Automatic identification of a location of focal source in atrial fibrillation (AF)
A system and method for automatically identifying a location of focal arrhythmogenic activity are provided. The method includes receiving, via a plurality of electrodes in a heart, a collection of acquisitions, each including a set of electrophysiological (EP) signals measured by the electrodes. A respective direction of arrival (DOA) and a respective distance relative to the electrodes from which the set of EP signals originated are estimated for each acquisition. The acquisitions are aggregated, to form a statistical distribution of the acquisitions as a function of estimated DOA and distance. Using a statistical test, it is checked whether the statistical distribution is consistent, in accordance with a predefined consistency criterion. If the statistical distribution is found consistent, an estimated location in the heart of a focal source of an arrhythmogenic activity that generated the received EP signals is derived from the statistical distribution. The estimated location of the focal source is overlaid on an anatomical map of at least a portion of the heart. The system includes an interface and a processor configured to perform said method.
1. A method for automatically identifying a location of focal arrhythmogenic activity, the method comprising:
receiving, via a plurality of electrodes in a heart, a collection of acquisitions, wherein each acquisition comprises a set of electrophysiological (EP) signals measured by the electrodes;
estimating for each of the acquisitions a respective direction of arrival (DOA) and a respective distance relative to the electrodes, from which the set of EP signals originated;
aggregating the acquisitions, to form a statistical distribution of the acquisitions as a function of estimated DOA and distance;
checking, using a statistical test, whether the statistical distribution of the acquisitions is consistent, in accordance with a predefined consistency criterion;
deriving from the statistical distribution an estimated location in the heart of a focal source of an arrhythmogenic activity that generated the received EP signals when the statistical distribution of the acquisitions is found consistent;
overlaying the estimated location of the focal source on an anatomical map of at least a portion of the heart; and
presenting to a user a notification that a location of focal arrhythmogenic activity was not identified when the statistical distribution of the acquisitions is found to be inconsistent.
2. The method according to claim 1 , wherein, for a given acquisition, estimating the DOA and distance comprises extracting from the set of EP signals in the given acquisition a respective set of relative times of arrival, and estimating the DOA and distance using the extracted relative times of arrival.
3. The method according to claim 2 , wherein aggregating the acquisitions comprises pre-filtering the acquisitions according to the respective set of relative times of arrival extracted from each acquisition, and including in the statistical distribution of the acquisitions only the pre-filtered acquisitions.
4. The method according to claim 3 , wherein pre-filtering the acquisitions according to the extracted set of relative times of arrival comprises the steps of:
using the estimated DOA and distance, calculating for each acquisition a modeled set of relative times of arrival that would have resulted from an EP wave originating from a focal source at the estimated DOA and distance; and
for each acquisition, determining, by applying a predefined geometrical test, a degree of similarity between the extracted set and modeled set of relative times of arrival.
5. The method according to claim 4 , wherein estimating the degree of similarity comprises calculating a cosine-similarity geometrical test between the two sets.
6. The method according to claim 4 , wherein estimating the degree of similarity comprises calculating an estimation error for each relative time of arrival and comparing the estimation error to a given threshold.
7. The method according to claim 4 , and comprising, using the modeled set of relative times of arrival, adjusting time values of annotations over EP signals for which a voltage-time slope of the EP signal is shallower than a prespecified slope.
8. The method according to claim 4 , wherein pre-filtering the acquisitions comprises discarding one or more acquisitions determined to have dissimilar sets of times of arrival.
9. The method according to claim 1 , wherein deriving the estimated location comprises fitting a curve to the statistical distribution and finding a maximum of the curve as a function of estimated DOA and distance.
10. The method according to claim 1 , wherein estimating the DOA and the distance comprises minimizing a cost-function.
11. The method according to claim 10 , wherein minimizing the cost-function comprises minimizing a weighted cost-function.
12. The method according to claim 10 , wherein minimizing the cost-function comprises minimizing the cost-function iteratively, by removing in each iteration an EP signal value having a largest estimation error.
13. The method according to claim 1 , wherein deriving the estimated location comprises applying k-means analysis to the statistical distribution, projecting estimated locations on anatomy and selecting the location having a projected distance that is less than a given value.
14. A system for automatically identifying a location of focal arrhythmogenic activity, the system comprising:
an interface, configured to receive a collection of acquisitions acquired by a plurality of electrodes in a heart, wherein each acquisition comprises a set of electrophysiological (EP) signals; and
a processor, which is configured to:
estimate for each of the acquisitions a respective direction of arrival (DOA) and a respective distance relative to the electrodes, from which the set of EP signals originated;
aggregate the acquisitions, to form a statistical distribution of the acquisitions as a function of estimated DOA and distance;
check, using a statistical test, whether the statistical distribution of the acquisitions is consistent, in accordance with a predefined consistency criterion;
derive from the statistical distribution an estimated location in the heart of a focal source of an arrhythmogenic activity that generated the received EP signals when the statistical distribution of the acquisitions is found consistent;
overlay the estimated location of the focal source on an anatomical map of at least a portion of the heart; and
present to a user a notification that a location of focal arrhythmogenic activity was not identified when the statistical distribution of the acquisitions is found to be inconsistent.
15. The system according to claim 14 , wherein, for a given acquisition, the processor is configured to estimate the DOA and distance by extracting from the set of EP signals in the given acquisition a respective set of relative times of arrival, and estimating the DOA and distance using the extracted relative times of arrival.
16. The system according to claim 15 , wherein, in aggregating the acquisitions, the processor is configured to pre-filter the acquisitions according to the respective set of relative times of arrival extracted from each acquisition, and to include in the statistical distribution of the acquisitions only the pre-filtered acquisitions.
17. The system according to claim 16 , wherein the processor is configured to pre-filter the acquisitions according to the extracted set of relative times of arrival by performing the steps of:
using the estimated DOA and distance, calculating for each acquisition a modeled set of relative times of arrival that would have resulted from an EP wave originating from a focal source at the estimated DOA and distance; and
for each acquisition, determining, by applying a predefined geometrical test, to what degree the extracted set and modeled set of relative times of arrival are similar.
18. The system according to claim 17 , wherein the processor is configured to estimate to what degree the extracted set and modeled set of relative times of arrival are similar by calculating a cosine-similarity geometrical test between the two sets.
19. The system according to claim 17 , wherein the processor is configured to estimate a degree of similarity between the extracted set and modeled set of relative times of arrival, by calculating an estimation error for each relative time of arrival and comparing the estimation error to a given threshold.
20. The system according to claim 17 , wherein the processor is further configured to, using the modeled set of relative times of arrival, adjust time values of annotations over EP signals for which a voltage-time slope of the EP signal is shallower than a prespecified slope.
21. The system according to claim 20 , wherein, in pre-filtering the acquisitions, the processor is configured to discard one or more acquisitions determined to have dissimilar sets of times of arrival.
22. The system according to claim 14 , wherein the processor is configured to derive the estimated location by fitting a curve to the statistical distribution and finding a maximum of the curve as a function of estimated DOA and distance.
23. The system according to claim 14 , wherein the processor is configured to estimate the DOA and the distance by minimizing a cost-function.
24. The system according to claim 23 , wherein the processor is configured to minimize the cost-function by minimizing a weighted cost-function.
25. The system according to claim 23 , wherein the processor is configured to minimize the cost-function iteratively, by removing in each iteration an EP signal value having largest estimation error.
26. The system according to claim 14 , wherein the processor is configured to derive the estimated location by applying k-means analysis to the statistical distribution, projecting estimated locations on anatomy and selecting the location having a projected distance that is less than a given value.