“One of the hardest parts of my job is enrolling patients in the study,” said Nicholas Borys, chief medical officer at Celsion, a biotechnology company in Lawrenceville, NJ. The company develops liver and ovarian cancer, as well as certain types of next-generation chemotherapy and immunotherapeutic agents. Of brain tumors. Borys estimates that less than 10% of cancer patients are enrolled in clinical trials. “If we could raise it to 20% or 30%, we would probably have overcome some cancers by now.”
Clinical trials test new drugs, devices, and procedures to determine if they are safe and effective before they are approved for general use. However, the journey from research planning to approval is long, winding and costly. Today, researchers are using artificial intelligence and advanced data analysis to speed up processes, reduce costs, and receive faster and more effective treatments for those who need them.And they are taking advantage of underutilized but rapidly growing resources: data on patients from past trials
Building external controls
Clinical trials usually include at least two groups, or “arms.” A test arm or experimental arm that receives treatment under investigation, and a control arm that does not. Controls may not be receiving treatment, placebo or current standard treatment for the disease being treated, depending on the type of treatment being studied and what is being compared under the study protocol. there is. Recruitment issues are easy to understand for researchers studying treatments for cancer and other deadly illnesses. Patients in a life-threatening condition are now in need of help. They may be willing to risk new treatments, but “the last thing they want is to be randomized to the control group,” says Boris. Combined with that reluctance and the need to recruit patients with relatively rare illnesses (eg, breast cancer forms characterized by specific genetic markers), it takes months to recruit a sufficient number of people. , In some cases it can last for years. Nine of the ten clinical trials worldwide cover all types of conditions, not just cancer, and we are unable to recruit a sufficient number of people within the target period. Some exams fail altogether due to lack of enough participants.
What if researchers could provide experimental treatment to everyone who agreed to participate in the study without having to employ a control group at all? Celsion is exploring such an approach at Medidata, headquartered in New York. Medidata provides management software and electronic data capture for more than half of the world’s clinical trials, servicing most major pharmaceutical and medical device companies and academic medical centers. Acquired by French software company Dassault Systèmes in 2019, Medidata has put together a huge amount of “big data” resources. More than 23,000 trials and detailed information from nearly 7 million patients about 10 years ago.
The idea is to reuse data from patients in past trials to create an “external control arm”. These groups perform the same functions as traditional control arms, but can be used in situations where it is difficult to employ control groups. For example, it can be used for very rare diseases or conditions such as imminent life-threatening cancer. They can also be effectively used in “single-arm” trials that make the control impractical: for example, to measure the effectiveness of an implanted device or surgical procedure. Perhaps their most valuable immediate use is to conduct a rapid preliminary trial to assess whether treatment is worth pursuing to the point of a complete clinical trial.
Medidata uses artificial intelligence to build a database, find patients who acted as controls in past trials of treatment for a particular condition, and create their own version of the external control arm. “We can carefully select these past patients and match the current study group with past study data,” said Arnaub Chatterjee, Senior Vice President of Acorn AI Products at Medidata. (Acorn AI is Medidata’s data and analytics department.) Trials and patients are the objectives of the study (so-called endpoints such as reduced mortality and how long patients remain cancer-free), and other aspects. Will be matched for. Study design, such as the type of data collected at the beginning and during the study.
Ruthie Davi, vice president of data science at Medidata’s Acorn AI, said when creating an external control arm, “we will do everything we can to mimic an ideal randomized controlled trial.” The first step is to search the database for possible control candidates using the primary eligibility criteria for the trial. For example, the type of cancer, the key characteristics and progression of the disease, and whether the patient is new to the disease. Being treated. This is basically the same process used to select control patients in standard clinical trials. The difference is that the data recorded at the start of the previous exam, rather than the current exam, is used to determine eligibility. “We are finding historic patients who would be the subject of the trial if they were present 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 MIT Technology Review.
With big data, clinical trials are better, faster, and cheaper
Source link With big data, clinical trials are better, faster, and cheaper