Quantification of interventricular dyssynchrony using a standard 12-lead ECG based solution to the inverse ECG problem
Translated title
Kvantificering af interventrikulær dyssynkroni ved brug af en standard 12-aflednings EKG baseret løsning af det inverse EKG problem
Authors
Møller, Maria Goncalves ; Grønborg, Gry Bruun ; Samuelsen, Frederik
Term
4. term
Publication year
2022
Submitted on
2022-06-01
Pages
103
Abstract
Cardiac resynchronization therapy (CRT) is used to treat severe left bundle branch block (LBBB), a condition where electrical signals reach the left ventricle late. Yet more than 30% of patients do not benefit, likely because their hearts are not truly electrically out of sync. To improve candidate selection, this study asked whether interventricular dyssynchrony—the difference in activation timing between the left and right ventricles—can be estimated from a standard 12‑lead electrocardiogram (ECG) using a generic heart‑torso model by solving the inverse ECG problem (inferring heart activation from body‑surface signals). We analyzed 12‑lead ECGs from 100 healthy individuals, 100 people with LBBB, and 135 patients evaluated for CRT. Ventricular activation was represented with a simplified six‑dipole model. We compared methods to stabilize the inverse solution (Gaussian regularization and Tikhonov regularization with λ=0 and λ=30). Dyssynchrony was quantified as left‑right ventricular uncoupling (LRVU) and compared with common ECG markers: QRS duration, QRS area, and Q‑LV (a timing measure used in CRT). Tikhonov regularization with λ=0 produced the most accurate ECG reconstructions in both healthy and LBBB groups. LRVU was higher in LBBB than in healthy subjects. In LBBB and CRT groups, LRVU showed stronger correlations with QRS duration and QRS area (r^2=0.359 and 0.791 in LBBB; r^2=0.328 and 0.630 in CRT) than in healthy individuals (r^2=0.178 and 0.093). In CRT patients, LRVU also correlated with Q‑LV (r^2=0.441). These findings show that interventricular electrical dyssynchrony can be quantified from standard 12‑lead ECGs using a generic heart‑torso model by solving the inverse ECG problem.
Cardiac resynkroniseringsterapi (CRT) bruges til at behandle udtalt venstresidigt grenblok (LBBB), hvor de elektriske signaler når venstre hjertekammer for sent. Alligevel har over 30% af patienterne ingen effekt, sandsynligvis fordi hjertet ikke er elektrisk ude af takt. For at forbedre udvælgelsen af patienter undersøgte vi, om interventrikulær dyssynkroni—forskellen i aktiveringstid mellem venstre og højre hjertekammer—kan estimeres ud fra et standard 12‑aflednings‑EKG ved hjælp af en generisk hjerte‑torso‑model ved at løse det inverse EKG‑problem (at udlede hjertets aktivering ud fra signaler på kroppens overflade). Vi analyserede 12‑aflednings‑EKG’er fra 100 raske personer, 100 med LBBB og 135 patienter vurderet til CRT. Ventrikulær aktivering blev beskrevet med en forenklet model med seks dipoler. Vi sammenlignede metoder til at stabilisere den inverse løsning (Gaussian‑regularisering og Tikhonov‑regularisering med λ=0 og λ=30). Dyssynkroni blev kvantificeret som left‑right ventricular uncoupling (LRVU) og sammenlignet med velkendte EKG‑mål: QRS‑varighed, QRS‑areal og Q‑LV (et tidsmål brugt i CRT). Tikhonov med λ=0 gav den mest nøjagtige rekonstruktion af EKG i både raske og LBBB‑grupper. LRVU var højere hos LBBB end hos raske. I LBBB‑ og CRT‑grupperne korrelerede LRVU stærkere med QRS‑varighed og QRS‑areal (r^2=0.359 og 0.791 i LBBB; r^2=0.328 og 0.630 i CRT) end hos raske (r^2=0.178 og 0.093). Hos CRT‑patienter korrelerede LRVU også med Q‑LV (r^2=0.441). Konklusionen er, at interventrikulær elektrisk dyssynkroni kan kvantificeres ved at løse det inverse EKG‑problem ud fra et standard 12‑aflednings‑EKG og en generisk hjerte‑torso‑model.
[This apstract has been rewritten with the help of AI based on the project's original abstract]
Keywords