Electrical Impedance Tomography for Hip Stem Implant Monitoring

Abstract As a first step towards a new method for noninvasive hip-stem implant monitoring, this work aims to detect bone defects from time- and frequency-difference electrical impedance tomography. As a proof-of concept, an in-silico model of a thigh with bone and implant is used, and convolutional neural networks are applied to predict size and position of areas where conductivity of the bone is increased. From both, time- and frequency-difference voltages, defect size and position can be predicted with high accuracy if signal-to-noise ratios are sufficiently high. This shows the potential for external impedance measurements using electrodes on a patient’s thigh to detect bone loss in the femur.