AI Reads Pre-Vaccine Blood To Predict Your Shot Response
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- Arizona State University researchers examined 8,687 blood samples from 4,089 participants, measuring antibodies against 185 antigens including SARS-CoV-2, common viruses and bacteria, and autoimmune-linked targets, to find pre-vaccination predictors of COVID-19 vaccine response.
- Joshua LaBaer and the team used deep learning to identify "sentinel" antibodies already present before vaccination, with higher levels of antibodies targeting Staphylococcus aureus, RSV, and human respirovirus 3 associated with stronger vaccine responses.
- Immunosuppressed groups including HIV, multiple myeloma, and organ transplant patients were more likely to show reduced vaccine responses, yet some immunosuppressed individuals still mounted strong responses while roughly 5%–6% of healthy participants showed weak ones — showing health category alone does not predict outcome.
- The method reads antibody patterns in blood rather than relying on genetic testing, which the researchers say could make it easier to translate into clinical practice than existing prediction approaches.
- The study, published in Cell Press Blue, could eventually help doctors identify people who need extra doses, closer follow-up, or other protective strategies, and extend to vaccines beyond COVID-19.
- The findings point toward vaccination decisions informed by an individual's pre-existing immune readiness rather than broad health categories, potentially reshaping how vulnerable populations are protected.
Why it matters: If validated, a blood-based antibody test could let clinicians flag the roughly 5%–6% of healthy people who respond poorly to vaccines — and the immunosuppressed patients whose category label currently masks their true response — before they get a shot, enabling extra doses or closer monitoring for those who need it.
Ask SkimNews
