’s (UK) Medicines and Healthcare Products Regulatory Agency (MHRA) has entered into a formal research agreement with a British AI technology company, in the hope of validating software algorithms used in digital health. They hope that this will accelerate medical research and drug development, whilst providing new methods of validation for AI.
Regulators worldwide are looking to harmonise efforts to enhance drug development through AI. The International Coalition of Medicines Regulatory Authorities (ICMRA) has released a report as part of a wider effort to decrease duplicative work and increase harmonisation among medicines regulators from the US, European Union (EU) and UK. To ensure that AI can be maximised to revolutionise drug development, drug regulators will need to ensure staff are trained on new product types, and have access to outside experts such as software engineers.
Exploiting digital technology and artificial intelligence in decision making can help to drive collaborative evidence generation, thereby improving the scientific quality of evaluations to enhance drug development.