Extent and Features of Late Gadolinium Enhancement Stratify Arrhythmic Risk in Patients With Biopsy-Proven Sarcoidosis
晚期钆增强程度与特征对活检确诊结节病患者心律失常风险的分层作用

We identified patients with biopsy-proven sarcoidosis who were investigated at the Royal Brompton Hospital from 2004 to 2023. Only patients with biopsy-proven sarcoidosis and good-quality CMR images were included. Patients with significant obstructive coronary artery disease, significant valvular heart disease, or suspicion of an alternative diagnosis of cardiomyopathy were excluded. Significant coronary artery disease was defined as previous percutaneous coronary intervention, coronary artery bypass graft, or known obstructive coronary artery disease. Significant valvular artery disease was defined as severe valvular stenosis or regurgitation or previous valve replacement or repair. An alternative diagnosis of cardiomyopathy was established by a multidisciplinary team based on clinical and imaging findings as well as genetic testing when appropriate.
我们筛选了 2004 年至 2023 年间在皇家布朗普顿医院就诊且经活检确诊的结节病患者。仅纳入经活检确诊且具有高质量心脏磁共振成像（CMR）图像的患者。排除存在明显阻塞性冠状动脉疾病、严重心脏瓣膜病或疑似其他类型心肌病的病例。显著冠状动脉疾病定义为既往接受过经皮冠状动脉介入治疗、冠状动脉旁路移植术或已知的阻塞性冠状动脉疾病。严重心脏瓣膜病定义为重度瓣膜狭窄/反流或既往接受过瓣膜置换/修复术。其他类型心肌病的诊断由多学科团队根据临床表现、影像学检查结果及必要时基因检测结果综合判定。

Patients’ medical records were searched for all clinical information. Medical imaging data sets of CMR studies were retrieved for independent review and analysis. Follow-up event data were collected from electronic hospital records. Time to first clinical event was defined as the time from the date of the index CMR scan to the date of the first event. Follow-up duration was calculated from the date of the index CMR scan until death or last clinical contact with the patient.
通过检索患者病历获取所有临床信息。心脏磁共振检查的医学影像数据集被调取用于独立审查和分析。随访事件数据从医院电子病历系统中收集。首次临床事件发生时间定义为从基线心脏磁共振扫描日期到首次事件发生日期的间隔。随访时长计算自基线心脏磁共振扫描日期直至患者死亡或末次临床接触日期。

The primary endpoint was a life-threatening ventricular arrhythmic event, including sustained ventricular tachycardia (VT) (>100 beats/min lasting >30 s or requiring earlier termination), ventricular fibrillation (VF), appropriate ICD shock, or appropriate antitachycardia pacing. The secondary endpoints of the study were hospitalization for heart failure or heart transplantation, and all-cause mortality. The study was approved by the Research Ethics Committee of the Health Research Authority, United Kingdom.
主要终点为危及生命的室性心律失常事件，包括持续性室性心动过速（心率>100 次/分钟且持续>30 秒或需提前终止）、心室颤动、适当的 ICD 电击或适当的抗心动过速起搏。研究的次要终点包括因心力衰竭住院或接受心脏移植，以及全因死亡率。该研究已获得英国健康研究管理局研究伦理委员会的批准。

Digitally archived multiscanner multicenter CMR studies, performed in our institution or imported from referring UK hospitals, were retrospectively reviewed. All imaging protocols included steady-state free precession breath-hold cine imaging for the assessment of ventricular volumes and function, and standard LGE sequences with magnitude inversion recovery and phase-sensitive inversion recovery reconstructions, according to standard recommendations. In patients with a cardiac implantable electronic device, gradient echocardiography sequences for cine imaging and wide-band LGE acquisitions were used to improve image quality.
我们对本机构实施或从英国转诊医院导入的数字化存档多扫描仪多中心心脏磁共振研究进行了回顾性分析。所有成像方案均包含稳态自由进动屏气电影成像用于评估心室容积和功能，以及符合标准建议的常规晚期钆增强序列（包括幅值反转恢复和相位敏感反转恢复重建）。对于植入心脏电子设备的患者，采用梯度回波电影成像序列和宽频带晚期钆增强采集技术以提升图像质量。

All studies were analyzed with the use of CVI42 software (version 5.13.4, Circle Cardiovascular Imaging) by 3 CMR-experienced operators blinded to clinical information. Patients were classified into 3 categories based on LV systolic function: LVEF ≤35%, LVEF 36%-50%, and LVEF >50%. LGE location in the left ventricle was described according to both the AHA 17 myocardial segments model¹³ and a wall segmentation model (anterior, septal, inferior, and lateral). Multifocal LGE was defined as the presence of enhancement in at least 2 nonconsecutive and nonadjacent AHA myocardial segments. The LGE pattern was described according to the type of involvement of myocardial layers as mid-wall, subepicardial, subendocardial, transmural, or mixed. In the presence of at least 1 segment of transmural LGE, patients were classified as having transmural LGE. Mixed pattern included a combination of at least 2 different patterns between mid-wall, subepicardial, and subendocardial LGE. Patients with only focal prominent LGE in the ventricular insertion points were identified. The presence of LGE in the right ventricular (RV) free wall or on the right side of the interventricular septum also was recorded.
所有研究均由三位对临床信息不知情且具有心脏磁共振经验的操作者使用 CVI42 软件（版本 5.13.4，Circle Cardiovascular Imaging）进行分析。根据左心室收缩功能将患者分为三类：左室射血分数（LVEF）≤35%、LVEF 36%-50%和 LVEF >50%。左心室晚期钆增强（LGE）的位置描述同时采用美国心脏协会（AHA）17 节段心肌模型 ¹³ 和壁段划分模型（前壁、室间隔、下壁和侧壁）。多灶性 LGE 定义为至少两个非连续且不相邻的 AHA 心肌节段出现增强。根据心肌层受累类型将 LGE 模式描述为中层、心外膜下、心内膜下、透壁性或混合性。当存在至少一个透壁性 LGE 节段时，患者被归类为透壁性 LGE。混合模式包括中层、心外膜下和心内膜下 LGE 中至少两种不同模式的组合。仅在心室插入点出现局灶性显著 LGE 的患者被单独标识。 右心室游离壁或室间隔右侧出现晚期钆增强的情况也被记录在案。

LGE extent was expressed as the number of LV segments involved (17-segment model) and as a percentage of the overall LV myocardial mass. Quantification of LGE extent was assessed using the signal threshold vs reference myocardium method, which analyses the signal intensity of enhanced myocardial areas compared with a reference region of interest placed over normal nulled myocardium. A signal threshold of >5 SDs above the mean signal of the reference myocardium was applied to calculate the total LGE mass. LGE extent, as a percentage, was automatically calculated by dividing the LGE mass by the total LV mass. The >5-SD threshold was chosen because it was the most used in previous CS cohorts^10,14-17 and provided the best agreement with visual assessment. In case of inaccurate tracking, manual adjustment was performed. Interobserver variability was assessed by a fourth independent CMR operator, who analyzed LGE mass and extent in 27 CMR studies. Coefficient of variation was 6.1% for LGE mass and 6.3% for LGE extent.
晚期钆增强（LGE）范围以受累左心室节段数量（17 节段模型）及占左心室整体心肌质量的百分比表示。采用信号阈值与参考心肌对比法对 LGE 范围进行量化评估，该方法通过比较增强心肌区域信号强度与置于正常抑制心肌上的参考感兴趣区来实现。应用高于参考心肌平均信号>5 个标准差的信号阈值计算 LGE 总质量。LGE 范围百分比由 LGE 质量除以左心室总质量自动得出。选择>5 个标准差阈值是因为该标准在既往心脏结节病队列研究中应用最广泛 ^1014-17，且与视觉评估结果一致性最佳。若自动追踪不准确，则进行手动调整。由第四位独立心脏磁共振操作员对 27 项心脏磁共振研究的 LGE 质量与范围进行分析以评估观察者间变异性，结果显示 LGE 质量的变异系数为 6.1%，LGE 范围的变异系数为 6.3%。

We present continuous variables as mean ± SD or median (Q1-Q3) as appropriate. We compared categoric variables between 2 or more groups with the chi-square test, and we compared continuous variables between groups with either Student’s t-test or Mann-Whitney U-test (for 2 groups, as appropriate) or by analysis of variance (for 3 groups). We tested correlations between continuous variables with either Pearson’s r or Spearman’s rank correlation coefficient, as appropriate. The prognostic value of LGE extent, LGE mass, and number of myocardial segments for the various study endpoints of ventricular arrhythmia, heart failure or heart transplantation, and all-cause mortality was explored in unadjusted Cox regression models and after adjustment for age, sex, and presence of LV systolic dysfunction (ie, LVEF <50%). The different LGE patterns for the arrhythmia endpoint were assessed in the overall population and in specific population subgroups, ie, without ventricular arrhythmia events at presentation, LVEF >35%, and LVEF >50%. The optimum cutoff for LGE extent was selected by identifying the value that maximized Youden’s J statistic (sum of sensitivity and specificity) on logistic regression analysis for the primary endpoint to ensure an optimum balance between sensitivity and specificity in our models. Cox proportional hazard regression models were fitted to test the association of each variable with the outcome of interest. The assumption of proportional hazards was checked with the use of the Schoenfeld residuals test. Kaplan-Meier survival curves for the study subgroups, ie, with high vs low LGE extent and presence or absence of high-risk LGE features, were plotted with numbers at risk shown below each graph. In the subgroup of patients who did not show arrhythmia at presentation, we fitted a Cox proportional hazards regression model with as an independent variable the HRS indication for an ICD (ie, no indication for ICD, Class IIa, or Class I). On top of this model, we added, as a second covariate, the proposed risk classification system (ie, low, intermediate, high, or very high risk category). The 2 nested models were compared by means of the change in the likelihood ratio test (chi-square). The C-index of the model was derived from time-dependent ROC curve analysis, and the improvement in risk discrimination was assessed by the change in the c-index (ΔAUC) of the nested models (from time-dependent ROC analysis). The improvement in risk reclassification was assessed in a risk classification table. Statistical analyses were performed with the use of R version 4.2.0. All tests were 2-sided and alpha was set at 0.05.
连续变量以均值±标准差或中位数（Q1-Q3）表示。分类变量比较采用卡方检验，连续变量组间比较根据情况采用 Student's t 检验或 Mann-Whitney U 检验（两组比较），或方差分析（三组比较）。连续变量相关性检验根据情况采用 Pearson's r 或 Spearman 秩相关系数。通过未调整的 Cox 回归模型及校正年龄、性别和左室收缩功能障碍（即 LVEF<50%）后的模型，评估晚期钆增强（LGE）范围、LGE 质量及心肌节段数量对室性心律失常、心力衰竭或心脏移植、全因死亡率等研究终点的预测价值。针对心律失常终点，在总体人群及特定亚组（即就诊时无室性心律失常事件、LVEF>35%、LVEF>50%的人群）中评估了不同 LGE 模式的预测价值。 通过识别使主要终点逻辑回归分析中约登指数（敏感性与特异性之和）最大化的数值，选择 LGE 范围的最佳截断值，以确保模型在敏感性和特异性之间达到最佳平衡。采用 Cox 比例风险回归模型检验各变量与目标结局的关联性，并利用 Schoenfeld 残差检验验证比例风险假设。绘制研究亚组（即高/低 LGE 范围组及存在/不存在高风险 LGE 特征组）的 Kaplan-Meier 生存曲线，各图表下方标注风险人数。针对初诊时未出现心律失常的患者亚组，我们构建了以 HRS 的 ICD 植入指征（即无指征、IIa 类或 I 类）为自变量的 Cox 比例风险回归模型。在此模型基础上，将提出的风险分级系统（即低危、中危、高危或极高危）作为第二协变量纳入。通过似然比检验（卡方检验）的变化比较这两个嵌套模型。 该模型的 C 指数源自时间依赖性 ROC 曲线分析，风险判别能力的提升通过嵌套模型 C 指数的变化（ΔAUC，来自时间依赖性 ROC 分析）进行评估。风险重分类的改善情况通过风险分类表进行评估。统计分析使用 R 4.2.0 版本完成，所有检验均为双侧检验，显著性水平α设定为 0.05。

Of the initial 385 patients with biopsy-proven sarcoidosis and available good-quality CMR images who were screened for eligibility, we excluded 23 patients with known significant coronary artery disease, 18 with significant valvular heart disease, and 12 with an alternative diagnosis of cardiomyopathy. In addition, we excluded 8 patients in whom LGE extent could not be quantified due to technical issues during acquisition. In total, 324 patients with biopsy-proven sarcoidosis were included in this study (Figure 1). The study cohort included 212 (65.4%) men and 112 (34.6%) women. Mean age at the time of the baseline CMR scan was 53.6 years. Demographics and clinical characteristics of the study population are reported in Table 1. Baseline CMR findings are presented in Table 2.
在最初接受筛查的 385 例经活检确诊结节病且具有高质量心脏磁共振成像（CMR）图像的患者中，我们排除了 23 例已知患有显著冠状动脉疾病的患者、18 例有明显瓣膜性心脏病的患者以及 12 例确诊为其他类型心肌病的患者。此外，由于图像采集技术问题导致晚期钆增强（LGE）范围无法量化的 8 例患者也被排除。最终共有 324 例经活检确诊的结节病患者纳入本研究（ 图 1）。研究队列包括 212 例（65.4%）男性和 112 例（34.6%）女性。基线 CMR 检查时的平均年龄为 53.6 岁。研究人群的人口统计学和临床特征见表 1，基线 CMR 检查结果见表 2。

During a median follow-up period of 4.6 years (Q1-Q3: 2.3-7.5 years), 30 patients (9.3%) reached the primary endpoint including sustained VT (n = 22) or VF (n = 8). Among them, appropriate ICD therapy occurred for VT in 14 patients and for VF in 6 patients. Hospitalization for heart failure was recorded in 15 patients (4.6%), and 3 (0.9%) underwent heart transplantation. Overall, 41 patients (12.7%) died of any cause. In total, 61 patients reached at least 1 study endpoint; and among them, 7 reached both the primary endpoint and 1 of the secondary endpoints. Survival curves for the study endpoints are shown in Figure 2.
在中位随访 4.6 年期间（四分位距：2.3-7.5 年），30 例患者（9.3%）达到主要终点，包括持续性室速（22 例）或室颤（8 例）。其中 14 例室速患者和 6 例室颤患者接受了适当的 ICD 治疗。15 例患者（4.6%）因心力衰竭住院，3 例（0.9%）接受了心脏移植。总体而言，41 例患者（12.7%）死于各种原因。共有 61 例患者达到至少 1 项研究终点；其中 7 例同时达到主要终点和 1 项次要终点。研究终点的生存曲线如图 2 所示。

There were no differences in patient demographics (age, sex) or prevalence of traditional risk factors (arterial hypertension, diabetes mellitus, dyslipidemia, or smoking history) among subgroups of patients based on the presence or absence of primary endpoint. In the whole study population, LGE extent (as % of LV mass) was significantly higher among patients who reached the primary endpoint during follow-up (16.8% [Q1-Q3: 7.9%-38.6%] vs 6.2% [Q1-Q3: 2.3%-11.7%]; P < 0.001) (Table 2). In univariate cox regression analysis, LGE extent was significantly associated with the primary endpoint (ie, the composite of sustained VT, VF, and appropriate ICD therapy; P < 0.001) and with the secondary endpoints of heart failure and heart transplant (P = 0.014), but not with all-cause mortality (P = NS). After adjustment for age, sex, and presence of LV systolic dysfunction (ie, LVEF <50%), LGE extent was still significantly associated with the arrhythmia endpoint (P = 0.047) but not with any of the other study endpoints (Table 3). ROC analysis showed that the optimum cutoff of LGE extent for the prediction of the arrhythmic endpoint in the whole cohort was 7.4%, providing a sensitivity of 83% and specificity of 70% (AUC: 0.795) (Supplemental Figure 1). In leave-one-out cross-validation, we obtained an average C-index of 0.792 across all iterations.
根据主要终点事件是否存在划分的各亚组间，患者人口统计学特征（年龄、性别）及传统危险因素（高血压、糖尿病、血脂异常或吸烟史）的患病率均无显著差异。在整个研究人群中，随访期间达到主要终点的患者其晚期钆增强范围（占左心室质量的百分比）显著更高（16.8%[四分位距：7.9%-38.6%] vs 6.2%[四分位距：2.3%-11.7%]；P＜0.001）（ 表 2）。单变量 Cox 回归分析显示，晚期钆增强范围与主要终点（即持续性室速、室颤及恰当 ICD 治疗的复合终点；P＜0.001）以及心力衰竭和心脏移植的次要终点（P=0.014）显著相关，但与全因死亡率无关（P= 无统计学意义）。校正年龄、性别及左室收缩功能障碍（即 LVEF＜50%）后，晚期钆增强范围仍与心律失常终点显著相关（P=0.047），但与其他研究终点均无显著关联（ 表 3）。 ROC 分析显示，在整个队列中预测心律失常终点的 LGE 范围最佳截断值为 7.4%，其敏感性为 83%，特异性为 70%（AUC：0.795）（ 补充图 1）。在留一法交叉验证中，我们获得所有迭代的平均 C 指数为 0.792。

In our cohort, 260 patients (80.2%) had LVEFs >50% in their index CMR scan, 43 (13.3%) showed LVEFs from 36% to 50%, and the remaining 21 subjects (6.5%) had LVEFs ≤35%. The subgroup of patients with LVEF ≤35% was significantly more likely to develop a ventricular arrhythmic event during follow-up (HR: 5.24 [95% CI: 2.16-12.7]; P < 0.001), although most arrhythmic events occurred in patients with LVEF >35% (n = 20 out of 30 arrhythmic events). Survival curves per LVEF subgroups are shown in Figure 3A.
在我们的研究队列中，260 名患者（80.2%）基线心脏磁共振检查显示左室射血分数（LVEF）＞50%，43 名（13.3%）患者 LVEF 介于 36%至 50%之间，其余 21 名受试者（6.5%）LVEF≤35%。虽然大多数心律失常事件发生在 LVEF＞35%的患者中（30 例心律失常事件中有 20 例），但 LVEF≤35%亚组患者在随访期间发生室性心律失常事件的风险显著更高（风险比：5.24 [95%置信区间：2.16-12.7]；P＜0.001）。不同 LVEF 亚组的生存曲线如图 3A 所示。

Next to LVEF, high LGE extent (≥7.4% of LV mass) was significantly associated with the primary endpoint during follow-up (HR: 6.24 [95% CI: 2.38-16.40]; P < 0.001) in the whole cohort. Survival curves are shown in Figure 3B. LGE on the right side of the interventricular septum, LGE in the RV free wall, multifocal LGE in the LV, and transmural LGE were all more frequent among patients who reached the arrhythmia endpoint (Table 2) and all associated with multifold increased risk for the arrhythmic endpoint, even in subgroup analyses for primary prevention of SCD (ie, after excluding from the whole cohort those patients who experienced sustained VT, VF, or aborted SCD at presentation) (Table 4). Among patients with LVEF >35% and those with LVEF >50%, LGE extent ≥7.4%, LGE on the right side of the septum, and multifocal LGE were still predictive of arrhythmia events (Table 4). Kaplan-Meier survival curves for the whole cohort for the presence and absence of each of these LGE locations and patterns are shown in Figures 3C to 3F. The same analyses in the subgroup of patients without sustained VT, VF, or aborted SCD at presentation are shown in Supplemental Figures 2 and 3. Illustrations of different types of LGE involvement within study participants are shown in Figure 4.
除左室射血分数（LVEF）外，在整个队列中，晚期钆增强（LGE）范围较大（≥左室质量的 7.4%）与随访期间主要终点事件显著相关（风险比：6.24 [95%置信区间：2.38-16.40]；P < 0.001）。生存曲线如 图 3B 所示。室间隔右侧 LGE、右室游离壁 LGE、左室多灶性 LGE 以及透壁性 LGE 在达到心律失常终点的患者中更为常见（ 表 2），且均与心律失常终点风险倍增相关，即使在针对心源性猝死（SCD）一级预防的亚组分析中（即从整个队列中排除就诊时出现持续性室速、室颤或抢救成功的 SCD 患者）亦是如此（ 表 4）。在 LVEF >35%和 LVEF >50%的患者中，LGE 范围≥7.4%、室间隔右侧 LGE 以及多灶性 LGE 仍可预测心律失常事件（ 表 4）。整个队列中不同 LGE 部位和模式存在与否的 Kaplan-Meier 生存曲线见图 3C 至 3F。 在初诊时未出现持续性室速、室颤或心脏骤停抢救成功的患者亚组中进行的相同分析结果见补充图 2 和 3。 图 4 展示了研究参与者中不同类型的晚期钆增强受累情况示意图。

Based on our observations, we defined high-risk LGE as the presence of at least 1 high-risk LGE feature, defined as LGE on the right side of the septum, LGE in the RV free wall, or multifocal LGE in the LV. Among the 324 study patients, we identified 4 risk group categories based on LVEF, LGE extent, and the presence of high-risk LGE features (Central Illustration A). Patients at very high risk of developing life-threatening ventricular arrhythmias were all those with LVEF ≤35% and LGE extent ≥7.4%. The high risk category included patients with LVEF ≤35% and LGE extent <7.4% with high-risk LGE features, and patients with LVEF 36%-50% and high-risk LGE features, independently from LGE extent. In the intermediate risk category were subjects with either LVEF ≤35% or LVEF 36%-50% and LGE <7.4% in the absence of high-risk LGE, and those with LVEF >50% and LGE ≥7.4%, or LGE <7.4% and high-risk LGE. Finally, the low risk category included only patients with LVEF >50% and LGE <7.4% in the absence of high-risk LGE. Based on this risk classification system, a total of 16 patients (5%) were in the very high risk, 39 (12%) in the high risk, 122 (38%) in the intermediate risk, and 147 (45%) in the low risk category of developing a life-threatening ventricular arrhythmia during follow-up. The respective survival curves of the risk subgroups are shown in Central Illustration B. The event rate was 44% (10.8% per year) in the very high risk group, 23% (or 5.6% per year) among high risk patients, and 11% (or 2.7% per year) in the intermediate risk group. Only 1% (0.2% per year) of patients in the low risk category reached the arrhythmic endpoint during the follow-up period (Central Illustration C).
根据我们的观察，我们将高危晚期钆增强定义为至少存在 1 项高危特征，包括室间隔右侧晚期钆增强、右室游离壁晚期钆增强或左室多灶性晚期钆增强。在 324 例研究患者中，我们根据左室射血分数、晚期钆增强范围及高危特征的存在情况划分了 4 个风险组别（ 核心图示 A）。发生致命性室性心律失常风险极高的患者均为左室射血分数≤35%且晚期钆增强范围≥7.4%者。高风险组包括左室射血分数≤35%伴晚期钆增强范围 <7.4%且具有高危特征的患者，以及左室射血分数36%-50%伴高危特征（无论晚期钆增强范围大小）的患者。中风险组包含左室射血分数≤35%或36%-50%伴晚期钆增强<7.4%但无高危特征的患者，以及左室射血分数>50%伴晚期钆增强≥7.4%或 <7.4%但具有高危特征的患者。低风险组仅包含左室射血分数>50%伴晚期钆增强<7.4%且无高危特征的患者。 根据该风险分级系统，随访期间共有 16 例患者（5%）处于发生致命性室性心律失常的极高危组，39 例（12%）为高危组，122 例（38%）为中危组，147 例（45%）为低危组。各风险亚组的生存曲线如核心图示 B 所示。极高危组事件发生率为 44%（年发生率 10.8%），高危组为 23%（年发生率 5.6%），中危组为 11%（年发生率 2.7%）。低危组患者在随访期间仅 1%（年发生率 0.2%）达到心律失常终点（ 核心图示 C）。

This risk stratification approach based on LVEF, LGE extent, and presence of high-risk LGE features was compared with the HRS recommendations for ICD placement for primary prevention of SCD in CS.⁷ Among the overall cohort of 324 patients, 27 presented with ventricular arrhythmias (23 with sustained VT and 4 with VF) and received an ICD for secondary prevention of SCD. In the subgroup of patients who were assessed for ICD placement for primary prevention of SCD (n = 297), compared with the HRS criteria, the proposed risk stratification algorithm based on LVEF and LGE features significantly enhanced the model’s prognostic performance (chi-square = 8.02; P = 0.046), risk discrimination (ΔAUC = 0.082; P = 0.019), and reclassification for the primary endpoint (Table 5). More specifically, a significant proportion of HRS Class IIa patients (43%) were reclassified to a lower risk category, and a smaller proportion to a higher risk category (11%).
这种基于左室射血分数（LVEF）、晚期钆增强（LGE）范围及高危 LGE 特征的风险分层方法，与心律学会（HRS）关于心脏结节病患者植入心律转复除颤器（ICD）一级预防心源性猝死（SCD）的建议进行了对比。⁷ 在 324 例患者队列中，27 例出现室性心律失常（23 例持续性室速，4 例室颤）并接受 ICD 二级预防 SCD。在评估 ICD 一级预防 SCD 的亚组患者（n=297）中，与 HRS 标准相比，基于 LVEF 和 LGE 特征提出的风险分层算法显著提升了模型的预后性能（卡方值=8.02；P=0.046）、风险区分度（ΔAUC=0.082；P=0.019）以及对主要终点的重分类能力（ 表 5）。具体而言，相当比例的 HRS IIa 类患者（43%）被重新划分为更低风险类别，而较小比例（11%）被划分为更高风险类别。

This was a retrospective observational study performed in a single tertiary center and as such is subject to selection and referral bias. The cohort is predominantly Caucasian, with a smaller percentage of Asian and Black patients. Although patients with poor-quality CMR studies were appropriately excluded from this study, the accuracy of LGE extent quantification might have been affected in patients who had a cardiac implantable electronic device at the time of CMR. The effect of immunosuppressive, antiarrhythmic, or heart failure medications on clinical outcomes was beyond the scope of this study. Finally, cardiovascular mortality was not assessed independently as the cause of death was not available for most patients. Given the relatively small number of arrhythmic events, external validation of the findings in other cohorts is desirable.
这是一项在单一三级医疗中心开展的回顾性观察研究，因此存在选择和转诊偏倚。研究队列以白种人为主，亚裔和非裔患者占比较小。尽管本研究已合理排除了心脏磁共振成像质量不佳的患者，但对于检查时携带心脏植入式电子设备的患者，晚期钆增强范围的量化准确性可能受到影响。免疫抑制剂、抗心律失常药物或心力衰竭药物对临床结局的影响不在本研究探讨范围内。最后，由于大多数患者缺乏明确死因记录，未能单独评估心血管死亡率。鉴于心律失常事件数量相对较少，建议在其他队列中对研究结果进行外部验证。