第一部分
基于 18F-FDG PET 的心脏功能评估对植入型心律转复除颤器 置入患者室性心律失常事件的预测作用
研究背景 室性心律失常是(Ventricular Arrhythmia，VA)是心脏性猝死(Sudden Cardiac Death，SCD)主要原因之一。对 VA 风险的准确预测和早期发现是预防 SCD 发生的重要策略之一。左室射血分数(Left Ventricular Ejection Fraction，LVEF)作 为评价心脏功能的影像学指标，常常用于识别 SCD 高危风险，但其对 V A 预测能力 较弱。多种心脏疾患导致的心室复杂或异常的解剖结构，心肌病变或瘢痕等病理状 态为 VA  供了触发因素和维持基质，更导致了心脏功能的异常。18 氟-氟脱氧葡萄 糖正电子发射断层显像(18F-fluorodeoxyglucose Positron Emission Tomography，18F- FDG PET)基于其代谢性示踪剂的特殊效能，在影像学评估心脏整体及局部功能中 显示独特优势。本研究拟利用 18F-FDG PET，对置入植入型心律转复除颤器 (Implantable Cardioverter-Defibrillator，ICD)的患者进行前瞻性研究，探究研究期 内发生 VA 事件患者心脏功能的影像学特征，探讨基于 18F-FDG PET 的心脏功能评 估预测 VA 发生的可能性，为 VA 乃至 SCD 的有效防治 供新技术途径。
研究方法 本研究为前瞻性研究。从 2017 年 11 月至 2019 年 1 月由阜外医院纳入拟 植入 ICD 的患者。所有患者在术前进行 99m 锝-甲氧基异丁基异腈( 99mTc-MIBI)门控 单光子发射计算机断层成像术(Single-Photon Emission Computed Tomography, SPECT)/ 18F-FDG PET 检查。图像分析采用美国心脏协会推荐的 17 节段分析法。 ICD 术后每 6~12 个月，或必要时对患者进行随访。通过对 ICD 程控或追踪医疗记 录明确 VA 事件。分析图像各参数结果与 VA 发生的相关性。
研究结果 本部分研究纳入51例患者(33名男性，53.9±17.2岁)，在平均随访12.2±6.4 月(所有患者随访时间在 1 个月~2 年)后有 17 例患者(33.3%)发生了 VA，其中 共有 57 人次室性心动过速(其中 39 阵接受抗心动过速起搏治疗，15 阵接受电击治 疗后转复)及 5 人次心室颤动(均系 ICD 电击治疗后转复)。与无 VA 事件患者相 比，VA 患者的相位标准差较大(51.4°±14.0°vs.34.0°±15.0°)、带宽更宽 (172.9°±39.8°vs.128.7°±49.9°)、总室壁增厚率评分更高(Sum Thickening Score, STS， 29.5±11.1 vs.17.8±13.2)、总室壁运动评分更高(42.9±11.5 vs.33.0±19.0)及瘢痕面 积(17.7±12.4%vs.7.0±7.9%)较大。其中 14 名患者同时进行了 SPECT/ 18F-FDG PET 及心脏核磁检查，将两种检查进行对比，18F-FDG PET 识别瘢痕与心肌钆对比剂延迟显像增强中等相关(McNemer 检验 P = 0.5; Kappa 系数为 0.44, P = 0.047)。Cox 回归分析示相位标准差、带宽、STS 及心肌瘢痕与 VA 相关。利用约登指数计算危 险因素参考临界值:相位标准差为 36.6° (敏感度 58.82% ，特异度 88.24%), 带宽 157.6° (敏感度 67.65%，特异度 76.47%)，STS 26.5 分( 敏感度 63.64% ，特异度 73.33%)，瘢痕面积 9.5% (敏感度 64.71%，特异度 76.47%)。相位标准差、带宽、 STS 及瘢痕面积的 ROC 曲线下面积分别为 0.78, 0.73, 0.72 和 0.71。VA 事件生存曲 线 示患者合并危险因素越多发生 VA 风险越高，合并 4 项，1~3 项及未合并危险 因素的风险分别为 64.7%, 26.3%及 6.67%(logrank P = 0.0011)。
研究结论 对置入 ICD 患者的前瞻性研究发现， VA 事件发生者心脏功能的 18F- FDG PET 评估影像学具有心脏不同步增加、STS 较高以及心肌瘢痕面积较大等重要 特征。 18F-FDG PET 心脏功能评估上述三项指标独立或合并存在， 示患者 VA 风 险程度，具有 VA 发生的预测作用。

第二部分
基于 18F-FDG PET 的节段性评估室壁增厚率对植入型心律转复 除颤器置入患者室性心律失常事件发生的预测价值
研究背景 评估室性心律失常(Ventricular Arrhythmia，VA)患者心肌特点，对于 发现室性心律失常危险因素，预防心脏性猝死(Sudden Cardiac Death， SCD)及选 择恰当的植入型心律转复除颤器(Implantable Cardioverter-Fibrillation，ICD)适应 证患者具有重要价值。由于致病病因不同，心肌基质不均一性等，对有 VA 风险 患者的心肌评价既应作为一个连续的整体来考虑，也要进行节段性分析。本研究 旨在通过 18 氟-氟脱氧葡萄糖正电子发射断层显像(18F-fluorodeoxyglucose Positron Emission Tomography，18F-FDG PET)评估节段性室壁增厚率对 ICD 患者室性心 律失常的预测价值。
研究方法 本部分研究纳入56名拟植入ICD的患者。术前均行99m锝-甲氧基异丁 基异腈( 99mTc-MIBI)门控单光子发射计算机断层成像术(Single-Photon Emission Computed Tomography, SPECT)心肌显像和 18F-FDG PET 心肌代谢显像。图像通过 美国心脏协会推荐的17节段分析法，应用QGS软件获得室壁增厚率(Wall thickening, WT)，室壁厚度(Wall motion, WM)，左室舒张末容积(Left Ventricular End Diastolic Volume, LVEDV)，左室收缩末容积(Left Ventricular End Systolic Volume, LVESV)， 左室射血分数(left ventricular ejection fraction，LVEF);应用 QPS 软件自动算法量 化分析心肌瘢痕及冬眠心肌面积等指标。患者在植入 ICD 后每 6-12 个月进行定期 随访，或当患者接收ICD治疗或必要时对患者进行额外随访。观察患者发生VA的 情况，分析患者 SPECT/18F-FDG PET 心肌影像学特点与发生室性心律失常的相关 性。
研究结果 在分析的 56 例患者(40 名男性，54.7±16.8 岁)中，18 例(32.14%)出 现 VA。对 952 个节段的局部功能(WT/WM)和心肌活性(正常、不匹配、瘢痕) 进行了评估。与无 VA 组相比，VA 组 WT 正常节段和 WM 正常节段比率明显降低 (分别为 19% vs. 37%，15% vs. 25%，均 P<0.001)，WT 消失节段和反向运动节 段发生率明显升高(分别为 28% vs. 12%，9% vs. 4%，P<0.001, P<0.05)。依据 节段活性分层，VA 组患者灌注正常节段中 WT 正常节段比率偏低(30% vs. 44%， P<0.05)，WT 消失节段的比例较高(18% vs. 8%，P<0.001)，且上述分布情况 在冬眠心肌组(11% vs. 33%，33% vs. 16%，P<0.001，P<0.05)，及心肌瘢痕组 (3% vs. 14%，44% vs. 24%，P 值均<0.001)一致。WM 节段分布显示，无 VA 患 者心肌瘢痕节段中正常 WM 节段比率更高(14%vs.0%，P<0.001)。以 Logistic 回归分析患者节段性心肌 SPECT/18F-FDG PET 影像学特点显示，WT 及心肌瘢痕(OR， 1.513，95%CI 1.014–2.258，P=0.043)与 VA 的发生有关。WT 分数越高，发生 VA 风险越大(1 至 3 分时 OR 和 95%CI 分别为 1.873[1.111-3.156]， 2.587[1.470-4.554] 和 4.907[2.596-9.276])。分析患者整体心肌特点，Cox 回归分析显示，室壁增厚率 总评分(Sum Thickening Score, STS)分数越高，瘢痕面积越大，VA 发生率越高(HR 分别为 1.049[1.003-1.097]及 1.040[1.003-1.079])。Kaplan-Meier 生存曲线显示，WT 消失的患者更易发生 VA，无论伴或不伴瘢痕(P=0.002 或 P<0.001)。存在心肌瘢 痕者合并 WT 异常更易发生 VA(P=0.004)，然而若 WT 正常，心肌瘢痕存在与否 患者的预后无显著差异(P=0.668)。
研究结论 本研究利用 18F-FDG PET 对拟置入 ICD 患者节段性评估 WT，可在 STS 评估心脏整体功能的基础之上更好的评估局部功能，以综合评估整体和局部心功能。 此外本研究还发现，WT 的评估有助于心肌瘢痕更好地识别存在 VA 高风险患者。

PartI
Cardiac function quantified by 18F-FDG PET scans can predict ventricular arrhythmia in patients with implantable cardioverter defibrillator
Background. Ventricular arrhythmia (VA) is one of the major causes of sudden cardiac death (SCD) for patients. Accurate prediction and early detection of risk of VA is one of the important strategies to prevent SCD. Left ventricular ejection fraction (LVEF) as the primary approach to identify patients with high risk of sudden cardiac death (SCD) has poor predictive value for VA. Ventricular complicated or abnormal anatomical structure, myocardial lesions or scars and other factors provide the trigger and maintenance substrate for VA, which lead cardiac dysfunction. 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) has unique advantages in evaluating Global and regional cardiac dysfunction. This prospective study sought to investigate the imaging characteristics of VA patients to predicting VA for patients with implantable cardioverter defibrillator (ICD) and the possibility of predicting the occurrence of VA based on 18F- FDG PET, which provide new technical ideas for effective prevention and treatment of VA.
Methods. We performed a prospective observational study and recruited patients who required ICD placement. Pre-procedure image scans were performed. The global and regional cardiac function parameters, left ventricular dyssynchrony parameters, scar burden was analyzed by Single-Photon Emission Computed Tomography (SPECT)/ 18F- FDG PET. For each patient assessment, a standard 17-segment of the left ventricle proposed by the American Heart Association was applied. Follow-up visits were scheduled every 6-12 months, or more often when clinically indicated. Occurrence of ventricular tachycardia (VT) and ventricular fibrillation (VF) was established through routine ICD interrogations and review of electronic medical records. Associations between quantitative parameters of imaging and VA were analyzed.
Results. In 51 patients (33 males, 53.9 ± 17.2 years) with mean follow-up of 12.2 ± 6.4 months (range 1 month- 2 years), 17 (33.3%) patients developed VA. There were 57 VTs (39 received antitachycardia pacing therapy and 15 received shock) and five VFs (all received therapy and shock). Compared with patients without VA, patients with VA had significantly larger values in phase standard deviation (51.4° ± 14.0° vs. 34.0° ± 15.0°), bandwidth (172.9° ± 39.8° vs. 128.7° ± 49.9°), sum thickening score (STS, 29.5 ± 11.1 vs. 17.8 ± 13.2), sum motion score (42.9 ± 11.5 vs. 33.0 ± 19.0) and scar area (17.7 ± 12.4% vs. 7.0 ± 7.9%). Fourteen patients received both PET and cardiac CMR scanning were performed comparison between the presence of Scar by PET and LGE by CMR. The presence of scar and LGE was moderate correlation (McNemer’s test P = 0.5; Kappa = 0.44, P = 0.047). Cox regression analysis showed that phase standard deviation, bandwidth, STS and scar size were associated with VA occurrence. Cut-off values for predicting VA, which identified by Youden’s index, for PSD, bandwidth, scar tissues and STS were 36.6° (sensitivity of 58.82% and specificity of 88.24%), 157.6° (sensitivity of 67.65% and specificity of 76.47%), 9.5% (sensitivity of 64.71% and specificity of 76.47%), and 26.5 (sensitivity of 63.64% and specificity of 73.33%), respectively. The area under curve (AUC)s of PSD, bandwidth, STS and scar area were 0.78, 0.73, 0.72 and 0.71, respectively. Survival analysis showed the patients with more risk factors had significantly higher risk for the prevalence of VA, with 64.7%, 26.3%, and 6.67% for those with 4, 1-3, and 0 risk factors, respectively (logrank P = 0.0011)
Conclusions. Increased dyssynchrony, higher STS and larger left ventricular scar burden quantified by 18F-FDG PET may indicate a higher VA incidence after ICD placement. 18F- FDG PET can be considered to assess these global and regional variables to identify the patients with high risk of VA.

Part II
Segmental myocardial wall thickening abnormalities assessment by 18F-FDG PET scans increase the risk of ventricular arrhythmia in patients with implantable cardioverter defibrillator
Background. It is important to evaluate the myocardial characteristics of patients with ventricular arrhythmia (VA) for identifing the risk factors of VA, preventing sudden cardiac death (SCD) and selecting proper patients who will benefit from implantable cardioverter defibrillator (ICD). Due to the different etiology and heterogeneity of myocardial substrate, the myocardial evaluation of patients with VA risk should not only be considered as a continuous whole, but also be performed segmental analysis. This present study was to assess the additional value of systolic wall thickening to viability assessment in predicting ventricular arrhythmia by 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) in patients with ICD placement.
Methods. We performed a prospective observational study and recruited 56 patients who required ICD placement. Both pre-procedure 99mTc-MIBI gated SPECT MPI and 18F-FDG PET myocardial metabolic image scans were performed. For each patient assessment, a standard 17-segment of the left ventricle proposed by the American Heart Association was applied. QGS software was used for the following cardiac function variables: wall thickening (WT), and wall motion (WM), left ventricular end diastolic volume (LVEDV), left ventricular end systolic volume (LVESV), left ventricular ejection fraction (LVEF). QPS software was used to grade 99mTc-MIBI and 18F-FDG uptakes in all segments and quantitative analysis of scar area, and hibernating myocardium. The patient’s follow-up was done every 6–12 months after ICD implantation, or when necessary. Associations between image results and VA were analyzed.
Results. In the 56 patients (40 males, 54.7 ± 16.8 years) analyzed, 18(32.14%) developed VA. Regional function (wall motion/thickening) and viability pattern (normal, mismatch, scar) were evaluated for 952 segments. Compared with patients without VA, patients with VA had significantly lower frequency of normal wall thickening segments and wall motion segments (19% vs.37% and 15% vs. 25%, all the P <0.001), higher frequency of absent wall thickening segments and dyskinetic segments (28% vs.12% and 9% vs. 4%, P<0.001 and P<0.05, respectively). In light of viability of segments, patients with VA has lower frequency of normal wall thickening and higher absent wall thickening segments in normal perfusion (30% vs. 44% and 18% vs. 8%, P<0.05 and P<0.001, respectively), hibernating myocardium (11% vs. 33% and 33% vs.16%, P<0.001 and P<0.05, respectively) and scar segments respectively (3% vs. 14% and 44% vs.24%, all the P values <0.001). The distribution of WM segments showed that more normal wall motion segments were found in scar segments of patients without VA (14% vs.0%, P<0.001). Logistic regression analysis for segmental assessment showed that wall thickening and scar (OR, 1.513, 95% CI 1.014 – 2.258, P=0.043) were associated with VA occurrence. The higher degree of wall thickening reduced, the greater association with VA (OR and 95% CI for score 1 vs. 2 vs.3 were 1.873[1.111-3.156] vs. 2.587[1.470-4.554] vs. 4.907[2.596-9.276], respectively). Cox regression analysis for global cardiac assessment showed that higher STS score and larger scar size were associated with VA occurrence (HR [95%CI], 1.049[1.003-1.097],1.040[1.003-1.079], respectively). Kaplan-Meier survival curve showed that patients with absent wall thickening more likely to experience VA with or without scar(P=0.002 or P< 0.001) . Patients with scar more likely to experience VA with wall thickening reduced (P=0.004), whereas outcomes not significantly different from patients without wall thickening abnormal whether with or without scar(P=0.668).
Conclusions. In this study, 18F-FDG PET was used to segmental evaluate the WT of ICD patients, which can better evaluate the regional cardiac function on the basis global cardiac function. Furthermore, assessment of myocardial wall thickening in addition to scar segments allowed better identified the patients with high risk factors of VA occurrence.