Late Gadolinium-enhanced Magnetic Resonance Imaging Research Articles

Biventricular modelling for investigating ventricular arrhythmias in silico

Abstract Background Ventricular arrhythmias have a significant incidence rate worldwide and may potentially lead to sudden cardiac death in severe cases. Despite therapeutic advancements, the intricate mechanisms dictating these arrhythmias remain elusive. Biventricular modelling and deep learning provide a way to investigate the complex electrophysiological substrates underlying cardiac arrhythmogenesis. Purpose This study aims to develop and validate a patient-specific biventricular computational modelling pipeline for simulating ventricular arrhythmias in silico. We provide a comprehensive platform for understanding the initiation, propagation, and treatment response of various ventricular arrhythmias. Methods To facilitate large-scale computational experiments and personalised model predictions within clinical timeframes, it is imperative to establish a rapid and reproducible modelling framework. First, we constructed biventricular volumetric models derived from late gadolinium-enhanced magnetic resonance imaging (LGE-MRI) data from the Cardiac Atlas Project (2018 Atria Segmentation Challenge), and then we calculate biventricular coordinates to facilitate data registration between measures, and fibre and fibrosis inclusion. We also apply the pipeline to cardiac magnetic resonance (CMR) data and electrocardiogram recordings from UK Biobank. Results We have successfully developed a pipeline for constructing biventricular volumetric models and efficiently processed biventricular coordinates (Figure). By systematically exploring interactions between ventricular regions, the model may identify critical factors contributing to arrhythmia susceptibility and guide personalised therapeutic strategies. Furthermore, the model may provide insights into the efficacy of antiarrhythmic interventions, ablation techniques, and device therapies, facilitating the development of precision medicine approaches. Conclusions We demonstrate a pipeline for constructing patient-specific biventricular computational models at scale from imaging data and apply this to LGE-MRI data and UK Biobank CMR data. These biventricular models enable investigation of the complex mechanisms of ventricular arrhythmias, providing novel insights for clinical decision-making and treatment strategies.

Residual post-infarct myocardial inflammation detected by 68Ga-DOTATATE PET/MRI predicts long-term adverse cardiac remodelling

Abstract Background Maladaptive inflammatory responses are associated with adverse cardiac remodelling after myocardial infarction (MI), but clinical methods for detecting residual myocardial inflammation post-MI are lacking. Purpose To investigate somatostatin receptor subtype 2 (SST2) positron emission tomography (PET)/magnetic resonance imaging (MRI) using 68Ga-DOTATATE after MI. Methods In a prospective observational cohort study, infarct-related myocardial inflammation detected by 68Ga-DOTATATE PET at 2wk and 3mo post-MI was compared with MRI characteristics including left ventricular (LV) volumetric changes at 1yr, as well circulating immune cell phenotyping and serum proteomic markers. Patients with a prior history of MI or heart failure were excluded. Histological SST2 expression was assessed in myocardial biopsies. Results 38 participants (mean age 60 [SD 9] years; 84% male) with recent MI were enrolled, of whom 58% had an ST-segment elevation MI. Mean peak Troponin I was 16,665 (range 408 to >25,000) ng/L and mean post-infarct LV ejection fraction was 51 (SD 9) %. 68Ga-DOTATATE maximum standardized uptake values (SUVmax) at 10 (SD 4) days accurately localised infarcted myocardial segments defined by late-gadolinium enhancement MRI (LGE; Fig. 1); were 14.2% higher in akinetic/hypokinetic vs. normo-kinetic segments (p<0.0001); and were correlated both with peak segmental strain (r=0.47, p=0.003) and segmental T1 values (r=0.51, p=0.02). Mean infarct SUVmax was 19.7% lower on repeat PET/MRI after a mean 102 (SD 22) days (p<0.0001), and this change was in keeping with parallel reductions in T2-weighted oedema signal on MRI and serum biomarkers of cardiac injury and systemic inflammation. Residual inflammation assessed by mean infarct SUVmax at 3mo was correlated with ΔLV indexed end-diastolic volume at 369 (SD 14) days, after adjustment for baseline infarct size and other clinical variables (p=0.038). The 3mo:2wk ratio of mean infarct SUVmax was an even stronger predictor of adverse myocardial remodelling (p=0.0025). Circulating SST2+ classical monocytes were correlated with infarct 68Ga-DOTATATE signal (r=0.47, p=0.02), and SST2 was co-expressed with CD68+ macrophages in inflamed myocardial specimens of patients with recent MI and ischaemic heart failure. After statistical feature selection and multiple comparisons adjustments, among the blood markers associated with 3mo:2wk infarct SUVmax was CCL25, a chemokine involved in macrophage recruitment which has been implicated in adverse cardiac remodelling. Conclusions Here, in the first prospective clinical study of serial SST2 PET/MRI after MI, we show that 68Ga-DOTATATE can identify and track resolving infarct-related myocardial inflammation. Moreover, residual inflammation detected by 68Ga-DOTATATE is associated with long-term adverse myocardial remodelling and could provide a potential future imaging biomarker to help stratify high-risk patients for emerging immunomodulatory therapies.Figure 1.68Ga-DOTATATE PET-MRI after MI

Are drivers recurring or ephemeral? observations from serial mapping of persistent atrial fibrillation.

Rotational re-entries and ectopic foci, or 'drivers', are proposed mechanisms for persistent atrial fibrillation (persAF), but driver-based interventions have had mixed success in clinical trials. Selective targeting of drivers with multi-month stability may improve these interventions, but no prior work has investigated whether drivers can be stable on such a long timescale. We hypothesized that drivers could recur even several months after initial observation. We performed serial electrophysiology studies on paced canines (n = 18, 27-35 kg) at 1-, 3-, and 6 months post-initiation of continual persAF. Using a high-density 64-electrode catheter, we captured endocardial electrograms in the left atrium (LA) and right atrium (RA) to determine the presence of drivers at each major anatomical site. We defined drivers that were repeatedly observed across consecutive studies to be recurrent. The mean probability that any driver would recur was 66% (LA: 73%, RA: 41%). We also found evidence of 'multi-recurring' drivers, i.e. those seen in all three studies. Multi-recurring drivers constituted 53% of initially observed drivers with at least one found in 92% of animals, and we found more multi-recurring drivers per animal than predicted by random chance (2.6 ± 1.5 vs. 1.2 ± 1.1, P < 0.001). Driver sites showed more enhancement than non-drivers during late gadolinium enhancement-magnetic resonance imaging (P = 0.04), but we observed no relationship between enhancement and driver recurrence type. We observed recurring drivers over a 6-month period at fixed locations, confirming our hypothesis. We also found drivers to be associated with fibrosis, implying a structural basis.

Unipolar voltage electroanatomic mapping detects structural atrial remodeling identified by LGE-MRI

BackgroundIn atrial fibrillation (AF) management, understanding left atrial (LA) substrate is crucial. While both electroanatomic mapping (EAM) and late gadolinium enhancement magnetic resonance imaging (LGE-MRI) are accepted methods for assessing the atrial substrate and are associated with ablation outcome, recent findings have highlighted discrepancies between low-voltage areas (LVAs) in EAM and LGE areas. ObjectiveThe purpose of this study was to explore the relationship between LGE regions and unipolar and bipolar LVAs using multipolar high-density mapping. MethodsTwenty patients scheduled for AF ablation underwent preablation LGE-MRI. LA segmentation was conducted using a deep learning approach, which subsequently generated a 3-dimensional mesh integrating the LGE data. High-density EAM was performed in sinus rhythm for each patient. The electroanatomic map and LGE-MRI mesh were coregistered. LVAs were defined using cutoffs of 0.5 mV for bipolar voltage and 2.5 mV for unipolar voltage. The correspondence between LGE areas and LVAs in the LA was analyzed using confusion matrices and performance metrics. ResultsA considerable 87.3% of LGE regions overlapped with unipolar LVAs, compared with only 16.2% overlap observed with bipolar LVAs. Across all performance metrics, unipolar LVAs outperformed bipolar LVAs in identifying LGE areas (precision: 78.6% vs 61.1%; sensitivity: 87.3% vs 16.2%; F1 score: 81.3% vs 26.0%; accuracy: 74.0% vs 35.3%). ConclusionOur findings demonstrate that unipolar LVAs strongly correlate with LGE regions. These findings support the integration of unipolar mapping alongside bipolar mapping into clinical practice. This would offer a nuanced approach to diagnose and manage AF by revealing critical insights into the complex architecture of the atrial substrate.

Adaptive dynamic inference for few-shot left atrium segmentation

Accurate segmentation of the left atrium (LA) from late gadolinium-enhanced cardiac magnetic resonance (LGE CMR) images is crucial for aiding the treatment of patients with atrial fibrillation. Few-shot learning holds significant potential for achieving accurate LA segmentation with low demand on high-cost labeled LGE CMR data and fast generalization across different centers. However, accurate LA segmentation with few-shot learning is a challenging task due to the low-intensity contrast between the LA and other neighboring organs in LGE CMR images. To address this issue, we propose an Adaptive Dynamic Inference Network (ADINet) that explicitly models the differences between the foreground and background. Specifically, ADINet leverages dynamic collaborative inference (DCI) and dynamic reverse inference (DRI) to adaptively allocate semantic-aware and spatial-specific convolution weights and indication information. These allocations are conditioned on the support foreground and background knowledge, utilizing pixel-wise correlations, for different spatial positions of query images. The convolution weights adapt to different visual patterns based on spatial positions, enabling effective encoding of differences between foreground and background regions. Meanwhile, the indication information adapts to the background visual pattern to reversely decode foreground LA regions, leveraging their spatial complementarity. To promote the learning of ADINet, we propose hierarchical supervision, which enforces spatial consistency and differences between the background and foreground regions through pixel-wise semantic supervision and pixel-pixel correlation supervision. We demonstrated the performance of ADINet on three LGE CMR datasets from different centers. Compared to state-of-the-art methods with ten available samples, ADINet yielded better segmentation performance in terms of four metrics.

History of stroke as a predictor of high left atrial fibrosis in patients with persistent atrial fibrillation-insight from the DECAAF II randomized trial.

There is a strong relationship between left atrial (LA) remodeling and ischemic stroke (IS) risk in atrial fibrillation (AF) patients. The Efficacy of Delayed Enhancement MRI-Guided Ablation vs. Conventional Catheter Ablation of Atrial Fibrillation (DECAAF-II) is the biggest MRI-based, randomized, multicenter clinical trial performed on persistent AF patients. The aim of this study is to evaluate the relationship between history of stroke and atrial fibrosis in the DECAAF II population. Persistent AF patients who underwent Late Gadolinium Enhancement Magnetic Resonance Imaging (LGE-MRI) were included in the study and divided into two different groups: those with a history of stroke and those without. Propensity score matching was performed to adjust for covariates. Atrial fibrosis was compared in both groups. Then, patients were divided into different fibrosis groups, using three different cut-offs of baseline atrial fibrosis: ≥ 15%, ≥ 20%, and ≥ 25%. Univariate logistic regression and adjusted multivariate analysis were performed to assess the effect of clinical characteristics and risk factors on baseline fibrosis. Eight-hundred forty-three patients were recruited in DECAAF II, of whom 70 (8.3%) had a history of stroke. Patients with history of stroke had a higher prevalence of hypertension (p = 0.043), diabetes (p = 0.014), and hyperlipidemia (p = 0.001). Seventy patients with no history of strokes were matched with patients with history of stroke to adjust for covariates using propensity score analysis. Patients in the stroke group had a significantly higher level of fibrosis than those without (20.2% vs. 8.1%, p = 0.017). Increased age was a significant predictor of all three baseline fibrosis classes (≥ 15%, ≥ 20%, and ≥ 25%). Additionally, history of stroke was found to be a predictor of baseline fibrosis ≥ 25% even after adjusting for other clinical characteristics and risk factors (OR = 1.98 [1.14-3.43], p = 0.01). Left atrial fibrosis level greater than 25% correlates with the history of previous stroke episodes in patients with persistent atrial fibrillation.

Prognostic role of late gadolinium-enhanced MRI in confirmed and suspected cardiac sarcoidosis: meta-analysis.

The prognostic implications of late gadolinium-enhanced (LGE) magnetic resonance imaging (MRI) in the context of cardiac sarcoidosis (CS) have attracted considerable attention. Nevertheless, a subset of studies has undistinguished confirmed and suspected CS cases, thereby engendering interpretative ambiguities. In this meta-analysis, we evaluated the differences in cardiac MRI findings and their prognostic utility between confirmed and suspected CS. A literature search was conducted using PubMed, Web of Science, and Cochrane libraries to compare the findings of cardiac MRI and its prognostic value in CS and suspected CS. A meta-analysis was performed to compare the prevalence of LGE MRI, odds ratios, and hazard ratios for predicting cardiac events in both groups. A total of 21 studies encompassing 24 different populations were included in the meta-analysis (CS: 393 cases, suspected CS: 2151 cases). CS had a higher frequency of LGE of the left ventricle (87.2% vs. 36.4%, p < 0.0001) and right ventricle (62.1% vs. 23.8%, p = 0.04) than suspected CS. In patients with suspected CS, the presence of left ventricular LGE was associated with higher all-cause mortality [odds ratio: 5.70 (95%CI: 2.51-12.93), p < 0.0001, I2 = 8%, p for heterogeneity = 0.37] and ventricular arrhythmia [odds ratio: 15.51 (95%CI: 5.65-42.55), p < 0.0001, I2 = 0, p for heterogeneity = 0.94]. In contrast, in CS, not the presence but extent of left ventricular LGE was a significant predictor of outcome (hazard ratio = 1.83 per 10% increase of %LGE (95%CI: 1.43-2.34, p < 0.001, I2 = 15, p for heterogeneity = 0.31). The presence of left ventricular LGE was a strong prognostic factor in suspected sarcoidosis. However, the extremely high prevalence of left ventricular LGE in confirmed CS suggests that the quantitative assessment of LGE is useful for prognostic estimation.

Advancements in the Study of Dilated Cardiomyopathy Biomarkers

Dilated cardiomyopathy (DCM) is the leading cause of sudden cardiac death (SCD) and heart failure (HF) in children and adults worldwide, and it is also the main indication of heart transplantation. DCM remains a significant cause of heart transplantation despite recent improvements in treatment, which is related to the enormous financial strain placed on the world's healthcare system. Therefore, determining the prognosis of individuals with DCM is crucial for providing them with a customized course of treatment. Genetic testing, microvolt T-wave alternation, and late gadolinium-enhanced cardiac magnetic resonance imaging have developed into potent methods for predicting the occurrence of sudden cardiac death and optimizing patient selection. Despite the availability of numerous novel diagnostic techniques, further tests are still required to enhance or partially substitute current methods of risk categorization. As a result, biomarkers are robust and straightforward tools that can identify people at risk for unfavorable cardiovascular events. Thus, this article aims to review the state of the research on biomarkers for DCM.

Antiarrhythmic and Anti-Inflammatory Effects of Sacubitril/Valsartan on Post-Myocardial Infarction Scar.

Sacubitril/valsartan (Sac/Val) is superior to angiotensin-converting enzyme inhibitors in reducing the risk of heart failure hospitalization and cardiovascular death, but its mechanistic data on myocardial scar after myocardial infarction (MI) are lacking. The objective of this work was to assess the effects of Sac/Val on inflammation, fibrosis, electrophysiological properties, and ventricular tachycardia inducibility in post-MI scar remodeling in swine. After MI, 22 pigs were randomized to receive β-blocker (BB; control, n=8) or BB+Sac/Val (Sac/Val, n=9). The systemic immune response was monitored. Cardiac magnetic resonance data were acquired at 2-day and 29-day post MI to assess ventricular remodeling. Programmed electrical stimulation and high-density mapping were performed at 30-day post MI to assess ventricular tachycardia inducibility. Myocardial samples were collected for histological analysis. Compared with BB, BB+Sac/Val reduced acute circulating leukocytes (P=0.009) and interleukin-12 levels (P=0.024) at 2-day post MI, decreased C-C chemokine receptor type 2 expression in monocytes (P=0.047) at 15-day post MI, and reduced scar mass (P=0.046) and border zone mass (P=0.043). It also lowered the number and mass of border zone corridors (P=0.009 and P=0.026, respectively), scar collagen I content (P=0.049), and collagen I/III ratio (P=0.040). Sac/Val reduced ventricular tachycardia inducibility (P=0.034) and the number of deceleration zones (P=0.016). After MI, compared with BB, BB+Sac/Val was associated with reduced acute systemic inflammatory markers, reduced total scar and border zone mass on late gadolinium-enhanced magnetic resonance imaging, and lower ventricular tachycardia inducibility.

RAS Dataset: A 3D Cardiac LGE-MRI Dataset for Segmentation of Right Atrial Cavity

The current challenge in effectively treating atrial fibrillation (AF) stems from a limited understanding of the intricate structure of the human atria. The objective and quantitative interpretation of the right atrium (RA) in late gadolinium-enhanced magnetic resonance imaging (LGE-MRI) scans relies heavily on its precise segmentation. Leveraging the potential of artificial intelligence (AI) for RA segmentation presents a promising solution. However, the successful implementation of AI in this context necessitates access to a substantial volume of annotated LGE-MRI images for model training. In this paper, we present a comprehensive 3D cardiac dataset comprising 50 high-resolution LGE-MRI scans, each meticulously annotated at the pixel level. The annotation process underwent rigorous standardization through crowdsourcing among a panel of medical experts, ensuring the accuracy and consistency of the annotations. Our dataset represents a significant contribution to the field, providing a valuable resource for advancing RA segmentation methods.

The benefit of a double ventricular extra stimuli protocol to improve functional substrate mapping during ablation of ventricular tachycardia. - The S3-VT-FREE project

Abstract Introduction Ablation of ventricular tachycardia (VT) based on substrate-based mapping has been proposed to itemize areas of interest identifying areas of low-voltage, late potentials (LPs), and deceleration zones (DZs). However, partial delineation of targets for ablation is likely responsible for a suboptimal long-term success rate. The abnormal electrograms (EGMs), featuring low-voltage near-field components may be hidden within the high-amplitude far-field signal. We hypothesized that the acquisition of functional mapping after double ventricular extrastimuli (S3 protocol) would uncover relevant areas of functional substrate that could be missed during sinus or S1 map. We aim to evaluate the potential benefit and the feasibility of functional substrate mapping using an S3 protocol for VT ablation and to assess its colocalization with detected corridors using late gadolinium enhancement magnetic resonance imaging (LGE-MRI). Methods We included consecutively 29 patients with structural cardiomypathy who underwent a VT ablation. The S3 protocol included: (1) assessment of the ventricular effective refractory period (ERP), (2) pacing protocol from the apex of the right ventricle with S1 x 6 (cycle length: 600ms) followed by S2 (ERP+30ms) and S3 (ERP+50ms). Electroanatomical mapping was acquired with the EnsiteX system (Abbott, IL) using HD Grid catheter and omnipolar signals. ILAMs were used to identify deceleration zones and areas of late potentials were identified for all maps. A non-invasive substrate assessment was done with LGE-MRI and post-processing with a commercially available system (ADAS3D, Spain) with automated identification of corridors. Results The S3 protocol was completed in 25 of the 29 procedures (86.2%). The repeated induction of ventricular arrhythmias did not allow for full protocol of the S3 protocol in 4 patients (13.8%). The functional substrate identified during the S3 activation mapping was significantly more extensive than the one identified using S1 and S2 activation mappings (p&amp;lt;0.05) (Table). The S3 protocol unmasked a higher number of DZs (p&amp;lt;0.01), a wider area of LPs (p&amp;lt;0.01), and a longer duration of the DZs (p=0.01). All DZ maps were compared with preprocedural cardiac magnetic resonance. The number of DZs matching with conduction corridors identified by LGE-MRI was higher in the activation mapping of S3 compared with the S1 and S2 ones (p=0.02). After VT ablation based on an S3 protocol, 88% of patients have been VT free during a median follow-up of 9 months. Conclusions This novel substrate-based protocol to assess the ventricular functional substrate is feasible in the majority of patients (86.2%) and allows for a better identification of targets for ablation and could facilitate VT ablation, and improve the prognosis of the patients after ablation.

Abstract 18093: History of Stroke as a Predictor of High Left Atrial Fibrosis in Patients With Persistent Atrial Fibrillation

Introduction: The relationship between left atrial (LA) remodeling and ischemic stroke (IS) risk in patients with atrial fibrillation (AF) is well-established. The DECAAF II trial is the largest MRI-based, multicenter, randomized study conducted on persistent AF patients. Hypothesis: Our research investigates whether a history of stroke correlates with the extent of atrial fibrosis, as determined by late gadolinium enhancement magnetic resonance imaging (LGE-MRI), among the DECAAF II trial population. Methods: Patients with persistent AF who underwent LGE-MRI were enrolled and categorized into two cohorts: those with a history of stroke and those without. Covariates were adjusted using propensity score matching, and atrial fibrosis was compared between groups. Participants were then subdivided according to three baseline atrial fibrosis cut-off values: ≥ 15%, ≥ 20%, ≥ 25%. Univariate logistic regression and adjusted multivariate analyses were applied to evaluate the impact of clinical characteristics and risk factors on baseline fibrosis. Results: From the 843 patients involved in the DECAAF II trial, 70 (8.3%) had a stroke history. Those with stroke history had higher rates of hypertension, diabetes, and hyperlipidemia. Seventy patients without a stroke history were matched to stroke patients using propensity score analysis, revealing a significantly higher fibrosis in the stroke group (20.2% vs. 8.1%, p = 0.017). Increasing age significantly predicted all three baseline fibrosis categories (≥15%, ≥20%, ≥25%). Notably, stroke history was as a predictor of baseline fibrosis ≥25%, even after adjusting for other risk factors and clinical characteristics (OR = 1.98 [1.14-3.43], p = 0.01). Conclusions: There is a significant correlation between high left atrial fibrosis level and an increased risk of ischemic stroke in patients with persistent atrial fibrillation. This suggests that atrial fibrosis may provide insights into stroke risk classification.

Abstract 19134: Association Between CHA 2 DS 2 -VASc and LA Volume Index: Insights From DECAAF II

Introduction: LA Volume is strongly associated with atrial fibrillation (AF) and underlying atrial myopathy. The CHA 2 DS 2 -VASc score predicts stroke risk in patients with AF. The aim of this study is to investigate the relationship between the two. Methods: In this subanalysis of the DECAAF II trial, patients with persistent AF undergoing first-time catheter ablation underwent a baseline late gadolinium enhancement magnetic resonance imaging (LGE-MRI) within one month prior to ablation. LA Volume Index (LAVI) was defined as LA Volume on the MRI prior to ablation divided by body surface area ( mL/m2 ). The CHA2 DS2 -VASc score was also calculated at baseline prior to ablation. Analysis of Variance (ANOVA) test and linear regression were used to study the association between the two. Results: 815 patients were included, (78.9% males, mean age 62.7 ± 9.1 years) (table). The mean LAVI was 62.2 ± 18.9 mL/m2, and mean CHA2 DS2 -VASc score was 1.7 ± 1.27. The ANOVA test showed significant differences between group (F = 7.71, p &lt;0.001). For every point increase in CHA2 DS2 -VASc score, LAVI increases by 3 (LAVI = 2.99 * CHA2 DS2 -VASc + 57; p &lt;0.001) (figure). Conclusions: Our data suggests that structural remodeling in patients with AF is likely driven by multiple factors. There was a clear association between CHA2 DS2 -VASc score and higher LAVI. Serial analysis of LA Volume at different time points is needed to further understand the impact of individual comorbidities on LA volume dilation.

Prognostic role of atrial fibrillation in patients with non-compaction cardiomyopathy

Aim. To evaluate the prognostic role of atrial fibrillation (AF) as a predictor of adverse events and outcomes in a cohort of patients with non-compaction cardiomyopathy (NCCM).Material and methods. We examined 216 patients with NCCM (140 men and 76 women, median age, 39 (30; 50) years). In addition to traditional clinical methods, all patients underwent late gadolinium-enhanced cardiac magnetic resonance imaging (MRI). The endpoints of the study included progression of NYHA class III heart failure (HF) with the need for hospitalization, ventricular tachyarrhythmias, and thromboembolic events (TEEs).Results. There were 54 out of 216 (23,6%) patients with AF, of which 18 had paroxysmal AF, 16 — persistent AF, and 20 — permanent AF. During the follow-up period (median follow-up, 36 (6; 72) months), 98 out of 216 (45,4%) patients with NCCM had adverse events and outcomes as follows: 16 (7,4%) had ventricular tachyarrhythmias, of which 12 (5,6%) — sudden cardiac death with successful resuscitation and implantation of an implantable cardioverter-defibrillator; 62 (28,7%) patients had NYHA III-IV class HF progression; 20 (9,3%) patients had TEEs. The rate of adverse cardiac events was significantly higher in patients with AF (74,1% vs 35,8%, χ2=23,93, p&lt;0,001) compared with patients without AF, including the incidence of TEEs (20,4% vs 5,6%, χ2=10,58, p=0,002) and HF progression to class III (46,3% vs 22,8%, χ2=10,9, p=0,002).Multivariate analysis showed that the following most significant predictors of HF progression risk: left ventricular ejection fraction (LVEF) &lt;50% according to cardiac MRI (hazard ratio (HR), 95,8; 95% confidence interval (CI), 10,2 -898,6; p=0,0001), presence of AF (HR, 8,2; 95% CI, 2,2-31,3; p=0,0022) and left atrial volume index (LAVI) &gt;43 ml/m2 (HR, 5,2; 95% CI, 2,1-12,8; p=0,0004); predictors of TEE risk were the presence of AF (HR, 6,5; 95% CI, 2,0-20,8; p=0,0020) and LAVI &gt;43 ml/m2 (HR, 6,0; 95% CI, 1,8-19,7; p=0,036). No association of AF with ventricular tachyarrhythmias was found in the study cohort of patients with NCCM. Predictors of ventricular tachyarrhythmias were LVEF &lt;50% (HR, 4,5; 95% CI, 2,950,4; p=0,0241) and the presence of non-sustained ventricular tachycardia (HR, 3,5; 95% CI, 1,3-9,3 p=0,0139).Conclusion. The present study shows that, along with the traditional predictor of adverse events in patients with NCCM (LVEF &lt;50%), the identified additional predictors (AF and LAVI &gt;43 ml/m2) can be used to identify patients at high risk of complicated NCCM for the timely prevention and treatment.

Differences in atrial substrate localization using late gadolinium enhancement-magnetic resonance imaging, electrogram voltage, and conduction velocity: a cohort study using a consistent anatomical reference frame in patients with persistent atrial fibrillation.

Electro-anatomical voltage, conduction velocity (CV) mapping, and late gadolinium enhancement (LGE) magnetic resonance imaging (MRI) have been correlated with atrial cardiomyopathy (ACM). However, the comparability between these modalities remains unclear. This study aims to (i) compare pathological substrate extent and location between current modalities, (ii) establish spatial histograms in a cohort, (iii) develop a new estimated optimized image intensity threshold (EOIIT) for LGE-MRI identifying patients with ACM, (iv) predict rhythm outcome after pulmonary vein isolation (PVI) for persistent atrial fibrillation (AF). Thirty-six ablation-naive persistent AF patients underwent LGE-MRI and high-definition electro-anatomical mapping in sinus rhythm. Late gadolinium enhancement areas were classified using the UTAH, image intensity ratio (IIR >1.20), and new EOIIT method for comparison to low-voltage substrate (LVS) and slow conduction areas <0.2 m/s. Receiver operating characteristic analysis was used to determine LGE thresholds optimally matching LVS. Atrial cardiomyopathy was defined as LVS extent ≥5% of the left atrium (LA) surface at <0.5 mV. The degree and distribution of detected pathological substrate (percentage of individual LA surface are) varied significantly (P < 0.001) across the mapping modalities: 10% (interquartile range 0-14%) of the LA displayed LVS <0.5 mV vs. 7% (0-12%) slow conduction areas <0.2 m/s vs. 15% (8-23%) LGE with the UTAH method vs. 13% (2-23%) using IIR >1.20, with most discrepancies on the posterior LA. Optimized image intensity thresholds and each patient's mean blood pool intensity correlated linearly (R2 = 0.89, P < 0.001). Concordance between LGE-MRI-based and LVS-based ACM diagnosis improved with the novel EOIIT applied at the anterior LA [83% sensitivity, 79% specificity, area under the curve (AUC): 0.89] in comparison to the UTAH method (67% sensitivity, 75% specificity, AUC: 0.81) and IIR >1.20 (75% sensitivity, 62% specificity, AUC: 0.67). Discordances in detected pathological substrate exist between LVS, CV, and LGE-MRI in the LA, irrespective of the LGE detection method. The new EOIIT method improves concordance of LGE-MRI-based ACM diagnosis with LVS in ablation-naive AF patients but discrepancy remains particularly on the posterior wall. All methods may enable the prediction of rhythm outcomes after PVI in patients with persistent AF.

CE-452775-2 MARS-HCM: MULTI-MODAL DEEP LEARNING METHOD FOR VENTRICULAR ARRHYTHMIA (VA) RISK STRATIFICATION IN HYPERTROPHIC CARDIOMYOPATHY (HCM) PATIENTS

Current clinical guidelines by the AHA/ACC and ESC remain inadequate for accurately identifying HCM patients at risk for sudden cardiac death (SCD) by VA. Although risk factors for SCD in HCM patients have been established previously, including demographics, lab tests, late gadolinium-enhanced magnetic resonance imaging (LGE-MRI) features, it is challenging to integrate these distinct data types in a predictive model. Whether deep learning can integrate clinical covariates and LGE-MRI images in a multi-modal predictor that stratifies VA risk in HCM patients better than current clinical guidelines has not been explored. To develop a multi-modal deep learning approach that stratifies the risk of VA (sustained VT or VF) in HCM patients using raw LGE-MRI images and clinical covariates. In a single-center retrospective study, LGE-MRI images and electronic health record (EHR) data (demographics, clinical risk factors, echocardiogram features, LGE-MRI features) were acquired from 819 HCM patients. We developed a deep learning model with two input branches for images and EHR. In the image branch, the left ventricle (LV) was automatically segmented from LGE-MRI images, and the LV with raw image intensities was fed into a 3D ResNet. We adopted a feedforward network for the EHR branch to learn patterns from clinical covariates. In the output, the model was trained to synergistically fuse knowledge from both branches for VA prediction. In a 5-fold cross-validation, the multi-modal model using both LGE-MRI and EHR identified HCM patients with VA with an area under the ROC curve (AUC) of 0.83 (Figure), outperforming single-modal models using either LGE-MRI (AUC=0.80) or EHR (AUC=0.79). As a comparison, AHA/ACC guidelines had an AUC of 0.65 with a sensitivity of 0.81 and specificity of 0.44; ESC guidelines had an AUC of 0.69 with a sensitivity of 0.94 and specificity of 0.22. While preserving the same level of sensitivity as AHA/ACC guideline, the multi-modal model achieved higher specificity, 0.66. The VA risk score by the multi-modal model also showed a higher correlation coefficient (r=0.16) with other future adverse events (ICD shock, death, etc.) than risk scores by AHA/ACC (r=0.06) and ESC (r=0.01); see Figure. Our multi-modal deep learning significantly improved the prediction of VA in HCM patients compared with single-modal models and current clinical guidelines. Our model has the potential to improve the precision of VA risk stratification in HCM patients.

Prediction of Recurrence of Atrial Fibrillation Post-ablation Based on Atrial Fibrosis Seen on Late Gadolinium Enhancement MRI: A Metaanalysis.

This meta-analysis aims to investigate the recurrence of atrial fibrillation (AF) post-ablation based on the various stages of fibrosis seen in the late gadolinium enhancement magnetic resonance imaging (LGE-MRI). Electronic databases were searched using specific terms and identified nine studies that met the inclusion criteria. A total of 1,787 patients underwent LGE-MRI to assess atrial fibrosis before catheter ablation for AF. We performed three analyses: first, we compared stage IV versus stage I (reference group). The second set examined the combined stages III and IV versus stages I and II (reference group). The third set compared stage IV versus combined stages I, II, and III. The metanalysis relied on a random-effects model to pool the odds ratios (OR) and 95% confidence intervals (CI) using the DerSimonian and Laird method. The data was analyzed using StatsDirect software in England. The study showed a higher rate of AF recurrence after ablation in stage IV atrial fibrosis than in stage I (OR, 9.54; 95% CI, 3.81 to 28.89; P<00001). Also, in patients with combined stages III & IV of atrial fibrosis, AF recurrence was significantly higher after ablation than in stages I & II groups (OR, 2.37; 95% CI, 1.61 to 3.50; P<00001). Similarly, compared to combined stages I, II, and III, patients with stage IV have higher odds of recurrence post-ablation (OR, 4.24; 95% CI, 2.39- 7.52, P < 0.001). This metanalysis demonstrates the strong association between left atrial fibrosis in LGE-MRI and AF post-ablation recurrence. The finding of this study will further assist clinicians in predicting the recurrence rate of AF based on the amount of fibrosis and tailor therapeutic decisions for further management.

PO-01-099 EARLY REMODELING OF THE LEFT ATRIUM FOLLOWING CATHETER ABLATION OF PERSISTENT ATRIAL FIBRILLATION: THE IMPACT OF SUBSTRATE MODIFICATION IN DECAAF II

Left Atrial (LA) enlargement is prevalent amongst atrial fibrillation (AF) patients and constitutes an important marker of underlying atrial myopathy. Several studies have described reduction in LA volume following catheter ablation (CA) of AF, however, none have investigated differences in remodeling related to additional ablation outside of the pulmonary veins. We sought to compare early remodeling of the LA between patients who received pulmonary vein isolation (PVI) only, and those who received additional fibrosis-guided extra-PV ablation. This study was a retrospective analysis of the DECAAF II randomized controlled trial. Patients with persistent AF (perAF) were randomized to receive PVI only or PVI + fibrosis-guided ablation. Late Gadolinium Enhancement Magnetic Resonance Imaging (LGE-MRI) was performed before and three months after CA to measure LA volume. Only patients in sinus rhythm at both MRIs are included. Patients were followed up with single-lead ECG devices for 12-18 months. Arrhythmic burden was calculated as the percentage of ECGs exhibiting AF. Univariate and multivariate regression were used to identify predictors of LA volume change. This analysis included 206 patients who met our inclusion criteria. The mean LA volume before ablation was 114.6 ± 36.9 mm3, and the mean volume post-ablation was 101.2 ± 32.3 mm3, with a mean reduction of 13.5 ± 22.3 mm3 (9.92% reduction, p <0.001). Patients randomized to receive MRI-guided fibrosis ablation had more LA volume reduction than patients who were randomized to PVI only (17.03 ± 25.04 mm3 vs 10.17 ± 18.89 mm3, p=0.03). This association persisted with multivariate analysis where pre-ablation left ventricular ejection fraction (LVEF) and randomization to receive substrate modification were the only predictors of LA volume reduction (table). LA volume reduction was associated with greater improvement in LVEF, regardless of treatment randomization (r = 0.242, p = 0.01). LA volume reduction did not predict arrhythmia recurrence when considered as a continuous variable. We noted more pronounced LA volume reductions in patients randomized to receive PVI + fibrosis-guided ablation when compared to those randomized to receive PVI only. We hypothesize worsened left atrial relaxation and diastolic dysfunction as an etiology for this difference, but further study is needed to better characterize the important structural impacts of extra-PV ablation.

Left atrial volume affects the correlation of voltage map with magnetic resonance imaging.

The low-voltage area detected by electroanatomic mapping (EAM) is a surrogate marker of left atrial fibrosis. However, the correlation between the EAM and late gadolinium enhancement magnetic resonance imaging (LGE-MRI) has been inconsistent among studies. This study aimed to investigate how LA size affects the correlation between EAM and LGE-MRI. High-density EAMs of the LA during sinus rhythm were collected in 22 patients undergoing AF ablation. The EAMs were co-registered with pre-ablation LGE-MRI models. Voltages in the areas with and without LGE were recorded. Left atrial volume index (LAVI) was calculated from MRI, and LAVI > 62ml/m2 was defined as significant LA enlargement (LAE). Atrial bipolar voltage negatively correlates with the left atrial volume index. The median voltages in areas without LGE were 1.1mV vs 2.0mV in patients with vs without significant LAE (p = 0.002). In areas of LGE, median voltages were 0.4mV vs 0.8mV in patients with vs without significant LAE (p = 0.02). A voltage threshold of 1.7mV predicted atrial LGE in patients with normal or mildly enlarged LA (sensitivity and specificity of 74% and 59%, respectively). In contrast, areas of voltage less than 0.75mV correlated with LGE in patients with significant LA enlargement (sensitivity 68% and specificity 66%). LAVI affects left atrial bipolar voltage, and the correlation between low-voltage areas and LGE-MRI. Distinct voltage thresholds according to the LAVI value might be considered to identify atrial scar by EAM.

Cryoballoon-Induced Circumferential Pulmonary Vein Fibrosis, Assessed by Late Gadolinium-Enhancement Cardiac Magnetic Resonance Imaging, and Its Correlation with Clinical Atrial Fibrillation Recurrence.

Prior studies evaluating post-atrial fibrillation (AF) ablation pulmonary vein (PV) ostial gaps via magnetic resonance imaging (MRI) have shown circumferential PV fibrosis in a minority of patients, and their correlation with AF recurrence was weak. These studies were mostly based on radio-frequency AF ablations. We aimed to assess cryoballoon ablation-induced PV fibrosis via MRI and its correlation with AF recurrence. This was a prospective study of consecutive patients with symptomatic AF who underwent pre- and post-ablation MRI to assess baseline and ablation-induced fibrosis, respectively. Post-ablation PV gaps were assessed by new semi-quantitative visual analysis assisted by computerized ADAS analysis. AF recurrence monitored via multiple ECGs and event monitoring at 6 and 12 months post ablation. Nineteen patients with 80 PVs were included, age 56 ± 11, with paroxysmal and persistent AF in 17/19 and 2/19 patients, respectively. Baseline MRI showed minimal LA fibrosis. All patients underwent successful cryoballoon PV electrical isolation. Post-ablation MRI revealed circumferential PV fibrosis among 63/80 (78.8%) PVs and partial fibrosis with major gaps among 17/80 (21.2%) PVs. AF recurred within one year in 5/9 (55.5%) patients with partial PV fibrosis, while no AF recurred among the 10 patients in whom all PVs had circumferential fibrosis (p < 0.01). Similarly, there were significantly more PVs without circumferential fibrosis (due to major gaps) among patients with AF recurrence as compared with patients without AF recurrence (42.9% vs. 13.5%; p < 0.01). Cryoballoon AF ablation results in circumferential PV fibrosis in the majority of PVs, as assessed by a new clinically relevant MRI-LGE analysis. Significant correlation was found between major PV gaps on post-ablation MRI and AF recurrence, suggesting that MRI might have the ability to predict AF recurrence.