An Optimization-Based Algorithm for the Construction of Cardiac Purkinje Network Models.

Objective: This work presents a new algorithm for the construction of a model for the Purkinje network (PN) of the heart.
Methods: The algorithm is based on a method called constructive constrained optimization (CCO), which was reformulated for the specific case of automatic PN generation. The proposed optimization-based algorithm is referred to as constructive optimization (CO). The CO method iteratively constructs the PN by minimizing the total length of the generated PN tree. In addition, it can take into account some important topological information of the PN, such as the location of the Purkinje-muscle junctions and the average bifurcation angle found in the literature.
Results: To validate the model, the new method was compared with the classical L-system method for generating PN models and to a recently proposed image-based technique.
Conclusion: The results show that the CO is able to construct PNs with geometric features and activation times that are in good agreement with those reported in the literature and to those obtained by the other aforementioned alternatives.