Validation
Since the outcome of the WPW24.com based ECG analysis does not point to
a single AP location but to a spectrum of anatomic locations with a
gradient of probability – validation of the method in a traditional
manner (i.e., the predicted site vs. the site of successful ablation)
was not feasible. Even apparently very odd AP locations (for a
particular QRS pattern) suggested by the WPW24.com might be validated as
correct when a sizable number of APs, that included such rare variants,
is analysed. Therefore, to validate WPW24.com, we examined whether the
distribution of APs used by the application to generate the maps is
representative. Validation of distribution patterns was performed only
for the full preexcitation as inherently more precise. For this,
comparison of AP distribution in the exploratory cohort that was used by
the application with the AP distribution in two external cohorts was
made. The first dataset used for validation was obtained from new
patients studied in years 2018-2023 in four centres. The second dataset
for validation of AP distribution was obtained from the seven published
studies that proposed new AP localization algorithms. The validation
phase was commenced once the work on the web-based application, and the
exploratory cohort, were completed.