Validation of a novel semi-automated ECG quantification tool, applied to a cardio-oncology : Semi-automated ECG Tool applied to cardio-oncology

Electrocardiogram (ECG) analysis is crucial to detect cardiotoxicity. Manual methods are time-consuming and limited by inter-reader variability, highlighting the need for precise, reproducible and rapid semi-automated digital tools in clinical practice.

This study evaluates the triplicate concatenation method (TCM) using a semi-automated ECG software (CalECG-4.2, AMPS®) by assessing intra- and inter-reader variabilities in two distinct cardio-oncology populations: breast cancer patients receiving ribociclib (a QT-prolonging drug) and patients admitted with severe immune checkpoint inhibitors (ICI)-myocarditis, a condition marked by QRS alterations.

A total of 420 ECG from 31 patients (21 ribociclib, and 10 ICI-myocarditis) were independently analyzed by two readers. Variability was assessed using Bland-Altman analyses and intraclass correlation coefficients (ICC). Linear mixed-effects modelling quantified time-dependent changes in heart rate (HR), PR, QTc (Fridericia's HR correction), QRS duration and voltage (Sokolow-Lyon) accounting for inter-reader variability.

Intra and inter-reader reproducibility was excellent (ICC > 0.98, including Sokolow-Lyon voltage; standard-deviation < 4 ms across all time-derived parameters). In ribociclib-treated patients (cycles of 21/28 days on drug), QTc peaked at day 14 (16 ± 1 ms, p < 0.001) before decreasing by day 28 (-6 ± 1 ms, p < 0.001) compared to baseline. In ICI-myocarditis, QRS duration increased at day 5 before returning to baseline starting day 28, while Sokolow-Lyon voltages increased progressively on immunosuppressive treatments, peaking at day 28 (458 ± 49 µV, p < 0.001) and remaining constant afterwards for the next month.

TCM with CalECG-4.2 ensures a high reproducibility while monitoring key parameters like QTc duration and Sokolow-Lyon voltage, making it a reliable and time-saving alternative for the ECG surveillance of drug toxicities in cardio-oncology.