Integration of electroanatomical mapping data with whole human heart histology to identify the morphological characteristics of VT substrate in patients with NICM

FUNDING ACKNOWLEDGEMENTS: Type of funding sources: None. INTRODUCTION: The dominant mechanism of ventricular tachycardia (VT) in patients with non-ischemic cardiomyopathy (NICM) is re-entry related to fibrosis. Fibrosis in NICM is fundamentally different from post-infarct scars. Fibrosis does not equal VT substrate. The amount and characteristics of fibrosis related to VT in NICM are unknown. Aim To describe histological characteristics associated with VT substrate in patients with NICM using integration of electroanatomical mapping (EAM) data and whole human heart histology. METHODS: EAM data of 10 patients with NICM who died or underwent heart transplantation after catheter ablation (CA) for VT were reviewed. VT related sites were categorized as class I (entrainment response/VT termination during CA), or class II (other ablation target sites). Explanted hearts were sliced and stained. Ablation lesions were identified and excluded from histological analysis. EAM and whole heart histology were integrated and class I/II sites projected onto the slices using reversed registration and assigned to a segment (32 segment model). VT exit sites (based on 12 lead VT-ECG) were indicated on a 16-segment model (fig 1). The amount and homogeneity of fibrosis was determined. A new method to measure pattern and inhomogeneity of fibrosis was developed with scores between 0 (inhomogeneous confluent) and 0.5 (homogeneous diffuse). RESULTS: In 10 patients (9 with a class IV/V mutation) a median of 11 (range 3-18) VTs were induced. A total of 37 class I sites could be identified and 182/320 segments contained class II sites. After exclusion of ablation lesions, the LV had a median of 24.7% fibrosis (range 11.3-39.3%). Higher percentages of fibrosis were found in the basal anteroseptal ring which also contained the majority of class 1 (21/37) sites. In contrast, VT exit areas were clustered in the apico-septal and basal anterolateral segments. Class 1 VT related site segments contained more fibrosis (median 35.3% IQR [29.1-48.7] vs 22.1% [13.2-32.8], p<0.001) and were more inhomogeneous confluent (median 0.22 IQR [0.185-0.257] vs 0.27 [0.217-0.312], p=0.003) compared to segments without class 1 sites (fig 2). These findings were true for class 2 sites (p<0.004). ROC analysis determined a cut-off of >31% fibrosis for the identification of class I VT related sites (AUC: 0.75, sensitivity 72% specificity 72%). Of note, in segments with <31% fibrosis and class 1 related VT sites, there was a tendency towards inhomogeneous confluent fibrosis compared to those without VT related sites (median score 0.27 vs 0.29). CONCLUSION: In human hearts with NICM and VT, fibrosis clusters in the basal anteroseptal ring. The majority of VT related sites are located in segments with >31% fibrosis and a higher degree of inhomogeneity but they are also found in segments with less fibrosis, but with inhomogeneous confluent fibrosis. This data suggest that location, amount and inhomogeneous confluence of fibrosis contribute to arrhythmogeneity. [Figure: see text] [Figure: see text]