⚠️ Auto-extracted by Claude on 2026-06-09 — not verified against the full PDF (DOI 10.1038/s41586-026-10242-y; closed-access, no local full text). Effect sizes below are drawn from the abstract and reputable secondary coverage (ASCO Post, press releases) and may be approximate; verify against the primary source before relying on quantitative claims. gap/no-fulltext-access

Thymic health consequences in adults

TL;DR

Companion population paper to Bernatz 2026 (immunotherapy). The group trained a deep-learning framework to score “thymic health” from routine chest CT — quantifying thymic size, shape, and composition (i.e., the degree of functional parenchyma vs. fatty involution) — and showed that higher thymic health predicts substantially lower all-cause and cardiovascular mortality and lower lung-cancer incidence in two large prospective cohorts of asymptomatic adults. The result repositions the adult thymus from “metabolically inert post-childhood remnant” to a measurable, prognostically meaningful axis of immune aging (immunosenescence, disabled-adaptive-immunity).

Design

  • Method: deep-learning model applied to routine chest CT, automatically segmenting and scoring the thymus on size/shape/composition into a continuous thymic health score. (Exact architecture, training labels, and segmentation pipeline not verifiable without full text. gap/no-fulltext-access)
  • Discovery/primary cohort: National Lung Screening Trial (NLST), n ≈ 25,031 screening-CT participants; ~12-year follow-up.
  • Independent validation cohort: Framingham Heart Study (FHS), n ≈ 2,581 (chest CT subset); longitudinal follow-up. Framingham provides external, non-lung-cancer-enriched replication of the cardiovascular-mortality association.
  • Total: ~27,612 individuals.
  • Adjustment: associations reported after adjustment for age, sex, smoking, and comorbidities.

This is observational imaging epidemiology — no intervention; thymic health is an exposure, not a manipulated variable. Causality is not established (a healthier thymus may mark, rather than cause, systemically lower risk).

Key findings (approximate — pending full-text verification)

High vs. low thymic health (per secondary coverage):

OutcomeEffectAbsolute (where reported)
All-cause mortality~50% lower risk12-yr mortality ~13.4% (high) vs ~25.5% (low), NLST
Cardiovascular mortality~63% lower risk (headline); replicated independently in Framingham
Lung-cancer incidence~36% lower6-yr incidence ~3.4% (high) vs ~5.3% (low)
  • The cardiovascular-mortality association is the one that replicated in the independent Framingham cohort (independent of age, sex, smoking) — the strongest cross-cohort signal, and notable because Framingham is not a lung-cancer-screening population.
  • Higher thymic health correlated inversely with systemic inflammation, including il-6 and other pro-inflammatory markers, and with metabolic dysregulation (elevated BMI/obesity) — consistent with a thymic-involution → reduced naive-T-cell output → inflammaging axis. gap/no-fulltext-access (specific inflammatory analytes and effect sizes unverified)

Why it matters for this wiki

  1. First large-scale demonstration that an adult organ-aging readout of the thymus is prognostic for hard endpoints. Prior thymic-involution evidence in this wiki (immunosenescence, disabled-adaptive-immunity) rested on mechanism and small interventional pilots (e.g. the Fahy 2019 TRIIM trial, n=9 completers). This adds population-scale outcome data.
  2. Establishes a candidate imaging biomarker of immune age — see thymic-health-score — derivable for free from CT scans already acquired for other indications (opportunistic imaging), comparable in spirit to opportunistic coronary-calcium or vertebral-density scoring.
  3. Supports thymic regeneration as a longevity target by showing the variable it would move (thymic composition) tracks mortality at population scale — though it does not show that intervening on thymic health changes outcomes. gap/no-mechanism (association, not causation)

Extrapolation / evidence quality

DimensionStatus
Human evidence?yes — two large human cohorts (n ≈ 27,600), no animal model needed
Independent replication?partial — CV-mortality replicated in Framingham; full panel single-group, single-method, brand-new (2026), not yet externally replicated by other teams gap/needs-replication
Causal (vs. associative)?no — observational; thymic health may be a marker of systemic health rather than a driver gap/no-mechanism

See also