Possibilities of Neural Networks for Personalization Approaches for Prevention of Complications After Endovascular Interventions

It is known that most of the diseases of the cardiovascular system are accompanied by disorders in the hemostatic system. The hemostatic system is one of the most complex systems. It has a hierarchical structure with a plurality of components. We analyze the results of thrombin generation test (TGT) which allows of estimating the actions of all components of the hemostatic system. The problem is complicated by the presence of too many various clinical cases. The simple statistical methods do not provide global assessments. We suggest the universal neural network approach for building hemostatic system models based on the factors which don’t have a statistically significant difference for various types of clinical post surgery cases. The neural network instruments allow of taking into account the nonlinear hierarchical nature of considered system and building individual models for each clinical cases. The aim of our study is to develop the neural network hemostatic system model for forecasting of disease progression and complications after endovascular interventions.

[1]  Ekaterina A. Blagoveshchenskaya,et al.  Neural Network Methods for Construction of Sociodynamic Models Hierarchy , 2016, ISNN.

[2]  Tatiana V. Lazovskaya,et al.  Neural network modeling of conditions of destruction of wood plank based on measurements , 2016 .

[3]  Audrey Borghi-Silva,et al.  Artificial Neural Networks Prognostic Evaluation of Post-Surgery Complications in Patients Underwent to Coronary Artery Bypass Graft Surgery , 2009, 2009 International Conference on Machine Learning and Applications.

[4]  Dmitriy A. Tarkhov,et al.  Basis Functions Comparative Analysis in Consecutive Data Smoothing Algorithms , 2016, ISNN.

[5]  H. Hemker,et al.  Continuous Registration of Thrombin Generation in Plasma, Its Use for the Determination of the Thrombin Potential , 1993, Thrombosis and Haemostasis.

[6]  H. Coenraad Hemker,et al.  Calibrated Automated Thrombin Generation Measurement in Clotting Plasma , 2003, Pathophysiology of Haemostasis and Thrombosis.

[7]  Moncef Gabbouj,et al.  Evaluation of Global and Local Training Techniques over Feed-forward Neural Network Architecture Spaces for Computer-aided Medical Diagnosis , 2022 .

[8]  A. Gori,et al.  Residual thrombin potential predicts cardiovascular death in acute coronary syndrome patients undergoing percutaneous coronary intervention. , 2016, Thrombosis research.

[9]  R. van Oerle,et al.  The hypercoagulable profile of patients with stent thrombosis , 2015, Heart.

[10]  I. Bolgov,et al.  Neural network model of rupture conditions for elastic material sample based on measurements at static loading under different strain rates , 2016 .

[11]  Tatiana V. Lazovskaya,et al.  Neural Network Technique in Some Inverse Problems of Mathematical Physics , 2016, ISNN.

[12]  Siti Mariyam Hj. Shamsuddin,et al.  Enhancement of artificial neural network learning using centripetal accelerated particle swarm optimization for medical diseases diagnosis , 2014, Soft Comput..

[13]  Sara Moein.  Medical Diagnosis Using Artificial Neural Networks , 2014 .

[14]  Martin A. Riedmiller,et al.  A direct adaptive method for faster backpropagation learning: the RPROP algorithm , 1993, IEEE International Conference on Neural Networks.

[15]  Daniel S. Levine,et al.  Neural Network Modeling , 2002 .

[16]  Tom Fawcett,et al.  An introduction to ROC analysis , 2006, Pattern Recognit. Lett..