Skip to main content
SCMR 29th Annual Scientific Sessions Main banner
Final Program


Program-At-A-Glance

Rapid Fire Session

Session: Rapid Fire Session 3: Arrhythmias (Monitor 8)

Regional Quantitative Analysis of Atrial Remodelling after Catheter Ablation by isotropic 3D Dixon Late Gadolinium Enhancement

Friday, February 6, 2026
10:29 AM - 10:36 AM  
Location: Galapagos / Aruba / Ilhabela (2nd Lower Level)
Background: CMR enables comprehensive evaluation of cardiac remodeling in atrial fibrillation (AF). Pulmonary vein isolation (PVI), the guideline-recommended therapy in selected AF patients, induces necrosis followed by scarring around the pulmonary veins. Late Gadolinium Enhancement (LGE), the non-invasive gold standard for assessing post-PVI scar formation and atrial remodeling is increasingly being used to transform the understanding of treatment failures and investigate and improve ablation techniques(1). However, conventional 3D LGE sequences often lack spatial resolution for the thin atrial wall. This study aimed to quantify PVI-induced fibrotic changes using a 3D LGE Dixon sequence with isotropic resolution of 1.3mm3 and a normalized image intensity (IIR) method(2), spatially correlated with ablation data.

Methods: 40 AF patients (median disease duration: 126 days; 21 with paroxysmal-AF, 19 with persistent-AF) and 20 healthy controls underwent CMR at 1.5T (MAGNETOM AvantoFit, Siemens Healthineers, Forchheim, Germany) with high resolution (1.3mm3) 3D LGE based on a GRE Dixon research sequence (3). In a subset of 28 patients, bipolar electroanatomic mapping (EAM) (1049±565 mapping sites) was performed prior to PVI. Those maintaining sinus rhythm at six months post-ablation - confirmed by 96-hour Holter monitoring - underwent follow-up CMR. 3D EAM and ablation sites were coregistered to LGE reconstructions by rigid and non-rigid surface deformation for regional analysis (Figure 1).

LA fibrosis was quantified as the percentage of the total LA surface. In addition, ablation induced changes in the matched periablational area (within ±7.5mm surface distance from ablation sites(4)) were compared to changes in the remote (non-ablated) LA wall in 16 cases where follow up reconstructions could be co-registered with EAM.

Results: Using data from the healthy control group (mean + 2SD), an IIR threshold of 1.34 to identify pathologically enhanced LA tissue was established. Persistent AF patients showed higher total LGE burden (median [interquartile range] of 0.63% [0.39-1.08]) than paroxysmal AF patients (2.16% [0.48-3.66]), p=0.030. At 6 months post PVI, an increased total left atrial LGE for paroxysmal AF patients (4.18% [2.36-11.74], p< 0.001) and persistent AF patients (6.56% [3.03-9.32], p=0.243) was observed. This corresponds to an ablation-induced change across both groups of +2.07% [0.52-7.89], p< 0.001. This increase in LGE was localized predominantly in the periablational area of the LA with (increase of 7.74% [2.08-10.07], p< 0.001) whereas no significant change was observed in the rest of the LA wall (increase of 0.33% [-0.85-1.43], p=0.375) (Figure 3).

Conclusion: Isotropic 3D Dixon LGE enables spatially resolved quantitative evaluation of ablation induced atrial remodeling. Ablation induced scarring indicated by LGE occurred predominantly in the periablational area but not in the remote LA myocardium.

Figure 1: Post-Processing Workflow (A) Manual segmentation of isotropic 3D Dixon Late Gadolinium Enhancement (LGE) images was performed to delineate the epicardial left atrial wall. Image intensities were projected onto a shell reconstruction (-3 voxels inward and +1 outward) from this segmentation and normalized by mean blood pool intensity to provide an image intensity ratio. Areas with an intensity ratio >1.34 are denoted in red, >1.2 are denoted in yellow. (B) The 3D electroanatomic mapping reconstruction and ablation sites (red) were (C) coregistered to the segmentation shell. (D) LGE was quantified as the percentage of the total left atrial surface above the threshold after clipping of the mitral valve, pulmonary veins and left atrial appendage (red areas indicate scar). (E) A separate analysis of the atrial wall region within ±7.5mm of the ablation sites (red dots) was performed to evaluate ablation-induced remodeling in comparison to the non-ablated (remote) LA wall (faded area).
Figure 2: Total LGE in paroxysmal and persistent AF at Baseline and 6 months after PVI. At baseline, isotropic 3D LGE Dixon imaging enabled differentiation between patients with persistent and paroxysmal atrial fibrillation based on total fibrotic substrate burden. Six months after the PVI procedure, an increase in total LGE percentage was observed in both groups (paroxysmal p<0.001, persistent: p=0.24, Wilcoxon signed rank test). Boxplots represent median (solid line inside the box), interquartile range (IQR, box) and Q1 - 1.5*IQR or Q3 + 1.5*IQR (whiskers) of the total left atrial LGE amount.
Figure 3: Spatial correlation of ablation sites from electroanatomic mapping (EAM) with isotropic Late Gadolinium Enhancement (LGE) enables quantitative assessment of ablation-induced fibrosis. EAM ablation sites spatially correlated with the LGE reconstruction can be seen in the left half of the plot. Red dots indicate ablation sites, image intensity ratios >1.34 marked in red, image intensity ratios >1.2 marked in yellow. Note that for quantification of scar amount in the remote (non-ablated) LA wall, pulmonary veins, LA appendage and mitral valve were excluded from the quantification. The right side shows the quantitative LGE analysis in the peri-ablational area and the remote LA wall. At 6 months post-PVI, changes in fibrotic substrate were observed predominantly in the peri-ablational area (right, p<0.001, Wilcoxon signed-rank test), with no significant changes in the remote LA wall (left, p=0.375, Wilcoxon signed-rank test). Boxplots represent median (solid line inside the box), interquartile range (IQR, box) and Q1 - 1.5*IQR or Q3 + 1.5*IQR (whiskers) of the left atrial LGE amount in the non-ablated remote LA-Wall (left) and periablational area (right).