Analysis of ultra-high-frequency electrocardiograms (UHF-ECG)

Cardiac dyssynchrony, characterized by temporal and sequential abnormalities in ventricular activation, affects approximately 25-30% of patients with advanced heart failure. Cardiac Resynchronization Therapy (CRT) has emerged as an established treatment modality for heart failure (HF) patients with prolonged QRS duration (QRSd) and left bundle branch block. However, approximately one third of patients fail to respond adequately to CRT. For this reason a novel technique, called Ultra-High-Frequency Electrocardiography (UHF-ECG) emerged. It extends frequency analysis, providing enhanced spatio-temporal resolution for the assessment of ventricular electrical dyssynchrony. The aim of this thesis is to conduct a systematic literature review following PRISMA guidelines to synthetize current knowledge regarding UHF-ECG methodologies and applications in CRT patients, and then, to implement and evaluate UHF-ECG analysis for the computation of electrical dyssynchrony and local activation duration parameters. The systematic literature review identified 15 studies from 49 records across four databases. The review revealed a progressive evolution, from 2 kHz to 5kHz. Two principal parameters were defined: electrical dyssynchony (e-dys) as the temporal difference between earliest and latest ventricular activations and local activation duration (Vd) as the UHF-QRS duration at half-maximum amplitude. In this study, data were collected by Czech Technical University in Prague from nineteen patients with CRT system implanted. Patients were characterized by a QRS duration ≥120 ms and New York Heart Association (NYHA) class II and III HF. Seven pacing configurations were acquired with ProCardio-8 system with sampling frequency equal to 1 kHz. The signal pre-processing and processing was performed in MATLAB environment, with differential preprocessing strategies applied based on pacing status. In fact, paced configurations required a three-phase artifact removal procedure prior to QRS detection. The experimental analysis revealed inter-subject variability and e-dys failed to demonstrate the physiologically expected during biventricular pacing. The observed discordance between expected and measured electrical dyssynchrony patterns suggests that e-dys, as currently defined and computed from UHF-ECG signals, may not reliably reflect the dyssynchrony. Future investigations should employ adaptive temporal windowing strategies, higher sampling frequencies to extend frequency band analysis.