Particle.news
Download on the App Store

AI Reads Chest CTs to Quantify Chronic Stress With New Imaging Biomarker

An adrenal volume index extracted from CT scans predicted future cardiovascular events in an older multi-ethnic cohort.

Overview

  • Johns Hopkins researchers used deep learning to segment adrenal glands on routine chest CTs and derive an Adrenal Volume Index normalized by height.
  • In 2,842 MESA participants (mean age about 69), higher index values aligned with perceived stress scores, multi-sample salivary cortisol exposure and greater allostatic load.
  • The imaging metric was associated with higher left ventricular mass index, indicating links to cardiac structural changes.
  • Each 1 cm³/m² increase in the index predicted higher risks of heart failure and all-cause mortality over as long as 10 years, with hazard ratios near 1.04 for both outcomes.
  • The results, presented for the RSNA meeting, highlight a potential opportunistic tool from existing scans without added testing, though evidence remains retrospective and needs prospective validation.