AI ECG Index Tracks Pubertal Maturation in Children and Adolescents

Puberty progresses gradually and can vary widely, making sex-based categories an imprecise proxy in large datasets. Routine electrocardiograms (ECGs) are widely available across inpatient and outpatient settings. To help address this challenge, researchers have developed an artificial intelligence–derived index from standard ECGs that tracks maturation across childhood and adolescence.

The Electrocardiographic Sex Index (ESI), created by investigators at Wake Forest University School of Medicine (Winston-Salem, NC, USA), generates a continuous score from routine ECG data that reflects biological development on a spectrum. The approach is designed to offer a standardized, scalable marker of maturation when Tanner staging or hormone measurements are unavailable. It leverages features contained in ECG waveforms to quantify developmental changes relevant to cardiovascular physiology.

In a retrospective analysis, an adult-trained ESI model was applied without retraining or recalibration to 61,930 ECGs from patients aged 0–18 years in the clinical ECG archive at the University of Tennessee Health Science Center. This design enabled direct comparison of pediatric ECG features against adult benchmarks and supported evaluation across a broad age range typical of hospital and health system datasets. The findings were published in European Heart Journal – Digital Health.

Results showed that ESI values were tightly centered in early childhood and then diverged in opposite directions during late childhood and adolescence, plateauing in mid-to-late adolescence. Similar age-related trends were observed across races, indicating generalizability in a diverse population. Model accuracy improved steadily with age, with older adolescents approaching adult-level performance, suggesting the score captures stepwise maturational changes rather than fixed categories.

The investigators report that ESI may provide a more precise way to account for developmental stage in pediatric cardiology and population studies when standard clinical staging or hormone data are missing. They recommend longitudinal research incorporating Tanner staging, hormone levels, and clinical outcomes to clarify biological significance and to examine how developmental maturity influences cardiovascular risk, treatment response, and long-term outcomes using ECGs already collected in routine care.

Related Links
Wake Forest University School of Medicine