How AI is shaping clinical research
Artificial intelligence (AI) is everywhere today, apparently being used in some capacity in almost every professional industry. And the life sciences are no exception.
In fact, it’s fair to say that AI has taken the biotech industry by storm, and the hype around it has caused many AI startups to emerge, using AI for everything from diagnostics to drug discovery and development. Even ChatGPT has been integrated into certain enterprise drug discovery processes.
As so many people in the industry have discovered, AI has many benefits, and making the most of the opportunities it provides can lead to faster success stories and important milestones in clinical research.
For example, a few months ago, AI helped scientists discover a new antibiotic called abaucin, which can kill Acinetobacter baumannii, superbug that causes fever, chills, and vomiting. It also occurs in multi-drug resistant ones.
In this case, the AI was able to rapidly screen out the 7,500 molecules it found to be inhibiting A. baumanniiand within an hour and a half narrowed down 250 potential compounds, after which abaucin was found to be the most potent.
Manually, this task would be very difficult. But, by using AI, the whole process is sped up, meaning that the antibiotics found can be used to treat A. baumannii much faster than if the process were done manually, which could end up saving people’s lives in the near future.
Key advantages of using AI in clinical research
Of course, as already mentioned, speed is one of the main advantages of using AI in clinical research.
This is very important, like, according to research from the Deloitte Center for Health Solutionsit often takes 10 to 12 years to bring a new drug to market, and the clinical trial stage itself averages five to seven years, thanks in large part to complications of manual effort, rework, and inefficiencies.
“The timeframe for developing a new drug from scratch is about 10 to 12 years, using $1.2 billion in a very conservative estimate. With AI, you can shorten it easily by more than half, to about 5 years. That’s what we see today. A lot of people in drug development will agree with that, especially for drug discovery to get the right drug and move on to the next phase,” said Gideon Ho, co-founder and chief executive officer (CEO) of HistoIndex, a company that uses AI augmented with stain-free pathology to view biopsy samples without staining tissue, before using a laser to scan and image biopsies, and using AI for readout data to assess the magnitude of the efficacy of treatments for fibrotic disease.
Besides speed, Ho also lists consistency, accuracy, repeatability, and scalability as the main advantages of using AI in the life sciences industry.
“This is very important because it can add to areas and has applications anywhere from drug discovery, testing of drugs in preclinical animal models, testing of safety and toxicity, assessment of efficacy during human clinical trials and even FDA approval of the drug itself,” he said.
Ho estimates that AI is used in about 50% to 60% of the entire drug development value chain. And, not only is it used for drug discovery, it can also be used in later stages, which involve testing the molecules found in pre-clinical animals or in vitro models, or in testing the safety or efficacy assessment of drugs in human trials.
Basically, AI can be used in almost the entire drug development process.
AI for drug design
AI is increasingly being used in preclinical and clinical trials, with advanced algorithms that can quickly analyze vast databases of chemical compounds and identify those compounds that are most likely to bind to targets, allowing drug developers to explore the world of chemistry much more quickly.
IKTOS is a company that leverages AI technology specifically for drug design, using it to quickly identify molecules – in this case, the company focuses on small molecules – that are suitable to become clinical candidates.
The Company’s CEO, Yann Gaston-Mathé, explains the drug design process: “So, basically, when you do small molecule drug discovery, you start with your target…we usually start by screening the known set of molecules on these biological assays, then identify what are called ‘hits’ – hits are the molecules that are active in your biologic assay – and then once you get a hit, you will start doing some chemistry, which means you will modify the hits to increase the level of activity or potency and also progressively optimize the molecules to make them. be selective on your targets.
Ultimately what you want are molecules that are active, selective, non-toxic, and novel.
And, again, the rationale behind wanting to use AI for this process is fundamentally about saving time, as it requires synthesizing and testing thousands of molecules. Done manually, it would require years of research, and the costs would be enormous.
However, by using its AI technology, IKTOS can carry out the drug design process much more quickly and efficiently, designing new molecules automatically.
The first drug produced entirely by AI enters clinical trials
In a recent success story for AI in clinical research, it was announced that the first drug completely generated by AI has entered human clinical trials.
The drug, INS018_055, is being developed by Hong Kong-based Insilico Medicine and has reached phase 2 trials for the treatment of idiopathic pulmonary fibrosis – a chronic disease that causes scarring in the lungs and makes breathing more difficult.
The company connects biology, chemistry, and clinical trial analysis using next-generation AI systems, having developed an AI platform that uses deep generative models, reinforcement learning, transformers, plus other modern machine learning techniques. They are used for discovery of new targets and fabrication of new molecular structures having desired properties.
And, while there are other AI-designed drugs in clinical trials, INS018_055 is the first drug with a new target that AI has discovered, as well as a new design that AI has come up with.
Insilico Medicine also has two more drugs in clinical phase that are produced in part by AI; one for COVID-19 in phase 1 trials, and another for solid tumors that recently received approval from the US Food and Drug Administration (FDA) to begin clinical trials.
The future of AI in clinical research
The use of AI in clinical research has really increased recently, and with technology seemingly advancing at a rapid pace, it’s likely that AI will take more responsibility in the drug development process to guarantee speed and efficiency.
“I would say that it will be an AI that inserts itself into different parts of the whole drug development process, as an aid into the whole diagnostic process, in decision making by doctors to make better decisions…,” commented Ho.
Moreover, the hype around AI is even expected to trigger a resurgence of investment into the biotech industry, after a difficult economic period where companies struggled to raise capital.
Overall, it looks like AI is the future of clinical research in the life sciences industry, and Ho believes that people within the industry should embrace it as a useful tool. “People who use AI will definitely replace people who don’t use AI. This is the natural evolution of things. If you have a new tool and you don’t use it, you’re essentially making yourself obsolete. If it’s a good tool, we should use it and embrace it so we’ll continue to be relevant to the industry.”
And, with the number of startups and AI technologies out there today, it’s fair to say that people are in the industry is really embrace it.
