Data, scripts, and functions from: Classifying human atrial electrograms and generating patient-specific models of the left atrial posterior wall using point cloud data to simulate electrograms arising from different tissue substrates

Public

  • Cardiovascular diseases cause the majority of deaths worldwide, increasing the need for physiologically accurate computational patient models to study these conditions. This project aimed to develop patient-specific models of the Left Atrial Posterior Wall (LAPW) using Point Cloud Data (PCD) and evaluate model performance using unipolar Electrogram (EGM) signals.

    Data recorded from 5 Paroxysmal and 5 Persistent Atrial Fibrillation (AFib) patients who were paced back to sinus rhythm was analyzed and used for this work. From this PCD data, all the EGMs from each different patient were characterized and classified into three different categories: smooth biphasic, complex multiphasic, and other. Differences were found in EGM characteristics and distribution of waveforms’ categories among patients of both types of AFib.
    Models for two patients were created using PCD LAPW points without further image segmentation. To evaluate model performance, a sample of EGMs coming from monodomain simulations with healthy tissue substrate was compared to measured EGMs classified as smooth biphasic. Overall, the comparison yielded mixed results, showing excellent matches for some EGMs while showing marked differences for others which origin remained undetermined.

    To evaluate EGM morphology changes under fibrotic tissue conditions, four patient models were generated using the same technique but with fibrotic tissue substrate. Different types and degrees of fibrosis were simulated, with simulation results showing increasing multiphasic morphology behavior as fibrosis degree increased. The comparison of these waveforms did not yield results strong enough to determine the specific tissue substrate present on each patient.

    Overall, the results show that PCD is sufficient to create patient-specific models of the LAPW. These models can simulate EGMs that are comparable to measured waveforms, whose different morphologies can potentially be used to determine atrial substrate modifications.

    These data, scripts, and functions are part of the work done/used for Nataren's PhD dissertation. More information on the data and how to use it can be found in his dissertation document titled: Classifying Human Atrial Electrograms and Generating Patient-specific Models of the Left Atrial Posterior Wall Using Point Cloud Data to Simulate Electrograms Arising from Different Tissue Substrates
    ... [Read More]

Total Size
8 files (1.35 GB)
Data Citation
  • Nataren Moran, J. (2024). Data, scripts, and functions from: Classifying human atrial electrograms and generating patient-specific models of the left atrial posterior wall using point cloud data to simulate electrograms arising from different tissue substrates. Duke Research Data Repository. https://doi.org/10.7924/r4br90c8z
DOI
  • 10.7924/r4br90c8z
Publication Date
ARK
  • ark:/87924/r4br90c8z
Type
Related Materials
Title
  • Data, scripts, and functions from: Classifying human atrial electrograms and generating patient-specific models of the left atrial posterior wall using point cloud data to simulate electrograms arising from different tissue substrates
This Dataset
Usage Stats