How in silico trials are making personalized medicine a reality

By Melissa Fassbender

- Last updated on GMT

How in silico trials are making personalized medicine a reality

Related tags in silico Clinical trials Personalized medicine AI Artificial intelligence Real world data

The biopharma industry is on the precipice of making personalized medicine a reality with access to massive amounts of data, computing power, and artificial intelligence to run in silico clinical trials, says GNS Healthcare CEO.

Powered by artificial intelligence (AI), in silico​ trials enable researchers to adapt quickly in a computer environment, transforming large data sets to better understand drug response, earlier predict outcomes, and discover new drug indications.

The technology is driving industry efforts to transition from resource-intensive clinical trials, to faster, more flexible trials – and while in the early stages of adoption, the value of AI continues to expand.

To further discuss how AI is affecting the use of adaptive trials and practice of precision medicine, Outsourcing-Pharma (OSP) caught up with GNS Healthcare co-founder and CEO Colin Hill (CH).

OSP: In what ways are in silico ​trials the key to adaptive trial success?

CH:​ Drugs have traditionally been evaluated for safety and efficacy through randomized controlled trials. The challenge with these types of trials is that they do not take into consideration how a drug acts in the real world, outside a controlled setting.

In silico ​trials offer a way for biopharma companies to test their drugs using diverse data from electronic health records, DNA sequencing, -omics, and real-world sources to create interactive computer models and simulations of clinical trials that can provide valuable information regarding safety and efficacy of drugs both pre- and post-launch. 

Adaptive trials by definition must continually evolve and learn from new data, so in silico ​trials provide the flexibility researchers need to “adapt” quickly and easily within a computer environment.

Ultimately, the more information we can collect and leverage in the clinical trial process, the better we are able to identify the types of patients who will respond to a drug and the greater the chance that a proposed drug has of reaching patients who need it.

In silico ​trials by their very nature can ingest and transform large volumes of data at scale which is key to improving the probability of success in adaptive trials.

Tailoring drugs and treatments to individuals is not only the right thing to do for patients, but it has the potential to save trillions of dollars globally by squeezing the cost out of the whole system by eliminating wasteful treatments.

OSP: What role does machine learning play in this?

CH: ​AI and machine learning have a huge role to play in making the drug discovery, development, and commercialization process more efficient and successful. But it’s important to understand that the term machine learning covers a lot of types of AI, and not all AI is equal.

Artificial intelligence is a branch of computer science that teaches machines to imitate intelligent human behavior. Machine learning is a set of algorithms that help machines learn concepts and algorithms from data and there are a number of different types of machine learning that can be used to achieve AI.

For in silico ​trials you need to use an AI that can tease out answers to questions that you didn’t even know you had. It has to be unbiased, scientifically rigorous, and explore all the possible options for outcomes using data as the fuel.

Causal machine learning, a powerful form of AI, is so far the only type of machine learning that is producing scientifically valid results to discover drivers of disease and drug response, break out subpopulations of patients who respond positively to certain medications, and perhaps most importantly explain why those outcomes are occurring.

The big question that any clinical trial is looking to solve it does this medication work for a given individual. And if so, why does it work for this patient but not another one?

Not all AI is able to answer the “why does this work” question. Causal machine learning enables researchers to conduct a greater number and wider variety of in silico ​experiments with large clinical trial and real-world data sets to understand the drivers of drug response, predict outcomes earlier in the process, and discover new drug indications.

OSP: What are the benefits of in silico ​trials?

CH: ​Currently it takes about 10 years and more than $2bn dollars to develop a drug. And yet only a fraction of drugs are approved by the FDA. We can do better than this. Biopharma is starting to embrace new tools, like AI, which can speed the discovery of new therapeutics, increase the probability of success of clinical trials, and identify new indications for current drugs.

At the end of the day, biopharma needs to transition from very resource intensive clinical trials to trials that have more flexibility, are faster, and that better identify the drivers that produce positive outcomes in patients. In silico ​trials can achieve that by building models directly from data at scale.

Researchers and clinicians can then interact with the models to simulate future events like what is the best line of therapy, what happens if you switch medications or doses, and are there other indications for a current drug. 

In the past, we’ve had the vision but have been hamstrung by missing data or insufficient technology. We now have massive amounts of deep data, the computing power needed to process it, and the AI to run in silico ​trials and deliver insights that impact patient lives.

OSP: Where does the use of in silico ​trials stand from a regulatory viewpoint?

CH: ​Recently, we’ve seen a number of encouraging signs from the government that they support the innovations happening in healthcare. The 21st​ Century Cures Act really paved the way for regulatory agencies to take a much more progressive stand on topics like AI technology, data, and novel treatments.

The FDA has also issued a number of announcements and recommendations this year, offering its vote of confidence to new technologies technology – like in silico ​trials – in medicine and drug development. 

From a data point of view, the NIH launched the All of Us Research Program, with the goal of gathering data from more than one million diverse Americans to incorporate it into new studies that help us understand health and disease.

OSP: How has the industry’s use of this type of trial evolved?

CH: ​It’s certainly early days in the adoption of in silico ​trials. Until recently, we didn’t have the right ingredients to make these models possible or accurate.

But now, with the terabytes of data being created in health care every day, the availability and cost-effectiveness of computing power, and new proven AI technologies the time is upon us.

We think we’ll see continued evolution of these types of trials as organizations solidify their data strategy, ensuring the data they collect is deep and meaningful and we expand the ways we are sharing and integrating the data as an industry.

We are on the precipice of making precision medicine a reality and continue to see the value of AI in discovering and developing drugs as well as in how they are commercialized.

OSP: What do you expect from the next five, ten years?

CH: ​We think we’re going to see major advancements in a couple of different areas: adoption of AI, more cross-industry collaboration, and finally, more diseases cured.

In terms of AI adoption, we’re seeing the conversation move from questioning whether the technology is real, to what are the use cases it is best suited for. Drug development is certainly one of those areas.

But there are many others in which AI can add value to so many other parts of the drug lifecycle. For example, our work has helped explain the drivers of metastatic colorectal cancer progression and treatment, how multiple myeloma patients may respond to stem cell treatment, and what drives Parkinson’s Disease motor progression.

We’re also seeing tectonic shifts in how different health care stakeholders interact and collaborate. For so long we’ve been talking about how siloed health care, but with the need to integrate more data and the renewed focus on making care patient-centric, we are seeing new partnerships and mergers across the industry. We think we will continue to see this over the course of the next five to ten years and that collaboration will be a strategic consideration for every player in the healthcare ecosystem.

Finally, we believe we are on a path to cure cancer and other devastating diseases in the near future. We are continually improving our understanding of human biology and disease progression at a more rapid pace. We are leveraging those insights to develop novel treatments and interventions that are precisely targeted to an individual’s biology. We are truly on the precipice of making personalized medicine a reality.

Related news

Show more

Related products

show more

Using Define-XML to build more efficient studies

Using Define-XML to build more efficient studies

Content provided by Formedix | 14-Nov-2023 | White Paper

It is commonly thought that Define-XML is simply a dataset descriptor: a way to document what datasets look like, including the names and labels of datasets...

Overcoming rapid growth challenges with process liquid preparation

Overcoming rapid growth challenges with process liquid preparation

Content provided by Thermo Fisher Scientific - Process Liquid Preparation Services | 01-Nov-2023 | Case Study

A growing contract development manufacturing organization (CDMO) was challenged with the need to quickly expand their process liquid and buffer preparation...

Related suppliers

Follow us


View more