Figure 1 | Scientific Reports

Figure 1

From: Predicting long-term time to cardiovascular incidents using myocardial perfusion imaging and deep convolutional neural networks

Figure 1

Schematic diagrams of this research. (A) The flowchart of patient selection for this study. (B) Schematic diagram of the proposed end-to-end survival training architecture designed to evaluate the risk of patients with respect to their subsequent time to cardiovascular incidents. Myocardial perfusion images presented as a series of two-dimensional gray-scale slices which are perpendicular to the short, long-vertical and long-horizontal axes of the heart. This 2D tomographic presentation is defined as the input format of our MPI AI model, trained end-to-end with optimization functions related to the survival analysis. The risk scores derived from MPI AI model reflect patients’ subsequent outcome. Patient strata by the risk score can be analyzed using Kaplan–Meier plots.

Back to article page