Researchers have developed artificial intelligence (AI)-based “tissue clocks” that can estimate the biological age of human organs from microscopic tissue images. Scientists at the CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences and the Ludwig Boltzmann Institute for Network Medicine (LBI-NetMed) at the University of Vienna analysed more than 25,000 tissue samples across 40 tissue types. Their findings, published in Nature Medicine, suggest that organs age at different rates and that some of these ageing patterns may even be detectable through blood samples.
People of the same chronological age can differ greatly in how old they appear or how well their bodies function. This raises an important question: do individual organs also age at different speeds? Led by CeMM and LBI-NetMed Principal Investigator André Rendeiro, the research team used AI to investigate how the microscopic architecture of human tissues changes with age, rather than focusing only on molecular markers such as DNA methylation or gene expression.
The researchers used data from the Genotype-Tissue Expression Project (GTEx), which collected samples from 983 individuals across 40 tissue types, including the brain, heart, lungs, pancreas, skin, and intestine. In total, they analysed 25,712 high-resolution tissue images, representing approximately 480 million individual image tiles, using advanced computer vision models.
The analysis showed that age was the strongest factor associated with changes in tissue appearance across all 40 tissue types. Building on this finding, the researchers developed “tissue clocks” capable of estimating biological age independently for different organs. These models predicted age with an average error of just 4.9 years. The estimated biological ages were also associated with established indicators of ageing, including shorter telomeres, tissue abnormalities, and the number of chronic diseases an individual had.
Importantly, the study found that organs do not follow the same ageing schedule. The lungs, kidneys, pancreas, and adrenal glands showed signs of accelerated ageing as early as ages 20 to 40, while other tissues followed more complex patterns later in life. The uterus showed a particularly notable change around menopause. Health conditions were also associated with organ-specific ageing: kidney failure was linked with accelerated ageing signals across several tissues, while diabetes had especially pronounced effects in the pancreas.
The tissue clocks also identified individuals whose organs appeared biologically older than their chronological age. According to the researchers, AI can detect subtle structural changes that may be difficult for the human eye to recognise, allowing ageing to be viewed as a process of architectural remodelling within tissues rather than simply the accumulation of molecular changes.
Because obtaining tissue samples is often invasive or impractical, the researchers investigated whether the same organ-ageing signals could be detected in blood. By linking blood-based gene expression with tissue age estimates from the same individuals, they developed predictors capable of estimating tissue-specific biological ageing using blood samples alone. This could potentially provide a much less invasive way to assess how individual organs are ageing.
The blood-based models detected ageing patterns associated with conditions including Alzheimer’s disease, Crohn’s disease, cystic fibrosis, vasculitis, diabetes, and stroke. Alzheimer’s disease produced its strongest ageing signal in the brain, while Crohn’s disease showed accelerated ageing across the gastrointestinal tract. Although further research is needed before these methods could become routine clinical tests, the findings suggest that combining AI, tissue imaging, and blood analysis could eventually help monitor organ health, detect disease-related changes earlier, and provide a more personalised picture of biological ageing.
More information: Ernesto Abila et al, Histological aging signatures for monitoring tissue-specific aging and disease, Nature Medicine. DOI: 10.1038/s41591-026-04566-5
Journal information: Nature Medicine Provided by CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences
