Photo by Markus Spiske from Pexels
COVID-19 has complicated oncology trials and oncology care in general, slowing down the success seen against cancer.
The cancer death rate fell 29% between 1991 and 2017, according to the American Cancer Society. Novel treatments such as immunotherapy helped, but the coronavirus pandemic has created challenges.
Researchers have had to adapt and innovate, with AI playing an important role.
COVID-19’s disproportionate impact on oncology
“Cancer patients have been disproportionately negatively affected by COVID-19,” said Jeff Elton, CEO of ConcertAI, an AI company specializing in oncology. Immunocompromised patients with hematological malignancies such as multiple myeloma, in particular, face outsized risks from COVID-19 infection.
To reduce the risk for patients, many oncology clinics in 2019 limited capacity while making significant treatment changes. “It wasn’t unusual to see people queuing through tents in the parking lot and being very segregated in different zones as they were going through treatment,” Elton said.
Such an environment made it difficult to run traditional clinical trials in the most disruptive pandemic phase in 2020. “The classic execution models of doing clinical research were very difficult to deploy” Elton said.
And yet, medical oncologists and biomedical research organizations are aware of the potentially life-saving option that clinical trials offer cancer patients. There’s a value to knowing whether experimental therapies are on track to winning regulatory approval and finding broader use.
Cancer trials adapt
To maintain clinical trial progress, clinical trials have adapted in the past year in several ways. First, sponsors have simplified study designs that reduce the burden to patients and clinical and medical oncology sites.
“You saw biopharma beginning to look to move things that had historically been largely at only the academic medical centers into high-capability community settings,” Elton said.
Such changes in venue have opened up access to clinical studies in the past year. “If we make clinical studies more like the standard of care, then it’s a lot easier to move them into community settings,” Elton said.
The pandemic has also played a role in driving interest in decentralized trial designs.
Assisting in this are machine learning and other AI tools for patient identification and scrutinizing study eligibility. Such tools are “now becoming almost de facto standard,” Elton said. They can optimize clinical trial design, analyzing, for instance, electronic medical record data for prospective clinical trials. AI tools can also help formulate inclusion and exclusion criteria for clinical trials. AI-based tools can also “minimize the number of patients required to get to a certain level of statistical power,” Elton said.
They can also boost patient recruitment by ensuring that a clinical trial is as close to the standard care treatment as possible. In that way, “the provider is going to decide what’s in the best interest of the patients, and patients are going to decide what’s in their own best interest,” Elton said.
While there has been a buzz around AI in healthcare and clinical trials in the past, COVID-19 is catalyzing change. “All of a sudden, you see the merits of alternative approaches,” Elton said. “I’m not aware of a single sponsor that would plan on returning back and just restoring all the old ways of working,” he added.
Healthcare providers committed to biomedical research are in a similar position. “In fact, if anything, providers are seeing the integration of localized AI and ML technologies to augment their research capacity as being super important,” Elton said. “We’re seeing this from the sponsors, too.”