Measuring polyethylene wear in total hip arthroplasty (THA) components using conventional radiographs is challenging for any measurement system because values less than 1 mm must be detected. Radiographic methods are essential because they are one of the few means of assessing prosthetic femoral head penetration into the acetabular cup to determine in vivo wear.20
Radiostereometric analysis (RSA) is the most accurate method for determining wear and femoral head penetration in vivo.10,11,13 However, the technique requires specialized radiographic equipment and implantation of tantalum beads into the patient, limiting its availability and practical use. Ein Bild Röntgen Analyse (EBRA) is a computer-aided method that determines in vivo wear by detecting prosthetic femoral head penetration into the acetabular component.10,11,14,22 Only one radiograph (an anteroposterior [AP] view of the pelvis) needs to be analyzed for a complete measurement at one time. Changes in head position are described as wear in the horizontal (X) and vertical (Y) directions. Ein Bild Röntgen Analyse also allows measurements of cup migration, abduction angle, and anteversion angle.10,14 Although less precise than RSA, EBRA has the potential to provide accurate wear information without the limitations of RSA.
Ilchmann et al investigated the precision of EBRA for measurements of femoral head penetration by comparing EBRA measurements with RSA measurements.10,11 However, because RSA was used as the reference measurement, the true femoral head penetration remained unknown.10,11 Wilkinson et al investigated EBRA using sets of clinical radiographs taken on the same day.22 The precision of the system was determined with the assumption that no displacements occurred.22 Although these studies10,11,22 provided an excellent foundation for initially validating the EBRA method, continued analysis of EBRA measurements under differing conditions is necessary.
To further validate the accuracy and precision of EBRA, it would be useful to precisely simulate femoral head displacement in vitro such that wear could be isolated and measured while controlling the variables that confound its measurement in clinical use. Using a wear-simulating device would allow for such an analysis. A wear-simulating device would allow for investigation of how onscreen image manipulation in EBRA might improve the accuracy of the digital measurement system.
We determined the accuracy of EBRA in measuring prosthetic femoral head penetration into the acetabular cup using a wear-simulating device and determined optimal conditions for performing these measurements in an experimental set-up. We sought to: (1) assess the effect of various onscreen-magnification factors used for measurements of head penetration; (2) identify the best combination of these factors to minimize measurement error and the 95% confidence intervals (CIs) in the X and Y directions; (3) determine the accuracy of EBRA in the X and Y directions using these optimal settings during controlled femoral head penetration in vitro; and (4) determine conditions for the least accurate measurements of head penetration and confidence intervals in the X and Y directions during controlled head penetration in vitro.
MATERIALS AND METHODS
To generate AP radiographs of a pelvis for EBRA analysis of precisely controlled femoral head movements, a previously developed femoral head migration simulator was used.18 The simulator was constructed from a postmortem-retrieved pelvis stripped of all soft tissue structures. The pelvic bone was aligned on the migration simulator so that the connecting line between the two anterior superior iliac spines was parallel to the base plate and the radiograph.18 The hemispheric acetabular component had an outer diameter of 52 mm. It was mounted without a liner together with a 32-mm metal ball head on a threaded rod with a thread lead of 1 mm. The cup and ball head were fixed on a device that allowed cup migration (together with the ball head) in 5 degrees of freedom (three translatory and two angles). The cup was positioned in 45° abduction and 15° anteversion. Femoral head displacement was simulated by advancing the femoral head on the threaded rod while the cup remained stationary. The displacements were performed in quarter turns (0.25 mm) and adjusted by markers at the ball head and the threaded rod. Head penetration was directed along the axis of the cup. To ensure that comparable radiographs were made, the simulator was placed in a stable upright position and all femoral head manipulations were performed in this orientation. Radiographs were made with the prosthetic femoral head in place at a 100-cm focus film distance.
Radiographs were digitized using the Diagnostix 2048 modular system (Fa GEMED, Ulm, Germany) and a digital camera. The images were saved uncompressed in tiff format, measuring 409.600 x 409.200 mm (2048 x 2048 pixels) using Corel Photo-Paint 8 (Corel Inc, Unterschleissheim, Germany). The file size was approximately 4 MB for each radiograph file. The measurements were performed on a 19-inch screen, working with 96 kHz and 0.26 pitch (Belinea, Fa Maxdata Computer GmbH and CoKG, Marl, Germany). The resolution was 1024 x 764 with 64,000 colors.
Images were burned on compact disks and measured on a computer using the EBRA software version 4.0 (Verfahren EBRA der Universität Innsbruck, Austria), which is based on the image manipulating software Optimate 6.51 (Fa Weiss Imaging and Solutions GmbH, Günding, Germany). A comparability algorithm in the EBRA software was used to compare and exclude any radiograph that contained more than an acceptable limit of measurement error from changes in pelvic positioning.14,22 This acceptable limit of positioning error, also called the comparability limit,14,22 was specified by the EBRA operator. The comparability limit was defined by the difference in measured distance between positioning landmarks, ranging from 1-4 mm.22 We used the most rigorous limit (1 mm). Using EBRA, all radio-graphs were found comparable, so none was excluded from analysis.
As suggested by EBRA, the obturator foramina (Fig 1) were chosen as landmarks when determining the pelvic reference line on the radiograph. Because of torsion in the anatomic specimen, this reference line was not exactly horizontal on the screen and was not parallel to the frame of the picture. This individual torsion was determined to be 2°, which was considered during the evaluation process (Fig 1).
All measurements were performed by one observer (AK) after performing several practice measurements to become familiar with the system. The manufacturer of EBRA software recommends various templates for measuring different types of pros-theses by schematically illustrating the outlined shapes of various types of hip endoprostheses. We used the Type 7 template. The system was calibrated using the femoral ball head with a 32 mm diameter. Its circumference was defined using five points, and the result was checked visually using a circle (Fig 1C).
The measurement of EBRA accuracy was separated into four investigations. In the first three investigations, the effect of magnification on accuracy was assessed by measuring observer error 10 times on the same radiograph using different magnification settings. In the first investigation (Series I), the pelvic reference lines were placed in the pelvis using a magnification factor of 50%, and the implant components were shown with a 100% magnification. In the second investigation (Series II), the pelvic reference lines and implant components were shown and measured with a 100% magnification. In the third investigation (Series III), the pelvic reference lines were shown at 100% and the implant components at 200% (Fig 1) (Table 1). When measuring 10 radiographs, nine values of femoral head penetration were calculated as the first image served as the reference. Deviations were recorded in relationship to the first image.
In the final investigation, the simulator was used to reproduce femoral head penetration directed along the axis of the cup from 0-1 mm in 0.25-mm increments (Series IV). This series of head penetration steps was simulated 10 times with radiographs taken at each position, resulting in a total of 50 radiographs for analysis. The radiographs illustrating the simulated femoral head penetration were measured according to the protocol for Series III. Because EBRA considers femoral head penetration in X and Y directions, evaluation of 0.25-mm intervals were decomposed into their X and Y directions to enable the set point versus actual value comparisons. The 2° individual specimen torsion was considered (Fig 1). The precision of positioning the femoral head while penetrating into the cup was estimated at ± 10°, which corresponded to approximately ± 0.03-mm linear femoral head movement in line with the threaded rod. For the X and Y coordinates, this equated to a precision of ± 0.02 mm in each direction.
Ten displacements of the femoral head allowed us to statistically evaluate the error in measuring femoral head penetration. Measurement error was defined as the difference between the mean value of head penetration measured via EBRA and the true penetration value known from the wear simulating device. The 95% CI was recorded under the premise that the measurement values were distributed normally.
The results of the measurements were entered into an EXCEL table (Excel 97 SR 2, Microsoft Corporation, Seattle, WA). Using the features of the EXCEL program, the means were determined and the 95% confidence intervals were calculated as ± 1.96 × SD/√n.17
The various magnification factors used for measurements had a distinct effect on accuracy (Table 1) (Fig 2), with the best accuracy recorded at the highest magnification factor. The accuracy when determining observer error using the highest magnification factor (Series III) was recorded as 0.056 mm for the X direction and 0.024 mm for the Y direction.
By increasing magnification it was also possible to reduce the 95% CIs for both directions. In Series III, 95% CIs were ± 0.013 mm for the X direction and ± 0.027 mm for the Y direction (Table 1) (Fig 2). All additional measurements were done using these settings.
The accuracy of EBRA in determining the X direction of the simulated femoral head penetration was from 0-0.029 mm. The 95% CIs were ± 0.035-0.067 mm (Table 2) (Fig 3). The accuracy of EBRA in determining the Y direction of femoral head penetration was from 0.001-0.013 mm. The 95% CIs were ± 0.046-0.079 mm (Table 2) (Fig 3). Although there was a trend for the software to underestimate the X direction, there was no trend of underestimating or overestimating the Y direction of femoral head penetration.
Femoral head penetration simulations showed the worst accuracy for EBRA (0.029 mm) in the X direction at a head penetration of 1 mm. As the femoral head penetration increased, the 95% CIs also increased to a maximum of ± 0.067 mm for the X direction and ± 0.079 mm for the Y direction. Assuming the worst case, and considering the maximum deviation of all measured values, the EBRA system could accurately record a linear femoral head penetration greater than 0.128 mm (calculated with a combined deviation of 0.096 mm for the X direction and of 0.084 mm for the Y direction).
Our main objective was to investigate the accuracy of the EBRA system under ideal conditions in vitro. This is a vital step toward error analysis of the EBRA system. There are several sources of error that influence total precision of an analytical system. In the clinical setting, sources of imprecision of EBRA measurements can be introduced by differences in patient positioning, radiographic technique, digitization method, the observer, or the EBRA system. Carefully controlled in vitro analysis is necessary to isolate and describe the imprecision of the method.
The major limitation of our study is that the in vitro results cannot be applied directly to the clinical situation. The in vitro situation created in this investigation was ideal: identical radiographic conditions, identical positioning of pelvic bone and implants for all radiographs, and excellent edge detection for the cup as the pelvic bone was stripped of all soft tissue. However, the uncontrolled nature of these variables in the clinical setting is certain to increase imprecision of the EBRA method and of all other methods in determining wear. This is the first study that describes the EBRA wear program under ideal conditions. Because our study is more similar to in vitro analyses of other radiographic techniques (Table 3) where confounding factors of the clinical environment have been excluded, it is more comparable than previous EBRA studies.10,11,22
Using in vitro analysis allowed us to observe the effect of magnification factors on the precision of EBRA measurements (Table 1) (Fig 2). It also allowed for specification of optimal settings, which minimized observer error on subsequent EBRA detection of femoral head penetration. We observed the best accuracy using a 100% magnification factor for the pelvic reference line and a 200% magnification factor for displaying the implant components. We found that the magnification had these upper limits because all the structures necessary for the evaluation steps had to be visible on the screen (Fig 1).
Ilchmann reported a measurement error of 0.19 mm for EBRA when an identical radiograph was measured several times (n = 25).10 This procedure corresponded closely to our procedure in Series I-III. We used the highest magnification factors in Series III and calculated a total deviation of 0.061 mm using the X and Y directions of femoral head penetration (√(0.0562 + 0.0242) (Table 1) for zero head movement. Compared with the values of Ilchmann, the use of magnification factors resulted in an improvement of 0.129 mm; corresponding to an improvement by a factor of three.10 Ilchmann et al reported a measurement error for femoral head penetration of 0.08 mm.11 Our measurement error ranged from 0.013-0.029 mm; 0.032 mm if the X and Y components were vectorally added together. Under ideal conditions femoral head penetration can be detected with a precision approximately 2.5 times greater than reported by Ilchmann et al.11 However, we used the highest possible magnification for our evaluations.
When assessing the precision of simulated wear measurements, we found the greatest deviations in accuracy at 1 mm of head penetration: −0.029 ± 0.067 mm (mean ± 95% CI) for the X direction and −0.001 ± 0.079 mm (mean ± 95% CI) for the Y direction (Table 2). These precision results are similar to the values reported for the RSA method.13 As reported by Kärrholm et al in a phantom study, the precision of RSA in determining a translational resolution was 0.007-0.025 mm.13 In another in vitro RSA study, Bragdon et al reported a precision of 33 μm in the medial direction and 22 μm in the cranial direction with 95% CIs of 8.4 μm and 5.5 μm, respectively.3 If these values are compared with values in our study, EBRA produces good results under optimum in vitro conditions. The smaller values for the 95% CIs as reported by Bragdon et al can be explained by their more precise testing equipment.3 This level of accuracy also compares favorably with the accuracies of other published methods (Table 3). The most comparable studies are those dealing with phantoms.2,4-7,12,15,16,19,21
The accuracy of EBRA in determining femoral head penetration as wear was increased by using magnification factors during the evaluation process on the screen. In the presence of ideal conditions and in the absence of distortions, the algorithm pinpoints variations in position between ball head and cup. Considering the worst accuracy for the X and Y directions, EBRA can accurately determine femoral head penetration greater than 0.128 mm under ideal conditions. Although the results are comparable with values reported in the literature for different systems, they are not applicable to the clinical situation.
Regarding the accuracy of EBRA applied clinically,22 the measurement error of a zero wear situation is similar to our phantom study, with the exception of larger confidence intervals in the clinical situation. However, our results indicate that using magnification factors on the screen can enhance the accuracy of EBRA wear measurements.
We thank Christi Sychterz Terefenko for revising the manuscript.
1. Barrack RL, Lavernia C, Szuszczewicz ES, Sawhney J. Radio-graphic wear measurements in a cementless metal-backed modular cobalt-chromium acetabular component. J Arthroplasty
2. Bragdon CR, Estok DM,Malchau H. Karrholm J, Yuan X, Bourne R, Veldhoven J, Harris WH. Comparison of two digital radiostereometric analysis methods in the determination of femoral head penetration in a total hip replacement phantom. J Orthop Res
3. Bragdon CR, Malchau H,Yuan X. Perinchief R, Karrholm J, Borlin N, Estok DM, Harris WH. Experimental assessment of precision and accuracy of radiostereometric analysis for the determination of polyethylene wear in a total hip replacement model. J Orthop Res
4. Collier MB, Kraay MJ, Rimnac CM, Goldberg VM. Evaluation of contemporary software methods used to quantify polyethylene wear after total hip arthroplasty. J Bone Joint Surg
5. Devane PA, Bourne RB,Rorabeck CH. Hardie RM, Horne JG. Measurement of polyethylene wear in metal-backed acetabular cups: I. three-dimensional technique. Clin Orthop Relat Res
6. Devane PA, Horne JG. Assessment of polyethylene wear in total hip replacement. Clin Orthop Relat Res
7. Ebramzadeh E, Sangiorgio SN,Lattuada F. Kang JS, Chiesa R, McKellop HA, Dorr LD. Accuracy of measurement of polyethylene wear with use of radiographs of total hip replacements. J Bone Joint Surg
8. Hall RM, Craig PS,Siney P. Wroblewski BM. Differences in the rates of penetration determined from radiographic and shadow-graphic measurements of acetabular sockets. J Arthroplasty
9. Hui AJ, McCalden RW,Martell JM. MacDonald SJ, Bourne RB, Rorabeck CH. Validation of two and three-dimensional radio-graphic techniques for measuring polyethylene wear after total hip arthroplasty. J Bone Joint Surg
10. Ilchmann T. Radiographic assessment of cup migration and wear after hip replacement. Acta Orthop Scand Suppl
11. Ilchmann T, Mjoberg B, Wingstrand H. Measurement accuracy in acetabular cup wear: three retrospective methods compared with Roentgen stereophotogrammetry. J Arthroplasty
12. Kang JS, Park SR, Ebramzadeh E, Dorr LD. Measurement of polyethylene wear in total hip arthroplasty: accuracy versus ease of use. Yonsei Med J
13. Karrholm J, Herberts P,Hultmark P. Malchau H, Nivbrant B, Thanner J. Radiostereometry of hip prostheses: review of methodology and clinical results. Clin Orthop Relat Res
14. Krismer M, Bauer R, Tschupik J, Mayrhofer P. EBRA: a method to measure migration of acetabular components. J Biomech
. 1995;28: 1225-1236.
15. Martell JM, Berdia S. Determination of polyethylene wear in total hip replacements with use of digital radiographs. J Bone Joint Surg
16. Martell JM, Berkson E, Berger R, Jacobs J. Comparison of two and three-dimensional computerized polyethylene wear analysis after total hip arthroplasty. J Bone Joint Surg
17. Ranstam J, Ryd L, Onsten I. Accurate accuracy assessment: review of basic principles. Acta Orthop Scand
18. Schwarz ML, Schroeder-Boersch H,Scheller G. Rappel B, Jani L. Simulator for migration processes and wear in the artificial hip acetabulum for radiologic measurement. [in German] Biomed Tech (Berl)
19. Shaver SM, Brown TD, Hillis SL, Callaghan JJ. Digital edge-detection measurement of polyethylene wear after total hip arthroplasty. J Bone Joint Surg
20. Sychterz CJ, Young AM, Engh CA. Effect of radiographic quality on computer-assisted head penetration measurements. Clin Orthop Relat Res
21. von Schewelov T, Sanzen L,Borlin N. Markusson P, Onsten I. Accuracy of radiographic and radiostereometric wear measurement of different hip prostheses: an experimental study. Acta Orthop Scand
22. Wilkinson JM, Hamer AJ,Elson RA. Stockley I, Eastell R. Precision of EBRA-digital software for monitoring implant migration after total hip arthroplasty. J Arthroplasty