Fri, 04.10.2024 10:30

Group Seminar: Mathematical Methods in Medicine and Life Sciences

Patricia Martinez Diaz, Karlsruhe Institute of Technology, Title: Digital-twin based assessment of atrial arrhythmias: influence of personalization strategies

TITLE: Digital-twin based assessment of atrial arrhythmias: influence of personalization strategies

ABSTRACT:

Introduction: Personalization means translating patient-specific data into model parameters reflecting specific properties of the individual. Yet, the impact of different personalization approaches, whether anatomical or functional, on the vulnerability to arrhythmia in personalized atrial computer models remains incompletely understood.


Methods: In a cohort of 8 patient-specific models, we assessed the effect of incorporating the right atrium (RA) on arrhythmia vulnerability across 3 substrate conditions: healthy (H), mild (M), andsevere (S) fibrotic remodeling. We developed 2 models: a monoatrial comprising only the left atrium (LA) and a biatrial incorporating both the RA and LA. In another cohort of 7 patient-specific models, we examined the impact of incorporating clinical effective refractory period (ERP) measurements on arrhythmia vulnerability across 4 configurations: homogeneous (A), heterogeneous (B), regional (C), and continuous (D). The first two configurations were non-personalized based on literature data, the latter two were personalized based on measurements.


Results: Incorporating the RA increased the mean LA vulnerability ratio by 115.8% in state M and 29.0% in state S. No arrhythmia was induced in the H models. RA inclusion increased LA inducibility revealing 5.5±3.0 new inducing stimulus points per patient in the LA for the biatrial model, which did not induce reentry in the monoatrial model. ERP scenario A was the least vulnerable to arrhythmia (3.4±3.9%), while scenario C was the most vulnerable (9.0±5.1%). Compared to the standard non-personalized approach (B), incorporating ERP as a continuous distribution (D) decreased the vulnerability slightly (7.6±3.4%
vs. 7.0±3.6%).

Conclusions: LA arrhythmia vulnerability in biatrial models is higher than in LA monoatrial models. Incorporating the RA unmasked potential inducing points in the LA. The incorporation of patient-specific ERP values impacts the assessment of arrhythmia vulnerability and the type of personalization affects the likelihood of arrhythmia inducibility. Using anatomically and functionally detailed models may enhance the evaluation of substrate vulnerability for patient-specific therapy planning tools.