Detection of left atrial fibrosis in LGE-MRI and high-density electroanatomical mapping: finding a threshold for the best match

Abstract

Abstract Background Fibrotic atrial cardiomyopathy plays an important role in determining the outcome of ablation in patients with atrial fibrillation (AF). Two main methods are being used for the evaluation of fibrosis: voltage based high-density (HD) electroanatomical mapping (EAM) and late gadolinium enhancement MRI (LGE-MRI). The comparability between both methods in detecting fibrosis has not been systematically investigated. Method LGE-MRIs of the left atrium (LA) were performed in 21 patients (pts). The extent and distribution of fibrosis were evaluated using a custom-designed software generating a 3D model of the LA. HD-electroanatomical maps were recorded in each patient using a 3D mapping software (CARTO3, Biosense Webster). The MRI-3D-model of the LA was then integrated in the 3D mapping system. After processing the HD maps and the MRI models by excluding the mitral valve, pulmonary veins and the left atrial appendage, the atrium was then divided into three independent areas: anterior and posterior wall, and intra-atrial-septum. The LGE areas were then measured using the 3D surface measurement tool. These were then compared to the low voltage areas (LVA) measured in the HD maps using three different cut-off values of 0.5 mV, 0.7 mV, and 1.0 mV. Results In MRI 5 pts had no significant fibrosis (fibrotic area ≤1 cm2). 16 pts showed fibrosis on the anterior wall (7.3 cm2±6.2 cm2). 15 pts showed fibrosis at the septum (5 cm2±2.9 cm2) and 9 pts showed fibrosis on the posterior wall (3.9 cm2±2.5 cm2). Using EAM with a cut-off voltage value of 0.7 mV, 3 pts had no fibrosis. 15 pts showed fibrosis on the anterior wall (7.7 cm2±6.8 cm2). 16 pts showed fibrosis at the septum (5.3 cm2±4.1 cm2) and 8 patient showed fibrosis on the posterior wall (3.5 cm2±1.8 cm2). Using a lower voltage cut-off value of 0.5 mV, fibrosis was evident in less patients with smaller fibrotic areas in LA. Whereas when using a cut-off value of 1.0 mV more patients showed a greater extent of the LA fibrosis. We found significant differences between the two methods (MRI vs. 3D-mapping) when using a cut-off value of 0.5 mV (Wilcoxon, p<0.001 for the anterior and posterior wall and the septum). A cut-off value of 1.0 mV also showed significant differences on the anterior and posterior walls (Wilcoxon, p<0.001). When using a cut-off value of 0.7 mV we did not find a significant difference between the two groups (Wilcoxon, p=0.365) with a correlation coefficient of 0.97 (Pearson; p<0.001) for the anterior wall, 0.69 (Pearson; p=0.007) for the posterior wall, and 0.81 (Pearson; p=0.002) for the septum. Conclusion A cut-off value of 0.7 mV in EAM correlated best with the findings in LGE-MRI. The currently used predefined cut-off value of 0.5 mV in EAM may lead to an underestimation of the actual extent of the fibrosis. Increasing the voltage cut-off value towards 0.7 mV may therefore lead to a better characterisation of fibrotic atrial cardiomyopathy. Funding Acknowledgement Type of funding sources: None.