Nature cardiovascular research最新文献

Medical imaging often captures multiple two-dimensional views of three-dimensional anatomic structures, but most artificial intelligence (AI) models analyze two-dimensional data. Here we show that integrating multiple imaging views using a single AI model can improve diagnostic performance. We developed a deep neural network (DNN) architecture that combines information from multiple video views simultaneously. Using echocardiogram data from the University of California, San Francisco, and the Montreal Heart Institute, we applied our multiview DNN approach for three primary demonstration tasks: detecting any left or right ventricular abnormality, diastolic dysfunction, and substantial valvular regurgitation. Across various tasks, our multiview DNNs improved discrimination as measured by the area under the receiver operating characteristic curve by 0.06-0.09 compared to DNNs trained on any single view. This demonstrates that AI models that can combine information from multiple imaging views simultaneously can better capture complex anatomy and physiology for certain tasks, underscoring the value of a multiview paradigm for AI in medical imaging.

Individuals with hypoplastic left heart syndrome (HLHS) have an underdeveloped left ventricle and require surgery to reconfigure blood flow for survival. Here we profiled the HLHS right-ventricular microenvironment by single-nucleus RNA sequencing and spatial transcriptomics at birth (before heart failure), after surgery with heart failure and after ventricular assist device (VAD) unloading (reduced hypoxia and volume overload). We show that HLHS cardiomyocytes, both within the heart and when derived from induced pluripotent stem cells, are intrinsically senescent. The HLHS myocardium contained a senescent microvascular niche with endothelial cells, pericytes and YAP-high fibroblasts, consistent with hypoxic and mechanical stress. This senescent niche is similar to adult myocardial infarction but not pediatric dilated cardiomyopathy with heart failure, pointing to a prominent role of hypoxia in senescence. The microvascular senescent niche was improved by VAD, providing insight into the potential to reverse cardiac cell states that lead to heart failure.

The regional enrollment of participants in pivotal randomized controlled trials (RCTs) often does not represent the regional distribution of cardiovascular diseases. Over the past four decades, trials have enrolled participants primarily from North America and Europe, limiting the global generalizability of findings. In this Perspective, we review the evolution of regional participation in RCTs, using heart failure as a case study to assess temporal trends, current gaps in representativeness and opportunities for improvement. We assess the regulatory, logistical and financial barriers to clinical trial enrollment in underrepresented regions. We examine the manner in which global regions have been classified in trials, and propose a standardized regional classification system for reporting and subgroup analysis. To improve regional representativeness, we suggest targeted strategies that address barriers faced at the national, regulatory, sponsor or funder, institution and patient level. We also recommend the use of a representativeness index during trial planning and site selection to enhance regional representativeness. Expanding trial participation beyond historically dominant regions could be a key step in improving trial efficiency, external validity and global health equity.