A two-step deep learning approach for patent foramen ovale diagnosis based on spatiotemporal sequence information of contrast transthoracic echocardiography

Mengjie Gu, Yingying Liu, Lin Wang, Yuanyuan Sheng, Junlong Zhu, Junqiang Yan, Mingchuan Zhang, Qingtao Wu

Abstract: Highlights•Clinical diagnosis of PFO is time-consuming and often affected by the doctor’s experience.•Deep learning has brought brand-new ideas and infinite possibilities to the diagnosis of PFO.•We implemented PFO diagnosis to help junior sonographers through a two-step intelligent diagnosis approach.