A Generative AI appears...Reduce the time and cost of developing new drugs by about half

Developing a new drug is a battle of time and cost. It takes about 10 to 12 years on average to develop a new drug based on one of the world's top 12 pharmaceutical companies, and it costs about $2.168 billion in research and development. Nevertheless, there is a low chance that a new drug will succeed.

Among them, it is predicted that the emergence of Generative AI will shorten the development period of new drugs by nearly half. The reason is that by using big data and AI, the discovery period of candidate substances can be shortened to 1-2 years, and approval can be shortened to 5-7 years from clinical trials 1 to approval.

According to related industries, attempts to use big data and AI in clinical development and approval as well as search for new drug candidates are continuing.

Morgan Stanley Research also predicted that the application of AI and machine learning could lead to 50 new treatments within 10 years.

In addition, according to market research firm Market & Market, the global AI-based new drug development market formed $413.2 million in 2021 and is expected to grow 46% annually by 2027.

AI utilization solutions currently used in the clinical trial process are being used in clinical trial planning and institutional selection, patient recruitment, and operation stages.

Among them, the key area of interest of researchers is the patient recruitment stage. This is because about 80% of clinical trials do not meet the patient recruitment schedule, and a third of phase 3 clinical trials are having difficulty recruiting patients.

Using AI can accelerate patient registration by solving unexpected problems in institutional selection and clinical operations.

AI is also used to model patient selection/exclusion criteria to select subgroups that are likely to respond to treatment, and to predict/prevent death or adverse reactions.

According to the ASCO 2022 published case, MediData also developed a model to predict cytokine release syndrome, one of the critical adverse reactions, with up to 90% accuracy, after creating a machine learning algorithm by applying past clinical trial datasets in chimeric antigen receptor T cell therapy (CAR-T).

Synthetic controls (Synthetic Control Arm) can also be used in the patient recruitment stage.

Synthetic controls are solutions that provide reliable external controls by matching historical clinical data with statistics in clinical trials where standard treatments are inadequate or lack control information due to patient scarcity problems such as life-threatening cancer and rare diseases.

The creation of a hypothetical control group has the advantage of reducing the total number of patients required to enroll in a clinical trial. Accordingly, recently, global regulatory agencies are also prospectively reviewing this type of clinical trial.

In fact, the U.S. Food and Drug Administration (FDA) has approved about 20 of all clinical trials to be conducted through external (synthetic) controls.

In October 2020, the FDA approved Medicena Therapeutics' phase 3 clinical trial for recurrent glioblastoma, an industry source said. "This is the first time a hybrid external control group has been approved rather than a conventional random control method. If phase 3 clinical trials are successfully completed, it could be the first new drug to be approved as a synthetic control group."