AI Shows COVID-19 Vaccines May Be Less Effective in Racial Minorities

Artificial intelligence tools examined a kind of vaccine similar to new COVID-19 vaccines and revealed that it could be less effective in people of black or Asian ancestry, according to a study conducted by researchers at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL).

In response to the White House’s Operation Warp Speed – an effort to produce and deliver 300 million doses of COVID-19 vaccines, with initial doses available by January 2021 – companies have raced to develop safe and effective treatments for the virus.

Currently, Moderna and Pfizer’s vaccine candidates are leading the pack. Both have achieved over 90 percent efficacy against the coronavirus, and Pfizer’s vaccine was just granted a temporary emergency use authorization in the UK.

However, the research from MIT suggests that the vaccines may not have the same impact among all patient populations.

CSAIL researchers used artificial intelligence and machine learning tools to examine a form of vaccine similar to Moderna and Pfizer’s.

The team found that among white participants, the amount of people whose immune systems didn’t strongly respond to the vaccine was less than half of one percent. Among Asian participants, that figure was nearly ten percent.

The results indicate that clinical trials measuring the efficacy of COVID-19 vaccines should include diverse cohorts of participants, researchers stated.

“There are obviously many other factors to consider, but our preliminary results suggest that, on average, people of Black or Asian ancestry could have a slightly increased risk of vaccine ineffectiveness,” said MIT professor David Gifford.

“Our work shows that clinical trials need to carefully consider ancestry in their study designs to ensure that efficacy is measured across an appropriate population.”

Unfortunately, the study’s results highlight longstanding trends in healthcare. Certain groups of people are continually underrepresented in clinical trials, often limiting the generalizability of trial results.

For vaccines protecting against COVID-19 infections, possible gaps in coverage among minority populations are especially concerning. Throughout the pandemic, existing racial disparities and health inequities have been intensified and accentuated, with the virus disproportionately affecting black and Hispanic communities.

If a vaccine proves less effective for minority groups, it will only widen the gaps that already plague the industry.

In an effort to prevent this from happening, the CSAIL team also presented a new machine learning-based strategy for improving the vaccines’ effectiveness in certain populations. The approach involves adding a small number of additional COVID-19 peptides to a given dose of the vaccine.

These augmentation vaccines use vaccine elements that have been known to cause immune system responses in COVID-19 patients. Adding between five and ten additional peptides to a particular dose can improve a vaccine’s effectiveness to nearly 100 percent in all patient populations, the team found.

Researchers judged a vaccine to be most effective if at least six COVID-19 vaccine peptides were shown to be displayed by the proteins that regulate the person’s immune system, known as HLE alleles.

This current project adds to MIT CSAIL’s previous work in using AI to design and improve COVID-19 vaccines. In July 2020, the lab developed a machine learning system called OptiVax that selects peptides that are predicted to offer high population coverage for a vaccine.

The OptiVax group plans to partner with other researchers to test the vaccine designs in animal models. The team will aim to discover if pieces of the virus that may cause immunity provide better protection than vaccines currently in clinical trials.

“We evaluated a common vaccine design based on the spike protein for COVID-19 that is currently in multiple clinical trials,” said Ge Liu and Brandon Carter, CSAIL PhD students and lead authors on the paper about OptiVax.

“Based on our analysis, we developed an augmentation to improve its population coverage by adding peptides. If this works in animal models, the design could move to human clinical trials.” 

As companies continue to develop and test COVID-19 vaccines, it will be critical to enroll diverse cohorts of individuals in clinical trials in order to truly support safety and efficacy for all populations.