Targeting special patient populations with the help of AI

By Melissa Fassbender

- Last updated on GMT

(Image: Getty/metamorworks)
(Image: Getty/metamorworks)

Related tags Artificial intelligence Technology Clinical trial data Clinical trial Data management

The industry is increasingly engaging in collaborations to gain access to technology, such as AI, which is being touted as a potential solution to some of the most challenging aspects of running clinical trials.

As opposed to data structured in electronic medical records, the majority of patient data is often unstructured, located across physician notes, pathology, genomic and medical imaging reports – unstructured data which traditional software cannot handle, said Stephen Gately, president and CEO, TD2.

As such, this data set is typically unsearchable and of no value to researchers looking to identify potential clinical trial participants. As Gately said, “The best-designed clinical trial goes nowhere if we fail to find the right patients.”

However, contract research organization (CRO), such as TD2, are increasingly collaborating with companies to gain access to new technology, which, until recently, has often been outside their purview.

Such “new” technology – while not new to other industries – is making inroads in drug development, enabling companies to identify specific patient populations for participation in clinical trials.

In fact, Mayo Clinic recently reported that IBM’s Watson for Clinical Trial Matching has shown an 80% increase in clinical trial enrollment​.

“The advent of targeted new oncology medicines has demanded the ability to search for patients with highly specific mutational contexts of drug sensitivity in real-time that are eligible to participate in clinical trials,”​ explained Gately.

To meet this demand, TD2 recently partnered with Deep 6 AI, which uses a variety of artificial intelligence (AI) and natural language processing (NLP) techniques to turn structured and unstructured clinical patient data into unified “patient graphs.”

“These patient graphs can be updated in real-time to display every clinical event in a patient’s life, and how these events relate to other events,”​ said Wout Brusselaers, Deep 6 AI cofounder and CEO.

“This makes it really easy to match patients’ graphs against clinical trial eligibility criteria, and find patients for studies, in minutes rather than the months it takes today,”​ he told us.

Brusselaers developed the platform with co-founder and chief scientist Brian Dolan and the rest of the Deep 6 AI team while working for the US Intelligence Community.

The platform allows comparison of a patient’s graph with other, similar patients to perform deeper analyses, such as “‘find patients like this’ or find patients who ‘look a lot like these early Alzheimers patients, but who have not been diagnosed with Alzheimer's yet,’” ​Brusselaers explained.

For TD2, Gately said the technology allows the CRO to that ensure genomically or incidence rare cancer patients are not excluded from new clinical trials based on inappropriate protocol criteria.

Researchers also can identify “exceptional responders”​ to treatment, Brusselaers explained – finding more, similar patients and fewer who might be more likely to encounter adverse events.

However, as Gately noted, “AI is not magic, but it’s a great accelerator, if harnessed and applied correctly.”

Workflows must allow for human validation and training of the software in order to resolve any data inconsistencies or conflicts, he explained, adding that AI cannot make up for lacking or incorrect data, but can prompt users to do so. 

“The ability to rely on real-time, real-world evidence to guide study decisions, interpretation and recruitment, unencumbered by HIPAA issues, is a game-changer,” ​Gately said, calling AI “a distinct competitive advantage.”

“CROs that cannot adapt and adopt new technologies, such as AI, and work with nimble, creative companies, such as Deep 6 AI, to guide the future of their industry, will be left behind,” ​he said.

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...

Increasing the Bioavailability of Oncology Drugs

Increasing the Bioavailability of Oncology Drugs

Content provided by Lonza Small Molecules | 13-Nov-2023 | White Paper

Oral tyrosine kinase inhibitors (TKIs) are a class of cancer drugs that can be highly susceptible to issues with solubility in the gastrointestinal tract

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