Acta Polytechnica Vol. 52 No. 5/2012 Femoral Test Bed for Impedance Controlled Surgical Instrumentation Christian Brendle1, Annegret Niesche2, Alexander Korff2, Klaus Radermacher2, Steffen Leonhardt1 1Philips Chair for Medical Information Technology, Helmholtz Institute for Biomedical Engineering, RWTH Aachen University, Pauwelsstrasse 20, 52074 Aachen, Germany 2Chair for Medical Engineering, Helmholtz Institute for Biomedical Engineering, RWTH Aachen University, Pauwelsstr. 20, 52074 Aachen, Germany Corresponding author: brendle@hia.rwth-aachen.de Abstract The risk for patients during the standard procedure of revision of cemented artificial hip joints is unsatisfactorily high due to its high level of invasiveness and limited access to the operative field. To reduce this risk we are developing an Impedance Controlled Surgical Instrumentation (ICOS) system, which aims to establish real-time control during a Bone Cement (BC) milling process. For this, the relationship between the thickness of the BC and its frequency-dependent electrical impedance is used to estimate the residual BC thickness. The aim is to avoid unintended cutting of bone by detecting the passage of the BC/bone boundary layer by the milling head. In a second step, an estimation of the residual BC thickness will be used to improve process control. As a first step towards demonstrating the feasibility of our approach, presented here are experimental studies to characterize the BC permittivity and to describe the process in detail. The results show that the permittivity properties of BC are dominated by its polymethyl methacrylate (PMMA) fraction. Thus, PMMA can be used as a substitute for future experiments. Furthermore, a Femoral Test Bed (FTB) was designed. Using this setup we show it is feasible to accurately distinguish between slightly different thicknesses of BC. Keywords: Bioimpedance spectroscopy, surgical instrumentation, artificial hip replacement, Bone Cement milling. 1 Introduction In industrialized countries procedures such as total hip replacement are frequently performed to improve quality of life and maintain mobility of the aging soci- ety. For example, in Germany in 2008 ≥ 150000 pri- mary total hip replacements were implanted and the number of cases is increasing [2]. The major part of the implanted artificial hip joint is cemented [6, 10]; however, a systematic review showed that only 88– 95% of the cemented artificial hips are revision-free after 10 years [4]. This implies that hip revision surgery has now become a standard procedure. For example, in Germany alone, in 2008, ≥ 22500 revi- sions of total hip arthroplasties took place [3]. During revision surgery, the old BC has to be re- moved to ensure fixation of the new implant [9]. Gen- erally this is done by manual control of surgical ham- mer and chisel. However, due to restricted access to the operative field, lateral femur windowing may be used to help resolve this problem. Unfortunately, these standard procedures are highly invasive and the risk of unintended damage to the bone or vessels is very high [14]. An alternative approach is to use milling tools for removal of BC. To acquire information on the spatial characteristics of BC, which is necessary for tool nav- igation, ultrasound or computed tomography (CT)- ~ IV Figure 1: Impedance Controlled Surgical Instru- mentation: planned operative scenario. based measurements are used. This requires an ad- ditional step before or during the operation, apart from the actual milling process [6, 5, 8]. However these additional steps cause stress, and CT imaging exposes the patient to ionizing radiation. To avoid these disadvantages, our approach uses real-time control of the milling process during BC removal. Figure 1 presents the planned surgical sce- nario. Through the milling head, an alternating cur- 17 Acta Polytechnica Vol. 52 No. 5/2012 rent is applied to the patient and the voltage between the milling head in the femur and the patient’s leg skin is measured. The measured values are used to calculate the spectral two-point bioimpedance from which we want to deduce the thickness of the resid- ual BC in the femur. 2 Permittivity characterization of Bone Cement BC consists mainly of PMMA. The ICOS system utilizes the high electrical impedance of PMMA, which is several orders of magnitude higher than the impedance of human tissue. To illustrate this, we compare the impedance fractions of the different lay- ers of tissue between the femur center and the skin surface of the leg using a two-point Bioimpedance Spectroscopy (BIS) measurement. Body tissue with the highest impedance fraction is dry skin, despite its relatively small thickness. Comparison of the abso- lute impedance values of two commensurate probes out of PMMA and dry skin shows that the PMMA value of 110 kHz is 400 times higher than the value of dry skin [11, 1]. Thus, the BC impedance presents the major part of the measured impedance in the ICOS system. To verify the characteristic BC permittivity and the relation between its thickness and electric impedance we analyzed probes from Biomet R© Bone Cement R using a capacitor measurement setup. The BC capacitor impedance was measured with an Ag- ilent E4890A LCR Meter in the range of 2 kHz to 2MHz. To calculate characteristic BC permittivity values, from the measured values we used eq. (1) (with sample dimensions: side length a = 10mm and thickness d = 1, 2, and 3mm). YBC = G+ j ·B = a2 d · j · ω · ε0 · (ε′r − j · ε′′r ) =⇒ ε′r = B · d a2 · ω · ε0 ε′′r = G · d a2 · ω · ε0 (1) Figure 2 presents the characteristic BC permittiv- ity. The red curves represent the mean values of 36 independent measurement series and the grey area in- dicates the standard deviation for each frequency. ε′r represents the stored energy and the ε′′r characterizes the loss of energy within the medium. The diagram confirms that BC dielectric properties are character- ized by the PMMA’s properties. In the complex εr- plane the curve of the measured BC (red) is shifted about 10 % to the right relative to the values reported in the literature for PMMA (black) [11]. This 10 % shift can also be observed in the frequency dependent [11] Figure 2: Permittivity measurements of BC (red) with standard deviation (grey) according to McCrum et al. permittivity values for PMMA (black) [11]: Complex permittivity plane (top), real part of the permittivity ε′r over frequency f (middle), imaginary part of the permittivity ε′′r over frequency f (bot- tom). illustration of ε′r in Figure 2 (middle band). The ε′′r of BC values are in accordance with the published values for PMMA. Based on these measurements, it is acceptable to use PMMA dummies for our experiments to demon- strate the functionality of the ICOS system and to describe the actual process. 3 Design of Femoral Test Bed The FTB presented in Figure 3 is used as a repre- sentation of the scenario for the adult human leg. Thus, the geometry of the FTB is defined as follows. The basis is a cylindrical bowl with an inner diame- ter of 15 cm as an approximation of thigh diameter. The relevant volume of the impedance measurement is determined by the path of the electrical current. For this reason, the inner height of the bowl is set at 4 cm. The anatomical femur position is 3.25 cm off-center. However, for the first measurements the center position is required for geometric invariance. Therefore, a carrier for the PMMA dummy (Figure 4, green) can be mounted in the middle of the bowl and 3.25 cm off-center. The dummy, representing the BC, is a hollow cylinder with varying wall thicknesses (1, 1.5, 2, and 2.5mm) in the upper part. In addition 18 Acta Polytechnica Vol. 52 No. 5/2012 Figure 3: FTB with 16 T-shaped electrodes of the inner wall. Figure 4: Cross-sectional view of the measurement setup: bowl (grey), T-shaped electrodes (blue), car- rier (green), dummy plate (yellow), PMMA dummy (red) and point electrode (cyan). a dummy plate (Figure 4, yellow) for the free carrier position is provided. The space between the dummy and bowl wall is filled with a conductive material as a substitute for human tissue. For these experiments we use NaCl solutions with an adjusted conductivity. For fu- ture experiments agar-agar gel compositions will be used [13]. These gels acquire a solid structure after a certain curing time. During the casting process of agar-agar gel, its temperature can rise to 80 ◦C. This has to be considered in the choice of bowl ma- terial. Moreover, the bowl needs to be a good isola- LCUR HCURLPOT HPOT Figure 5: Wiring diagram for the first BIS measure- ment: the two-terminal method. tor to avoid unwanted effects on the measurements. Therefore, polyoxymethylene (POM) is a good choice for the bowl material due to its negligible electrical conductivity and its thermal stability in the required temperature range [7]. Furthermore, the T-shaped electrodes in the wall should be replaceable for up- dates in the experimental setting. In addition, we need 16 uniformly distributed electrodes to be able to use this FTB in Electrical Impedance Tomography (EIT) experiments. The first measurement setup will be BIS measure- ments between the point electrode (Figure 4, cyan) in the middle, and one T-shaped electrode (Figure 4, blue). The bowl will be filled with NaCl solution with a conductivity of 850µS/cm, similar to the conduc- tivity of cancellous bone [1]. 3.1 Electrical contacting The Agilent E4890A LCR Meter has four BNC con- nectors used for impedance measurements, two for the low connector pair and two for the high connector pair. In each case, one of these two BNC connectors is used to apply the measurement current and the other for the electrical potential measurement. The LCR Meter is connected to the FTB using a two- terminal measurement configuration (Figure 5) with four 0.5m coaxial cables. The current and the poten- tial cable to the FTB electrodes are connected using BNC T-Connectors (Figure 5, grey). The cladding contacts of the two BNC T-Connectors are connected to limit the effects of measurement incidents. The measurement range is set between 200 kHz and 2MHz. The LCR Meter low connector pair is con- nected to one T-shaped electrode (Figure 4, blue) and the high connector pair is connected to the point elec- trode (Figure 4, cyan). This is realized using screwed solder terminal lugs, which are connected to the inner 19 Acta Polytechnica Vol. 52 No. 5/2012 Two-sample t-test p-value Significant difference |Z1,0mm| < |Z1,5mm| 0.38 % yes |Z1,5mm| < |Z2,0mm| 1.21 % yes |Z2,0mm| < |Z2,5mm| 5.06 % no Table 1: Two-sample t-test with significance level α = 5%. conductor of the two BNC T-connectors (Figure 5, grey). 4 First results The impedance curves in the Z-plane (Figure 6, top) show the capacitive characteristic of the PMMA dummy wall. Assuming a similar scale for both dia- gram axes, the curves are almost parallel to the neg- ative y-axis and aspire to the origin of the Z-plane with increasing frequency. This characteristic is con- firmed by the decrease of the absolute value of the mean impedance value from four measurement se- ries over the frequency (Figure 6, bottom left). This plot also shows that our measurement range extends to MΩ values. Figure 6 (bottom right) presents the mean impedance with standard deviation at 2MHz as a function of the wall thickness. This plot shows that the probability distributions of the measurement series for different thicknesses intersect within their range of the single standard deviation for greater wall thicknesses. With smaller wall thicknesses the relative standard deviation decreases from 8.31% at 2.5mm to 4.34% at 1mm. To characterize the sensitivity of the first experi- ments, a two-sample t-test was performed with the absolute impedance values at 2MHz. For this we used a left-tail test taking into account that the mean impedance of smaller wall thickness is lower [12]. In addition, we assume unequal variances. The results of the t-test (Table 1) show that the measured differ- ence of the impedance values is significant for smaller wall thicknesses. Due to the higher standard devia- tions for greater wall thicknesses it is not possible to show significant differences. 5 Conclusion The FTB represents the surgical operative setting. It allows repeatable arrangements of experimental set- tings with PMMA dummies as a substitute for the BC. For medical application, the sensitivity of the ICOS system at smaller wall thicknesses (< 2mm) is important because the risk of damage to patient tis- sue is increased in this range. The results of the first Figure 6: Impedance measurements in the FTB with a hollow cylindrical PMMA dummy with dif- ferent wall thicknesses d. Top: Z-plane with mean impedance curves. Bottom left: Absolute value of the mean impedance as function of the frequency. Bottom right: Absolute impedance mean values with standard deviation at 2MHz as a function of wall thickness. experiments with a non-rotating point electrode in- dicate the suitability of the measurement system for this purpose. For more valid values, a larger sample size is required. In the future, additional experiments will support further development of the ICOS system. For exam- ple, a theoretical model of the electrical conduction in the setup is needed which also takes into account the effects of surface conduction. In addition, the influ- ence of current transmission on the rotating milling shaft and the transmission from the rotating milling head to the BC needs to be characterized. Acknowledgements The authors thank the Federal Ministry of Education and Research for financial support of the research project Impedance Controlled Surgical Instrumenta- tion (07EZ1005). 20 Acta Polytechnica Vol. 52 No. 5/2012 References [1] R. Andreuccetti, D.and Fossi, C. Petrucci. Di- electric properties of body tissues: in the fre- quency range 10 Hz–100 GHz, 1997–2007. http: //niremf.ifac.cnr.it/tissprop/, [2012-09- 10]. [2] O. Boy, S. Hahn, E. Kociemba. Hüft Endoprothesen Erstimplantation. In BQS- Qualitätsreport, pp. 146–153. 2008. [3] O. Boy, S. Hahn, E. Kociemba. Hüft-Endo- prothesenwechsel und -komponentenwechsel. In BQS-Qualitätsreport, pp. 154–160. 2008. [4] Kelly L. Corbett, Elena Losina, Akosua A. Nti, et al. Population-based rates of revision of pri- mary total hip arthroplasty: A systematic re- view. PLoS ONE 5(10):e13520, 2010. [5] R.E. Ellis. From scans to sutures: Computer- assisted orthopedic surgery in the twenty-first century. In Engineering in Medicine and Biology Society, 2005. IEEE-EMBS 2005. 27th Annual International Conference of the, pp. 7234–7237. 2005. [6] M. de la Fuente, J.A.K. Ohnsorge, E. Schkom- modau, et al. Fluoroscopy-based 3-d reconstruc- tion of femoral bone cement: a new approach for revision total hip replacement. Biomedical Engi- neering, IEEE Transactions on 52(4):664–675, 2005. [7] Jurima Dichtungen GmbH. POM – C Daten- blatt, 2012. http://www.jurima-gmbh.de/ downloads/pomcdatenblatt.pdf, [2012-09-10]. [8] S. Heger, T. Vollborn, M. Niggemeyer, et al. Robot integrated ultrasound geometry-scanning for trackerless bone cement detection in rthr. In Ultrasonics Symposium (IUS), 2010 IEEE, pp. 845–848. 2010. [9] P. L. S. Li, P. J. Ingle, J. K. Dowell. Cement- within-cement revision hip arthroplasty; should it be done? Journal of Bone & Joint Surgery, British Volume 78-B(5):809–811, 1996. [10] Henrik Malchau, Peter Herberts, Peter Söder- man, Anders Odén. Prognosis of total hip re- placement: Update and validation of results from the swedish national hip arthroplasty reg- istry 1979–1998. In 67th Annual Meeting of the American Academy of Orthopaedic Surgeons, Orlando, USA, pp. 1–16. 2000. [11] N. G. McCrum, B. E. Read, G. Williams. Anelastic and Dielectric Effects in Polymeric Solids. John Wiley & Sons Ltd, 1967. [12] MathWorks R©. Matlab product doc- umentation: Two-sample t-test, 2012. http://www.mathworks.de/help/toolbox/ stats/ttest2.html, [2012-09-10]. [13] Tong In Oh, Hwan Koo, Kyung Heon Lee, et al. Validation of a multi-frequency elec- trical impedance tomography (mfEIT) system KHUMark1: impedance spectroscopy and time- difference imaging. Physiological Measurement 29(3):295, 2008. [14] J. Pfeil. Hüftchirurgie. 1st ed., Steinkopff, 2008. 21