DeepPlanner4Cardio: an automatic multi-view planning tool for CMR

Motivation: Manual planning of the 4 standard cardiac views (2-chamber, 3-chamber, 4-chamber, short axis) is currently a time-consuming process that demands significant operator interaction and expertise. Goal(s): Our aim is to streamline and automate this process to reduce examination times and minimize dependency on highly trained operators. Approach: We developed DeepPlanner4Cardio, a deep learning model designed to predict the standard cardiac views simultaneously. Results: DeepPlanner4Cardio achieved a clinically acceptable accuracy of 94% on the test set with a mean orientation error of 11º and a mean displacement error of 5.9 mm, both within the inter-operator variability ranges. Impact: DeepPlanner4Cardio effectively supports CMR operators by providing an accessible, automated solution to enable fast and reproducible cardiac view planning, demonstrating great potential to be applied in a clinical setting.