To develop tissue clocks that estimate biological age from histology images and identify organ-specific aging patterns linked to disease.
Approach:
Study Design: Analyzed 25,712 whole-slide images from 40 tissue types collected from 983 postmortem donors.
Model Development: Computer models examined tissue features to predict biological age and calculate the tissue age gap.
Validation: Tested models in independent cohorts involving 295 donors and combined histology findings with gene expression data from 1,205 blood samples.
Key Findings:
Models estimated biological age with an average error of approximately five years.
Larger age gaps were associated with shorter telomeres, more comorbidities, and subclinical pathological changes.
In the cerebellum, greater age gaps were linked to myelin loss and ischemic changes; in the aorta, they were associated with wall thickening and structural damage related to vascular disease.
Blood-based models corresponded with age gaps in specific organs related to chronic diseases.
Interpretation:
Histology images and blood gene expression may contain measurable signs of tissue-specific aging.
Limitations:
Study did not determine if age gaps could predict disease before symptoms or diagnosis.
Used postmortem samples with unequal gender representation.
Lacked matching tissue samples in external blood cohorts.
Conclusion:
Prospective studies using prediagnostic samples are needed to assess clinical utility.
