MIT Technology Review
“One of the hardest parts of my job is getting patients into trials,” said Nicholas Borys, chief medical officer of Lawrenceville, NJ, the biotechnology company Celsion, which is developing next generation chemotherapy and immunotherapies for liver and ovarian cancer and certain types. of brain tumors. Borys estimates that less than 10% of cancer patients participate in clinical trials. “If we could have achieved that to 20% or 30%, we would probably have beaten several cancers by now.”
Clinical trials test new drugs, devices, and procedures to determine if they’re safe and effective before they’re approved for general use. But the path from study design to approval is long, winding and expensive. Today, researchers are leveraging artificial intelligence and advanced data analytics to speed the process, reduce costs, and deliver effective treatments faster to those who need them. And they draw from an underutilized but rapidly growing resource: data on patients from past studies
Establishing external controls
Clinical trials usually involve at least two groups, or “arms,” a test or experimental arm that receives the treatment under study and a control arm that does not. A control arm may receive no treatment, placebo, or the current standard of care for the disease being treated, depending on the type of treatment being investigated and what it is compared to according to the study protocol. The problem of recruiting researchers studying therapies for cancer and other deadly diseases is easy to see: Patients with a life-threatening disease need help now. Although they may be willing to take a risk for a new treatment, “the last thing they want is to be randomized to a control arm,” says Borys. Combine that reluctance to recruit patients with relatively rare diseases – for example, a form of breast cancer characterized by a specific genetic marker – and the time to recruit enough people can take months or even years . Nine out of ten clinical studies worldwide – not just for cancer but for all types of diseases – fail to recruit enough people within their target period. Some attempts fail completely because there are not enough participants.
What if the researchers didn’t have to recruit a control group at all and could offer the experimental treatment to anyone who agreed to participate in the study? Celsion is exploring such an approach with New York-based Medidata, which provides management software and electronic data capture for more than half of the world’s clinical trials and services most major pharmaceutical, medical device and academic medical centers. Medidata was acquired by the French software company Dassault Systèmes in 2019 and has compiled an enormous “Big Data” resource: detailed information from more than 23,000 studies and almost 7 million patients dating back about 10 years.
The idea is to reuse patient data in previous studies to create “external control arms”. These groups perform the same function as traditional control arms, but can be used in situations where a control group is difficult to recruit: for example, extremely rare diseases or conditions such as cancer that are directly life-threatening. They can also be used effectively for “one-armed” studies that make a control group impractical: for example, to measure the effectiveness of an implanted device or a surgical procedure. Perhaps the most valuable immediate benefit is running rapid preliminary studies to assess whether any treatment up to a full clinical trial is worth pursuing.
Medidata uses artificial intelligence to crawl its database and find patients who have served as controls in previous attempts at treatment for a specific condition to create its proprietary version of external control arms. “We can carefully select these historical patients and match the current experimental arm with the historical study data,” said Arnaub Chatterjee, senior vice president of products, Acorn AI at Medidata. (Acorn AI is the data and analysis department of Medidata.) The studies and the patients are designed in terms of the goals of the study – the so-called endpoints, such as reduced mortality or how long patients remain cancer-free – and other aspects of the aspects of the study, such as the type of Data collected at the start of the study and during the course of the study.
In creating an external control arm, “we do everything we can to mimic the ideal randomized controlled trial,” said Ruthie Davi, vice president of data science, Acorn AI at Medidata. The first step is to search the database for possible control arm candidates based on the main eligibility criteria from the trial study: for example, the type of cancer, the main characteristics of the disease and how far it has progressed, and whether the patient is the first to be treated. It’s essentially the same process used in selecting control patients in a standard clinical study – except that the data recorded at the start of the last study, rather than the current one, is used to determine eligibility, says Davi. “We’re finding historical patients who would qualify for the study if they existed today.”
Download the full report.
This content was created by Insights, the custom content division of MIT Technology Review. It was not written by the editorial staff of the MIT Technology Review.