BackgroundIdentification of infrequent nonpulmonary vein trigger premature atrial contractions (PACs) is challenging. We hypothesized that pace mapping (PM) assessed by correlation scores calculated by an intracardiac pattern matching (ICPM) module was useful for locating PAC origins, and conducted a validation study to assess the accuracy of ICPM‐guided PM.MethodsAnalyzed were 30 patients with atrial fibrillation. After pulmonary vein isolation, atrial pacing was performed at one or two of four sites on the anterior and posterior aspects of the left atrium (LA, n = 10/10), LA septum (n = 10), and lateral RA (n = 10), which was arbitrarily determined as PAC. The intracardiac activation obtained from each pacing was set as an ICPM reference consisting of six CS unipolar electrograms (CS group) or six CS unipolar electrograms and four RA electrograms (CS–RA group).ResultsThe PM was performed at 193 ± 107 sites for each reference pacing site. All reference pacing sites corresponded to sites where the maximal ICPM correlation score was obtained. Sites with a correlation score ≥98% were rarely obtained in the CS‐RA than CS group (33% vs. 55%, P = .04), but those ≥95% were similarly obtained between the two groups (93% vs. 88%, P = .71), and those ≥90% were obtained in all. The surface areas with correlation scores ≥98% (0[0,10] vs. 10[0,35] mm2, P = .02), ≥95% (10[10,30] vs. 50[10,180] mm2, P = .002) and ≥90% (60[30,100] vs. 170[100,560] mm2, P = .0002) were smaller in the CS‐RA than CS group.ConclusionsICPM‐guided PM was useful for identifying the reference pacing sites. Combined use of RA and CS electrograms may improve the mapping quality.