Method for selection of femoral component in total knee arthroplasty (tka)

Abstract

A method is proposed enabling a surgeon to preoperatively determine the preeminent type and size of prosthesis, from those available, to be used in a particular patient undergoing knee replacement surgery. Parameters of healthy knee geometry were estimated by employing an unsupervised neural network. These estimated parameters were then applied in a χ(2) goodness of fit (GoF) test to determine which femoral prosthesis type and size delivers the most appropriate fit. This approach was used to determine the most suitable match of three implants for 34 different cases. Implant C performed the best and was the optimal fit in 59% of the cases, Implant A was the best fit in 38% of the cases and Implant B the best fit in 3% of the cases. This method shows promise in aiding a surgeon to select the optimal prosthesis type and size from an array of different conventional total knee replacements.