Clinical use of nuclear cardiology in the assessment of heart failure

A nuclear cardiology test is the most commonly performed non-invasive cardiac imaging test in patients with heart failure, and it plays a pivotal role in their assessment and management. Quantitative gated single positron emission computed tomography (QGS) is used to assess quantitatively cardiac volume, left ventricular ejection fraction (LVEF), stroke volume, and cardiac diastolic function. Resting and stress myocardial perfusion imaging, with exercise or pharmacologic stress, plays a fundamental role in distinguishing ischemic from nonischemic etiology of heart failure, and in demonstrating myocardial viability. Diastolic heart failure also termed as heart failure with a preserved LVEF is readily identified by nuclear cardiology techniques and can accurately be estimated by peak filling rate (PFR) and time to PFR. Movement of the left ventricle can also be readily assessed by QGS, with newer techniques such as threedimensional, wall thickening evaluation aiding its assessment. Myocardial perfusion imaging is also commonly used to identify candidates for implantable cardiac defibrillator and cardiac resynchronization therapies. Neurotransmitter imaging using 123 I-metaiodobenzylguanidine offers prognostic information in patients with heart failure. Metabolism and function in the heart are closely related, and energy substrate metabolism is a potential target of medical therapies to improve cardiac function in patients with heart failure. Cardiac metabolic imaging using 123 I-15-(p-iodophenyl)3-R, S-methylpentadecacoic acid is a commonly used tracer in clinical studies to diagnose metabolic heart failure. Nuclear cardiology tests, including neurotransmitter imaging and metabolic imaging, are now easily preformed with new tracers to refine heart failure diagnosis. Nuclear cardiology studies contribute significantly to guiding management decisions for identifying cardiac risk in patients with heart failure.

心肌模型和分析软件的研发及其在心肌SPECT显像标准化中作用的多中心研究

目的该课题组自主研发了一款心肌模型和分析软件，通过标准化的定量指标来客观评价显像质量。研究目的是明确该模型和软件是否适用于显像的标准化。方法为对心肌SPECT显像进行标准化，课题组研发了心肌显像的评价分析软件（EMIT）和模型，利用该模型对胸部进行模拟，其内设置了肺和心脏。心肌模型包括5个放射性正常区域和8个放射性缺损区域。

利用日本多中心数据库对人工神经网络进行再训练以探测心肌缺血

目的人工神经网络（ANN）已被应用于探测心肌灌注缺损和缺血。该研究的目的是比较最新1.1版本的ANN与最初1.0版本的诊断准确性。方法研究了106例冠状动脉（简称冠脉）造影阳性的患者[平均年龄（77±10）岁]，其中多支冠脉病变（狭窄≥50%）者占52%，陈旧性心肌梗死患者占27%，行冠脉重建术者占30%。1.0和1.1版本的ANN已分别在瑞典（1 051例）和日本（1 001例）用于99Tcm-甲氧基异丁基异腈（MIBI）心肌灌注显像的诊断训练，计算其诊断当地患者负荷态心肌灌注缺损和缺血的可能性（0.0～1.0）。以相关专家的阅片结果作为＂金标准＂，计算受试者工作特征（ROC）曲线下面积（AUC），比较2个版本ANN的诊断准确性。结果1.1和1.0版本的ANN诊断负荷态心肌灌注缺损的AUC分别为0.95和0.93（P＝0.27），但1.1版本诊断心肌缺血的能力显著提高（P＝0.005 5），其AUC为0.96（灵敏度87%，特异性96%），1.0版本的AUC为0.89（灵敏度78%，特异性97%）。在未行冠脉重建且无陈旧性心肌梗死的患者中，1.1版本的AUC也有显著提高（P＝0.009 3），其AUC为0.98（灵敏度88%，特异性100%），而1.0版本的AUC为0.88（灵敏度76%，特异性100%）。与1.0版本相比，1.1版本的中位ANN可能性常为0.1～0.7，这使得1.1版本的诊断准确性更高。以冠脉狭窄作为＂金标准＂时，1.1版本的诊断准确性在单支冠脉病变或无冠脉狭窄的患者中也有提高（47例；AUC：0.81与0.66，P＝0.006 0）。结论1.1版本的ANN在日本人数据库中训练后，其诊断能力得到提高，尤其在诊断心肌缺血方面。