Sensitivity of CIPS-computed PVC location to measurement errors in ECG electrode position: the need for the 3D camera

Area of Science:

• Cardiovascular Electrophysiology
• Medical Imaging and Modeling
• Biomedical Engineering

Background:

• The Cardiac Isochrone Positioning System (CIPS) is a non-invasive method for localizing cardiac arrhythmias like PVCs, VT, and WPW using 12-lead ECG and MRI-derived torso models.
• CIPS currently uses assumed virtual ECG electrode positions, not accounting for patient-specific anatomy, potentially introducing errors in arrhythmia origin localization.

Purpose of the Study:

• To determine the sensitivity of the CIPS model to variations in virtually represented ECG electrode positions.
• To quantify the degree of error introduced by electrode displacement in CIPS-computed arrhythmia origins.

Main Methods:

• A model-based study using previously obtained PVC ablation patient data as controls.
• Simulations involved moving virtual precordial electrodes (V1-6) simultaneously and individually to recalculate PVC origins.
• The distance between the recalculated and control PVC origins served as the primary metric for error assessment.

Main Results:

• Both simultaneous and independent movements of virtual precordial electrodes led to non-linear and unpredictable shifts in the computed PVC origin.
• Shifts ranged from 0 to 62mm for simultaneous electrode movement and 0 to 40mm for independent movement (>10mm displacement).
• The impact of electrode displacement on CIPS accuracy was more pronounced when the virtual electrode was closer to the actual PVC origin.

Conclusions:

• Minor variations in virtual precordial electrode placement significantly alter CIPS-computed PVC origins, rendering the model potentially inaccurate.
• Accurate, patient-specific measurement of ECG electrode positions relative to torso geometry is essential for CIPS reliability.
• Integrating a 3D camera for precise electrode localization is proposed as a solution to enhance CIPS accuracy in arrhythmia mapping.