Neural Network as an Assisting Tool in Designing Talus Implant

Abstract

The design of current talus implant are focusing too much on mechanical simplicity and usually based on certain population which tends to ignore the anatomically difference between populations. An anatomically talus implant design is known can reduce the contact pressure but one of the constraints for designing implant anatomically is to get bone parameters. This is due to the difficulty to get enough volunteers in getting bone parameters using hazardous method (X-ray or CT scan) .Thus, the talus implant (TI) for particular population was developed based on artificial neural network (ANN) prediction. By using Finite Element Method (FEM), numerical models that include mainly the talus bone and the talus implants are created to compare the contact pressure distribution of the newly develop talus implant with the three different kind of current talus implant designs (BOX, STAR & TNK). For FEM results, only BOX and the newly develop talus implant exceeded the contact stress recommended for the superior articular surface compared to the others. The results also showed that the stress increased near the resected surface. Thus, it is agreed that excessive bone resection may not support the force at the ankle which consequently may contribute to early loosening and subsidence of the talus implant. It is concluded that the excessive bone resection can be avoided by perfectly match talus implant which only can be achieved by designing talus implant for a particular population.