After kicking off the new year by expanding their year-old data analytics-based research collaboration, Mayo Clinic and nference’s union continues to bear fruit.
The latest product of the partnership is a pair of artificial intelligence-powered studies examining the use of monoclonal antibodies to treat COVID-19 in patients with an elevated risk of requiring hospitalization for the virus due to comorbidities such as hypertension, obesity, diabetes and chronic lung disease.
Monoclonal antibody treatments from Big Pharmas such as Regeneron and GlaxoSmithKline have received emergency use authorizations from the FDA since the start of the pandemic, clearing them for use in treating high-risk patients with mild to moderate cases of COVID—even as the FDA maintains that the three authorized vaccines offer the most effective protection against the virus.
The treatments comprise lab-developed coronavirus antibodies that are designed to bulk up the immune system’s response to the virus, ideally keeping the infection from worsening by stopping its spread throughout the body—a theory lent further credibility by the new studies from nference and the Mayo Clinic.
RELATED: Mayo Clinic, nference to expand data partnership spanning COVID-19 and more
The first of the two studies focused on casirivimab-imdevimab, the monoclonal antibody cocktail present in Regenerton’s REGEN-COV. Using nference’s natural language processing AI software, researchers combed through the lab tests, clinical notes and electronic health records of nearly 1,400 high-risk Mayo Clinic patients with mild to moderate cases of COVID, half of whom were given the antibody treatment.
The analysis found that throughout the four weeks after receiving treatment, the patients who received the antibody infusions were hospitalized at increasingly lower rates than those in the control group. After two weeks, only 1.3% of patients treated with antibodies had been hospitalized, compared to 3.3% of the control group. By the four-week mark, that gap had further widened between the two groups, with hospitalization rates of 1.6% and 4.8%, respectively.
The second study used nference’s AI to examine the effects of bamlanivimab, another monoclonal antibody that is no longer authorized to be administered on its own—after the FDA revoked its EUA in April—but may still be used to treat COVID when combined with etesevimab.
That analysis compared the results of bamlanivimab treatment in 2,335 high-risk COVID patients with those of an untreated control group of the same size. Once again, all-cause hospitalization rates were shown to be significantly lower in the antibody-treated group, clocking in between 1.4 and two percentage points lower than those in the control group after two, three and four weeks.
The researchers also studied ICU admission rates for both groups. The AI analysis found that while the control group’s rate hovered around 1% at each interval, the antibody group’s average admission rate maxed out at 0.56% after four weeks.
“While trials with small cohorts led to the authorization of casirivimab-imdevimab treatment for COVID-19 with monoclonal antibodies, clinical data has been needed to prove its effectiveness,” said AJ Venkatakrishnan, Ph.D., nference’s vice president of scientific research and the co-author of both studies. “These two papers are the latest evidence of how our platform and our scientists are able to provide the real-world evidence necessary to accelerate research and development of important drugs.”
The antibody studies are the latest additions to nference and Mayo Clinic’s growing list of joint research projects, with the bulk of this year’s focusing on COVID-19.
In June, for example, the partners performed a similar AI analysis on the health records related to nearly 133,000 doses of the three FDA-authorized coronavirus vaccines from Pfizer-BioNTech, Moderna and Johnson & Johnson. The study found no significant link between any of the vaccines and cerebral venous sinus thrombosis, a rare condition that causes blood clots in the brain and that was the cause of the 10-day pause on the J&J vaccine in April.
The following month, nference and Mayo Clinic published another study outlining links between certain pre-existing conditions and COVID complications. By analyzing the clinical notes of more than 1,800 Mayo Clinic patients hospitalized due to the coronavirus, the researchers were able to draw definitive lines between 21 risk factors—led by hypertension—and potential complications, including pleural effusion, cardiac arrhythmia, anemia and more.