INTRODUCTION: Atrial late gadolinium enhancement (Atrial-LGE) and electroanatomic voltage mapping (Atrial-EAVM) quantify the anatomical and functional extent of atrial cardiomyopathy. We aimed to explore the relationships between, and outcomes from, these modalities in patients with atrial fibrillation undergoing ablation.METHODS: Patients undergoing first-time ablation had disease severities quantified using both Atrial-LGE and Atrial-EAVM. Correlations between modalities and their relationships with clinical features and arrhythmia recurrence were assessed.RESULTS: In 123 atrial fibrillation patients (60 ± 10 years), Atrial-EAVM was moderately correlated with Atrial-LGE (r = .34, p < .001), with a mean fibrosis burden of 47.2\% ± 14.91\%. Agreement was strongest in the highest tertile of fibrosis burden (mean of differences 16.8\% (95\% CI = -24.4\% to 57.9\%, p = .433). Fibrosis burden was greater for Atrial-LGE than Atrial-EAVM (50.7\% ± 10.7\% vs. 13.7\% ± 7.13\%, p < .005) for patients in the lowest tertile who were younger, had smaller atria and a greater frequency of paroxysmal atrial fibrillation. Both Atrial EAVM and Atrial LGE were associated with recurrence of arrhythmia following ablation (Atrial-LGE HR = 1.02 (95\% CI = 1.01-1.04), p = .047; Atrial-EAVM HR = 1.02 (95\% CI = 1.005-1.03), p = .007). A low fibrosis burden (<15\%) by Atrial-EAVM identified patients with very low arrhythmia recurrence. In contrast, a much higher fibrosis burden (>66\%) by Atrial-LGE identified patients failing to respond to ablation.CONCLUSIONS: We demonstrate for the first time that the level of agreement between Atrial-EAVM and Atrial-LGE is dependent on the level of atrial cardiomyopathy disease severity. The functional consequences of atrial cardiomyopathy are most evident in patients with the highest anatomical extent of disease.