A few years ago, we started a project and collaboration between Harvard (Prof. Schwab and Dr. Ghaednia), MIT (Prof. A. John Hart) and Purdue (Prof. Tyler Tallman) before we joined Cedars Sinai to develop the tools for early detection of implant loosening in orthopaedic joint replacements. Currently we are continuing this work at Cedars Sinai and CSIE. In this system we developed a piezo resistive bone cement by imparting conductive fibers such as carbon micro and nano fibers and silver nano wires into the PMMA bone cement with a polymer used to attach implants to bone. Around 1-1.5% volume ratio of fibers to polymer we get to the percolation point where even small deformations cause large electrical conductivity changes. Now assume we have this bone cement in the joint. As the patient moves and applies force on the joint the mechanical deformations on the cement cause electrical conductivity changes, therefore the cement acts as a sensor. We then integrate this with Electrical Impedance Tomography where through skin electrodes we construct conductivity maps of the domain. Therefore, we can monitor the interface condition using a non-invasive wearable. This started a series of projects that led to applications of machine learning on EIT signals better post-processing of EIT signals and finally EIT for muscle monitoring (Zhu et al 2021, Zhu et al. 2022, Zhu et al., 2022, Zhu et al., 2021, Zhu et al., 2021). The following is the story of our projects on screening implant failure (Ghaednia et al., 2019, Ghaednia et al. 2020, Keiderling 2023 ).
Non-invasive monitoring of implant-bonn cement interface:
Similar to interface failure between any two materials, the mechanical failure for cemented joint replacements initiates with the propagation of micro-cracks within the bone cement caused by contact/interfacial stresses. The contact stresses are at a maximum beneath the interface (within the cement) rather than at the interface. This causes the initiation of micro-cracks within the bone cement volume as illustrated in Fig. 1. Even though at this stage the defects are not detectable by conventional diagnosis devices, the stress distribution within the bone cement is altered by the micro-cracks. The accumulation and propagation of these micro cracks continues and can yield two failure mechanisms: i) when cracks reach the interface and cause loss of fixation (aseptic loosening) or ii) as the micro-cracks merge they create larger cracks and hence breakage of the fixation. At this stage the defects are visible however the failure is already in advanced stages and will require revision. It is for these reasons that traditional imaging modalities such as plain radiographs, CT scans, arthrography, and nuclear imaging are largely limited to diagnosis of mechanical failures in advanced stages. This is why we decided to create a new system for early detection of implant failure
Therefore, we developed the idea we talked about earlier to integrate the piezoresistive bone cement with EIT. We then developed a test setup where we 3D printed samples that simulated the implant-bone interaction and used a mechanical loading device to apply compressive loading to the samples.
While applying the compressive loads we simultaneously performed the EIT measurements and constructed time difference conductivity maps. The figure on the right shows the conductivity maps as the loading increases for 3 different volumetric ratios of the conductive fibers.
We continued this work by applying machine learning on the data and shows ML is a viable method for understanding the EIT signals! For more details on this work make sure to check out our papers (Ghaednia et al., 2019, Ghaednia et al. 2020, Keiderling 2023 )
Ongoing work:
We are currently very interested in applications of EIT in injuries and muscle monitoring. We will be sharing some new findings as soon as we are ready but if you are interested in this area feel free to reach out.