Li Yanan, Zheng Qiang, Cui Cunying, Liu Yuanyuan, Hu Yanbin, Huang Danqing, Wang Ying, Liu Jun, Liu Lin
Department of Ultrasound, Henan Province People's Hospital, Fuwai Central China Cardiovascular Hospital, People's Hospital of Zhengzhou University, Zhengzhou, China.
School of Computer and Control Engineering, Yantai University, Yantai, China.
Quant Imaging Med Surg. 2022 Jan;12(1):244-256. doi: 10.21037/qims-20-1038.
Accurate evaluation of left ventricular (LV) systolic function is the premise for diagnosing and treating chronic heart failure. This study aimed to explore the incremental value of echocardiographic myocardial work in evaluating the LV systolic dysfunction in patients with chronic heart failure.
A total of 206 participants were enrolled, including 155 patients with chronic heart failure and 51 healthy controls (HC). The chronic heart failure patients were divided into three groups according to LV ejection fraction (LVEF): Heart failure with preserved ejection fraction (HFpEF group, 54 cases, LVEF ≥50%), heart failure with mid-range ejection fraction (HFmrEF group, 50 cases, 40%≤ LVEF <50%), and heart failure with reduced ejection fraction (HFrEF group, 51 cases, LVEF <40%). Except for the conventional echocardiographic parameters, the left ventricular myocardial work parameters, including the global myocardial work index (GWI), global constructive work (GCW), global wasted work (GWW), and global work efficiency (GWE), were calculated in the study participants. One-way analysis of variance test followed by Fisher's least significant difference (LSD) -test were used to obtain parameters with significant differences, which were then fed into a machine learning model established for subsequent multi-classification of the four groups. The selected myocardial work parameters with high importance rankings resulting from the machine learning model were further compared with the traditional LVEF in the multi-classification of the four groups.
All conventional echocardiographic parameters were significantly different between the HFmrEF and HFrEF groups, but only E/e', left atrium showed notable differences between the HFpEF and HC groups (P<0.05). All myocardial work parameters were markedly different between the four groups (P<0.05). LVEF and GWI were more important than the other parameters according to the multi-classification machine learning model. The multi-classification diagnostic performances of LVEF, GWI, and LVEF + GWI were 82%, 88%, and 98%, respectively, which confirmed that GWI + LVEF could complementarily improve the diagnosis accuracy in classifying the four groups, with a performance increase of approximately 10% than each individually.
GWI can play a complementary role to LVEF in the early diagnosis of HFpEF patients from the HC group and improve the clinical evaluation accuracy in chronic heart failure patients. Echocardiographic myocardial work should be utilized along with conventional LVEF to evaluate the systolic function of chronic heart failure patients in clinical practice.
准确评估左心室（LV）收缩功能是慢性心力衰竭诊断和治疗的前提。本研究旨在探讨超声心动图心肌做功在评估慢性心力衰竭患者左心室收缩功能障碍中的增量价值。
共纳入206名参与者，包括155例慢性心力衰竭患者和51名健康对照者（HC）。慢性心力衰竭患者根据左心室射血分数（LVEF）分为三组：射血分数保留的心力衰竭（HFpEF组，54例，LVEF≥50%）、射血分数中度降低的心力衰竭（HFmrEF组，50例，40%≤LVEF<50%）和射血分数降低的心力衰竭（HFrEF组，51例，LVEF<40%）。除常规超声心动图参数外，还计算了研究参与者的左心室心肌做功参数，包括整体心肌做功指数（GWI）、整体建设性做功（GCW）、整体浪费做功（GWW）和整体做功效率（GWE）。采用单因素方差分析，随后进行Fisher最小显著差异（LSD）检验以获得有显著差异的参数，然后将这些参数输入为四组后续多分类建立的机器学习模型。将机器学习模型得出的具有高重要性排名的选定心肌做功参数在四组多分类中与传统LVEF进一步比较。
HFmrEF组和HFrEF组之间所有常规超声心动图参数均有显著差异，但HFpEF组和HC组之间仅E/e'、左心房有显著差异（P<0.05）。四组之间所有心肌做功参数均有显著差异（P<0.05）。根据多分类机器学习模型，LVEF和GWI比其他参数更重要。LVEF、GWI和LVEF+GWI的多分类诊断性能分别为82%、88%和98%，这证实GWI+LVEF可互补提高四组分类的诊断准确性，性能比各自单独使用时提高约10%。
GWI在从HC组早期诊断HFpEF患者中可对LVEF起到互补作用，并提高慢性心力衰竭患者的临床评估准确性。在临床实践中，应将超声心动图心肌做功与传统LVEF一起用于评估慢性心力衰竭患者的收缩功能。
