Recurrence-quantification analysis of heart-rate variability and QT-interval variability in pediatric cancer therapy

Abstract Pediatric cancer therapy may perturb autonomic and ventricular repolarization control. We tested whether short-term electrocardiograms (ECGs) reveal stage-specific nonlinear cardiac dynamics between active treatment ( T ) and post-treatment surveillance ( S ), and whether recurrence-quantification analysis (RQA) adds physiological information beyond linear indices. In 40 children diagnosed with cancer ( T = 20, S = 20), we recorded 5-min, supine, lead-II ECGs during routine visits. R-peaks, QRS onset, and T -wave offset were identified to construct beat-to-beat RR and QT-interval series; QTc was derived beatwise using a power-law rate correction. We computed linear indices (RR mean , SDRR, RMSSD, pNN50; QT mean , SDQT; QTc mean , SDQTc) and RQA indices (DET, L, ENTR, LAM, TT, VMAX, T1, T2) for RR, QT, and QTc. The S group showed a longer RR interval (slower sinus rate) and higher short-term vagal indices (RR mean , p = 0.0112; RMSSD, p = 0.0375; pNN50, p = 0.0277). QT and QTc were longer in S (QTc mean , p = 0.0375). In contrast, RR RQA indicated more regular/persistent dynamics in T , with the strongest separation in T2 RR ( p = 0.0047), consistent with slower returns to distinct states. QT-RQA showed no differences, but QTc–RQA revealed longer trapping structure in T (TT QTc , p = 0.0171), implying rate-masked repolarization organization that emerges after correction. Thus, active therapy was associated with lower RR complexity, whereas follow-up was characterized by higher vagal-related indices and longer repolarization intervals. Crucially, QTc analysis revealed the only nonlinear differences, underscoring the need for beatwise rate correction. Combined RR–QTc–RQA from short ECGs may provide a stage-aware, noninvasive biomarker of cardiac well-being in pediatric oncology and warrants validation in larger, exposure-stratified longitudinal cohorts.