Objective To evaluate the accuracy and value of measurement of left ventricular systolic function in dilated cardiomyopathy (DCM) and hypertrophic cardiomyopathy (HCM) patients with stereo three-dimensional echocardiography (S3 DE). Methods End-diastolic volume (EDV), end-systolic volume (ESV),stroke volume(SV) and ejection fraction(EF) of the left ventricle were measured with M-mode echocardiography(ME),two-dimensional echocardiography(2DE) and S3DE in DCM patientsC20 cases). HCM patients(20 cases),and normal controls(20 cases). The different results among the three groups or three methods were analyzed. Results (①In all the three groups,the results of EDV,ESV,and SV obtained with ME were significantly higher than those obtained with S3DE( P <0. 01). Only in normal group( P <0. 01) and HCM group ( P <0. 05) ,the results of EF obtained with ME and 2DE were significantly higher than that obtained with S3DE. ②By S3DE,compared with normal group,EDV,ESV were increased and EF was decreased obviously in DCM group (all P <0. 01); while in HCM group, only SV was significantly higher( P < 0. 01). ③EDV, ESV, and EF measured by S3DE were correlated and fit well with those measured by 2DE(r = 0.778,0.876, 0.932;R2 =0.605,0.767,0.869;all P <0.01). ④Within HCM group,excluding the impact of heart rate,cardiac output (CO) was highly correlated with SV( r = 0. 987,P < 0. 01). Conclusions S3DE can real-time display the stereo structure of the heart, and accurately and reliably assess the left ventricular systolic function, with a priority over traditional ME and 2DE methods. EDV,ESV, and EF are still effective indicators for the clinical assessment of left ventricular systolic function. SV obtained with S3DE will be expected to be the more sensitive and accurate value in assessing left ventricular systolic function in patients with early-stage cardiomyopathy.

Excellent lead compounds have a profound influence on drug development and can improve the success rate of product launch. It is expensive and time-consuming to discover lead compounds by traditional methods, yet artificial intelligence (AI) can discover good lead compounds efficiently.This article systematically summarizes the research progress of obtaining lead compounds through the screening and generation models of AI, classifies different models according to the type of information input, focuses on drug repurposing by screening model and multi-objective drug design by generation model, and discusses the development prospect of AI in the research field of lead compounds, aiming to provide new research ideas for the application of AI in lead compounds.