An Artificial Neural Network Approach to the Predictive Modeling of Tensile Force during Renal Suturing

Abstract

Further understanding of the mechanical deformation and tear behavior of kidney during suturing helps to enhance surgical task execution that requires fine suture manipulation. This paper aims to develop a model to describe the relationship between the tensile force acting on suture line and the resulted kidney deformation in the tension of suture line. The tensile force was recorded during multiple suture procedures, filtered, and we extracted the force data between the beginnings of tensioning suture line and tearing renal remnant. The extracted data were interpolated to obtain the same data sample length for each suture procedure, and then used to train the back propagation (BP) neural network. In order to predict the tearing force, force data were collected in sequence and interpolated to be input into BP network. Suture processes were performed on in vitro porcine kidneys, and experimental results verified the effectiveness of predictive modeling and the accuracy of tearing force prediction. The relative error for tearing force prediction was about 15%. The predictive modeling method can be used to forecast the tearing force before the suture line tears through soft tissues.