Deep learning applied to more than 25,000 histopathology slides revealed tissue-specific structural signatures of biological aging that tracked telomere attrition, pathology, comorbidity, and molecular changes more closely than chronological age alone.