Software startup Deep 6 AI has developed a system that analyzes a range of patient health records to accelerate the process of identifying candidates for clinical trial recruitment—a longstanding bottleneck in the drug development process.
The Pasadena, CA-based company, founded in 2015, recently raised $17M in Series A funding to advance its software system, which aims to use artificial intelligence tools to resolve one of research’s biggest challenges: enrolling patients to participate in trials.
About 80 percent of clinical trials miss enrollment timelines, and nearly one-third of Phase 3 trials don’t take place due to a lack of patients, according to various reports.
While the move from paper to electronic medical records has improved the recruitment and retention process, most patient data is in physician notes and pathology, genomic, and medical imaging reports. Traditional software struggles to handle this type of “unstructured” data, which isn’t formatted in a way that computers can easily understand.
That’s where the software tools Deep 6 AI has developed can make a difference, the company says.
Deep 6 AI uses advanced algorithms to analyze the qualification criteria for clinical trials and to find prospective participants who fit within those guidelines by processing patient data—structured and unstructured—at participating sites. The system is in use at health systems including Cedars-Sinai Medical Center in Los Angeles and Texas Medical Center in Houston, as well as life sciences companies, medical device makers, and contract research organizations.
A number of companies are offering similar services. Mayo Clinic used the IBM (NYSE: IBM) Watson for Clinical Trial Matching software for 18 months to measure its effect on the recruitment of breast cancer patients for trials testing new systemic therapies. The software system was trained by Mayo subject matter experts before implementation, and a team of coordinators screened the Watson-derived trial matches. Using the system plus the screening coordinator team, Mayo reported an average monthly increase of 80 percent in the number of patients recruited.
Deep 6 AI aims to use its software to improve aspects of the clinical trial process in addition to patient recruitment.
“What we have built is a real-time, real-world evidence platform,” CEO Wout Brusselaers told Xconomy. “We are trying to make sure that we can leverage that data to improve every part of what we call the clinical trial value chain.” He says the system could improve trial design and feasibility, as well as study site selection.
The software also tracks patient outcomes. That data could be analyzed and potentially used to improve future trial designs and enable what are known as adaptive studies, in which researchers modify aspects of a trial while it is underway based on data already gathered. To encourage the use of this type of study, the FDA late last year released guidance documents detailing how to best design, conduct, and report results from such trials. The agency believes using those and other newer clinical trial designs could reduce development costs.
Deep 6 AI’s Series A was led by Point72 Ventures, the VC arm of the eponymous asset management firm. Chinese private equity firm GSR Ventures and unnamed “strategic partners” also invested. The money will go to hire more engineers, data scientists, and customer support staff.
Brusselaers says the company next aims to develop a system that patients could use to find clinical trials for which they may be eligible, launching it by the end of 2020.