Longevity
A blood score built on AI may track how many 'zombie cells' you carry
By Life and Health Today Staff, . Life and Health Today.
A team of researchers has built and tested a blood-based score designed to measure how much cellular senescence a person is carrying, and published the results in the journal Aging Cell in October 2026.
Cellular senescence is what happens when a cell sustains enough damage that it stops dividing but does not die. These cells, sometimes called zombie cells, are a recognised feature of biological ageing. They secrete a cocktail of inflammatory proteins known as the senescence-associated secretory phenotype, or SASP, which can push neighbouring and distant cells into the same damaged state. The idea behind the new score is that measuring SASP proteins in blood gives a practical read on how much of this cellular burden a person is carrying at any given time.
The researchers built their tool using a method called a Guided autoencoder with Transformer, or GAET, a type of deep-learning model that learns patterns across large datasets. They trained it on protein data from the UK Biobank Pharma Proteomics Project, a large-scale population study, and then tested it internally within the broader UK Biobank dataset. The resulting composite SASP Score was, according to the paper, a strong and independent predictor of mortality risk and of the first occurrence of several serious chronic conditions, including dementia, chronic obstructive pulmonary disease, heart attack and stroke.
The researchers then took the score to an independent cohort drawn from a randomised clinical trial, which is the study design that can establish cause and effect rather than just association. In that separate group, participants who underwent what the paper describes as multimodal exercise showed a significant change in their SASP Score trajectory over 18 months.
What the study does not establish is whether changing the SASP Score actually changes health outcomes. A score that moves in response to exercise is a surrogate marker, meaning it is a measurable signal that researchers hope tracks something meaningful, not direct evidence that the people exercising lived longer or avoided the conditions the score predicts. That distinction matters enormously for anyone tempted to read the exercise finding as proof that exercise reduces senescence burden in a clinically meaningful way. It may, but this study does not show it.
The score is also not available as a clinical test. It was developed and validated in research settings using large proteomics datasets, and the paper frames it as a tool for ageing research and for evaluating interventions, not as something a clinician or a patient can order today.
The open question the authors themselves point toward is whether the SASP Score can be used to measure whether senolytic drugs, which are experimental compounds designed to clear senescent cells, are actually working in people. That would require trials that use the score as an endpoint and then track whether patients whose scores improve do better over time. Those trials have not been reported here.
The score's cross-platform utility, which the paper highlights, matters because proteomics data, the large-scale measurement of proteins in blood, is collected differently across research sites and clinical labs. A tool that works across platforms would be more useful than one tied to a single measurement method. Whether it holds up across the full range of clinical settings remains to be tested.
For a reader weighing whether any of this is actionable today, the honest answer is that it is not, at least not directly. The finding that exercise shifted the score is consistent with a large body of evidence that physical activity affects inflammatory biology, but the SASP Score itself is a research instrument at this stage. Whether it will become a standard clinical biomarker depends on whether future trials can show that moving the score moves outcomes. Anyone considering interventions aimed at senescence, whether lifestyle or pharmaceutical, should discuss the current state of the evidence with a clinician who knows their individual history.