BACKGROUND: Atrial fibrillation (AF) recurrence is common after catheter ablation. Pulmonary vein (PV) isolation is the cornerstone of AF ablation, but PV remodeling has been associated with the risk of AF recurrence. We aimed to evaluate whether artificial intelligence–based morphological features of primary and secondary PV branches on computed tomography images are associated with AF recurrence post-ablation. METHODS: Two artificial intelligence models were trained for the segmentation of computed tomography images, enabling the isolation of PV branches. Patients from Cleveland Clinic (N=135) and Vanderbilt University (N=594) were combined and divided into 2 sets for training and cross-validation (D 1 , n=218) and internal testing (D 2 , n=511). An independent validation set (D 3 , N=80) was obtained from University Hospitals of Cleveland. We extracted 48 fractal-based and 12 shape-based radiomic features from primary and secondary PV branches of patients with AF recurrence (AF+) and without recurrence after catheter ablation of AF (AF-). To predict AFrecurrence, 3 Gradient Boosting classification models based on significant features from primary (M p ), secondary (M s ), and combined (M c ) PV branches were built. RESULTS: Features relating to primary PVs were found to be associated with AF recurrence. The M p classifier achieved area under the curve values of 0.73, 0.71, and 0.70 across the 3 datasets. AF+ cases exhibited greater surface complexity in their primary PV area, as evidenced by higher fractal dimension values compared with AF- cases. The M s classifier results revealed a weaker association with AF+, suggesting higher relevance to AF recurrence post-ablation from primary PV branch morphology. CONCLUSIONS: This largest multi-institutional study to date revealed associations between artificial intelligence–extracted morphological features of the primary PV branches with AF recurrence in 809 patients from 3 sites. Future work will focus on enhancing the predictive ability of the classifier by integrating clinical, structural, and morphological features, including left atrial appendage and left atrium–related characteristics.
Genome-wide association studies have identified a locus on chromosome 10q22, where many co-inherited single nucleotide polymorphisms (SNPs) are associated with atrial fibrillation (AF). This study seeks to identify the impact of this locus on gene expression at the transcript isoform level in human left atria and to gain insight into potential causal variants. Bulk RNA sequencing was analyzed to identify myozenin 1 (MYOZ1) and synaptopodin 2-like (SYNPO2L) transcript isoforms and the association of common SNPs in this region with transcript isoform expression levels. Chromatin marks were used to suggest candidate regulatory SNPs in this region. Protein amino acid changes were examined for predicted functional consequences. Transfection of MYOZ1 and two SYNPO2L isoforms were performed to localize their encoded proteins in cardiomyocytes derived from stem cells. We identified one MYOZ1 transcript isoform and four SYNPO2L transcript isoforms, two of which encode proteins, while the other two encode long noncoding RNAs (lncRNAs). The risk allele of the strongest AF susceptibility SNP on chromosome 10q22 is associated with decreased MYOZ1 expression and increased expression of the two SNYPO2L lncRNA isoforms. There are many SNPs co-inherited with the top AF-associated SNP due to linkage disequilibrium (LD), including rs11000728, which we propose as the MYOZ1 regulatory SNP, confirmed by reporter gene transfection. In addition, this LD block includes three missense SNPs in the SYNPO2L gene, with the minor protective haplotype predicted to be detrimental to protein function. MYOZ1 and both protein isoforms of SYNPO2L were localized to the sarcomere. This is a complex locus with the potential for several SNPs in a haplotype to alter AF susceptibility by opposing effects on MYOZ1 and SYNPO2L lncRNA expression, along with effects on SYNPO2L protein function.
Variable selection is critical in competing risks regression with high‐dimensional data. Although penalized variable selection methods and other machine learning‐based approaches have been developed, many of these methods often suffer from instability in practice. This paper proposes a novel method named Random Approximate Elastic Net (RAEN). Under the proportional subdistribution hazards model, RAEN provides a stable and generalizable solution to the large‐p‐small‐n variable selection problem for competing risks data. Our general framework allows the proposed algorithm to be applicable to other time‐to‐event regression models, including competing risks quantile regression and accelerated failure time models. We show that variable selection and parameter estimation improved markedly using the new computationally intensive algorithm through extensive simulations. A user‐friendly R package RAEN is developed for public use. We also apply our method to a cancer study to identify influential genes associated with the death or progression from bladder cancer.
1 October 2024 in Circulation: Genomic and Precision Medicine
BACKGROUND: Atrial fibrillation GWAS (genome-wide association studies) identified significant associations for rs1152591 and linked variants in the SYNE2 gene encoding Nesprin-2, which connects the nuclear membrane with the cytoskeleton. METHODS: Reporter gene vector transfection and CRISPR-Cas9 editing were used to identify the causal variant regulating the expression of SYNE2α1 . After SYNE2 knockdown or SYNE2α1 overexpression in human stem cell–derived cardiomyocytes, nuclear phenotypes were assessed by imaging and atomic force microscopy. Gene expression was assessed by RNAseq and gene set enrichment analysis. Fura-2 AM staining assessed calcium transients. Optical mapping assessed action potential duration and conduction velocity. RESULTS: The risk allele of rs1152591 had lower promoter and enhancer activity and was significantly associated with lower expression of the short SYNE2α1 isoform in human stem cell–derived cardiomyocytes, without an effect on the expression of the full-length SYNE2 mRNA. SYNE2α1 overexpression had dominant negative effects on the nucleus with its overexpression or SYNE2 knockdown leading to increased nuclear area and decreased nuclear stiffness. Gene expression results from SYNE2α1 overexpression demonstrated both concordant and nonconcordant effects with SYNE2 knockdown. SYNE2α1 overexpression had a gain of function on electrophysiology, leading to significantly faster calcium reuptake and decreased assessed action potential duration, while SYNE2 knockdown showed both shortened assessed action potential duration and decreased conduction velocity. CONCLUSIONS: rs1152591 was identified as a causal atrial fibrillation variant, with the risk allele decreasing SYNE2α1 expression. Downstream effects of SYNE2α1 overexpression include changes in nuclear stiffness and electrophysiology, which may contribute to the mechanism for the risk allele’s association with AF.