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

Publication/Presentation Date



BACKGROUND: The Cardiac Isochrone Positioning System (CIPS) is a non-invasive method able to localize the origins of PVCs, VT and WPW from the 12 lead ECG. The CIPS model integrates a standard 12-lead ECG with an MRI derived model of the heart, lungs, and torso in order to compute the precise electrical origin of a PVC from within the myocardium. To make these calculations, CIPS uses virtually represented ECG electrode positions. These virtual electrode positions, however, are currently assumed to represent the standard 12 lead positions in the model without taking into account the actual, anatomical locations on a patient. The degree of error introduced into the CIPS model by movement of the virtual electrodes is unknown. Therefore, we conducted a model-based study to determine the sensitivity of CIPS to changes in its virtually represented ECG electrode positions.

METHODS: Previously, CIPS was tested on 9 patients undergoing PVC ablation, producing a precise myocardial PVC location for each patient. These initial results were used as controls in two different simulation experiments. The first moved all virtual precordial leads in CIPS simultaneously up and down to recalculate a PVC origin. The second moved each virtual precordial lead individually, using 8 points on multiple concentric circles of increasing radius to recalculate a PVC origin. The distance of the newly calculated PVC origin from the control origin was used as a metric.

RESULTS: Moving either all electrodes simultaneously or each V1-6 precordial electrode independently resulted in non-linear and unpredictable shifts of the CIPS-computed PVC origin. Simultaneously moving all V1-6 precordial electrodes by 10mm increments produced a shift in CIPS-computed PVC origin between 0 and 62mm. Independently moving an electrode, a shift of more than 10mm resulted in an unpredictable CIPS-computed PVC origin relocation between 0 and 40mm. The effect of moving the virtual electrodes on CIPS modeling more pronounced the closer the virtual electrode was positioned to the actual PVC origin.

CONCLUSIONS: Slight changes in the virtual positions of the V1-6 precordial electrodes produce marked, non-linear and unpredictable shifts in the CIPS-computed PVC origin. Thus, any variation in the physical ECG electrode placement on a patient can result in significant error within the CIPS model. These large errors would make CIPS useless and underscore the need for accurate, patient specific measurement of electrode position relative to the patient specific torso geometries. A potential solution to this problem could be the introduction of a 3D camera to incorporate accurate measurement of physical electrode placement into the CIPS model. Since the 3D camera software integrates the 3D imaged position of the electrode with the MRI derived torso model, it is conveniently incorporated in the next generation CIPS software to decrease the errors in modeled location of the electrodes. Thus, the 3D camera will be the III(rd) component of the CIPS to increase its accuracy in PVC, VT, and WPW localization.





First Page


Last Page





Medicine and Health Sciences




Department of Medicine, Cardiology Division

Document Type