Radiographic Analysis of Femoroacetabular Impingement with Hip 2 Norm Reliable and Validated

Transcription

1 Radiographic Analysis of Femoroacetabular Impingement with Hip 2 Norm Reliable and Validated Moritz Tannast, 1 Sapan Mistry, 1 Simon D. Steppacher, 1 Stephan Reichenbach, 2 Frank Langlotz, 3 Klaus A. Siebenrock, 1 Guoyan Zheng 3 1 Department of Orthopaedic Surgery, Inselspital, University of Bern, Freiburgstrasse, 3010 Bern, Switzerland, 2 Department of Social and Preventive Medicine, University of Bern, Bern, Switzerland, 3 MEM Research Center for Orthopaedic Surgery, Institute for Surgical Technologies and Biomechanics, University of Bern, Bern, Switzerland Received 3 July 2007; accepted 17 January 2008 Published online in Wiley InterScience ( DOI /jor ABSTRACT: The purpose of this study was to validate the accuracy, consistency, and reproducibility/reliability of a new method for correction of pelvic tilt and rotation of radiographic hip parameters for pincer type of femoroacetabular impingement on an anteroposterior pelvic radiograph. Thirty cadaver hips and 100 randomized, blinded AP pelvic radiographs were used for investigation. To detect the software accuracy, the calculated femoral head coverage and classic hip parameters determined with our software were compared to reference measurements based on CT scans or conventional radiographs in a neutral orientation as gold standard. To investigate software consistency, differences among the different parameters for each cadaver pelvis were calculated when reckoned back from a random to the neutral orientation. Intra- and interobserver comparisons were used to analyze the reliability and reproducibility of all parameters. All but two parameters showed a good-to-very good accuracy with the reference measurements. No relevant systematic errors were detected in the Bland Altman analysis. Software consistency was good-to-very good for all parameters. A good-to-very good reliability and reproducibility was found for a substantial number of the evaluated radiographic acetabular parameters. The software appears to be an accurate, consistent, reliable, and reproducible method for analysis of acetabular pathomorphologies. ß 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 26: , 2008 Keywords: hip; femoroacetabular impingement; radiographic analysis; computer method Pelvic tilt and rotation are a natural component of the patient s posture, but are difficult to assess during radiographic examination. 1,2 The individual orientation of the pelvis can vary significantly among patients 3,4 and over time. 5 Consequently, the radiographic appearance of the hip on planar x-rays is strongly dependent on pelvic orientation. 1,6,7 It was suggested that radiographic examinations of the pelvis should be standardized to ensure a neutral starting point and reproducible readings, particularly in preoperative planning of jointpreserving surgical interventions and follow-ups of patients or in epidemiological and clinical studies. 1,8 Most available computerized methods for correction of radiographic parameters of the native hip were developed for evaluation of cranio-caudal femoral head coverage in hips with a deficient acetabular roof allowing calculation of the amount of correction achieved with acetabular rotational osteotomies However, none of them represents an appropriate tool for the evaluation of hips with pincer type of femoroacetabular impingement (FAI), which was recently identified as a very common mechanical cause of early primary osteoarthritis. 13 Pincer FAI is a focal or general overcoverage of the femoral head with early pathologic linear contact between the too prominent acetabular rim and the head neck junction. The main reasons for the inability to use existing methods comprise the lack of thorough validation, 9 the inability to correct for pelvic tilt 9,11 or rotation, 9 12 and the use of pelvic tilt indicators that Additional Supporting Information may be found in the online version of this article. Correspondence to: Moritz Tannast (T: ; F: ; moritz.tannast@insel.ch) ß 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. could be proven unreliable. 4 In particular, none provides FAI specific parameters such as the calculation of head coverage in the AP direction or the determination of the retroversion sign and index, respectively. Based on comprehensive cadaver trials and theoretical error analysis, a noncommercial software, Hip 2- Norm, was recently developed for anatomically based correction of tilt and rotation of pelvic radiographs in hips with FAI and associated calculation of relevant radiographic parameters for describing hip morphology. 2,6,14 The aim of this study was to perform a thorough validation of Hip 2 Norm. We hypothesized that this software represents an accurate, consistent, reliable, and reproducible tool for the correction of selected and important radiographic hip parameters. METHODS Radiographic Technique In our department, AP pelvic radiographs are taken in supine position. Since pelvic orientation differs in supine and standing position, this has the advantage that intraoperative and immediately postoperative radiographs can be compared reliably with preoperative x-rays. The source-to-image distance is 120 cm. The central beam is directed to the midpoint between the symphysis and the center between both anterior superior iliac spines. 4,6 The legs are rotated about 158 internally to compensate for femoral antetorsion. An additional one-time strong lateral radiograph 8,14 is centered on the cranial tip of the greater trochanter without patient repositioning resulting in a simultaneous lateral view of the pelvis allowing calibration for individual pelvic tilt as described later. Hip 2 Norm Software Hip 2 Norm is previously developed noncommercial software for standardized evaluation of radiographic parameters for description of acetabular morphology on AP pelvic radiographs. 2,14 Based on a spherical projection and on extended 1199

2 1200 TANNAST ET AL. Figure 1. The Hip 2 Norm software calculates the 3D configuration of the hip joint out of an anteroposterior (AP) pelvic x-ray based on a cone projection. Individual pelvic malpositioning can be corrected with the help of the vertical/horizontal distances between the symphysis and the sacrococcygeal joint for tilt/rotation. An additional one-time lateral radiograph is necessary to calibrate the vertical distance a. cadaver calibration experiments, it allows the correction of the projected acetabular rim and associated important parameters for individual pelvic tilt and rotation. The geometrical model comprises the assumption that all points of the acetabular rim are lying on a sphere (Fig. 1). The mathematical background has been described. 14 Two linear distances are used to estimate tilt and rotation. The vertical distance between the sacrococcygeal joint and the superior border of the symphysis ( a ) has to be calibrated with a lateral pelvic radiograph where pelvic inclination is measured. Pelvic inclination describes the angle between a horizontal line and a line connecting the sacral promontory and the pubic symphysis. Once the individual correlation is established, the tilt of any subsequent or previous radiograph of the pelvis of the same patient can be calculated prospectively and retrospectively because of the intraindividual linear correlation between a and pelvic tilt. 2,6 In comparison to other pelvic tilt parameters, a represents the most reliable predictor. 4 However, as could be shown in a theoretical error analysis, due to variation in individual sacral morphology, the calibration with a lateral radiograph is necessary because small variations can lead to significant changes in the appearance of acetabular morphology and femoral head coverage. 2 For pelvic rotation, the horizontal distance of the sacrococcygeal joint and the symphysis ( b, Fig. 4) has been proven a reliable indicator. 2 The neutral orientation is defined by a pelvic inclination angle of 60 degrees 15,16 and b of 0 cm. As a horizontal reference, the interteardrop line is used. The software then provides the user with a set of parameters (Table 1). Validation Study The validation study to analyze Hip 2 Norm was divided into three subparts: (1) software accuracy; (2) software consistency; and (3) inter-/intraobserver analysis. Software Accuracy Software accuracy was determined by comparing the calculated measurements from Hip 2 Norm on study radiographs with the reference measurements obtained from AP radiographs taken in neutral rotation, false profile radiograph, and CT scan ( reference radiographs, Fig. 3). Ten cadaveric pelves (20 hips) with no macroscopic abnormalities were obtained. In both hips, the acetabular rims were marked with 1-mm metal Table 1. Definition of Evaluated Parameters Parameters Cranio-caudal coverage Anterior coverage Posterior coverage Lateral center edge angle (LCE) Acetabular index ACM-angle Extrusion index Cross-over sign Retroversion index Posterior wall sign Anterior center edge angle (ACE) Definition The percentage of femoral head covered by the acetabulum in cranio-caudal direction The percentage of femoral head covered by the acetabulum in AP direction The percentage of femoral head covered by the acetabulum in postero-anterior direction Angle formed by a line parallel to the longitudinal pelvic axis and by the line connecting the center of the femoral head with the lateral edge of the acetabulum Angle formed by a horizontal line and a tangent from the lowest point of the sclerotic zone of the acetabular roof to the lateral edge of the acetabulum Angle constructed by the following points: (A) superolateral acetabular edge, (M) midpoint of a line connecting the superolateral and the inferolateral acetabular edge, (C) point of the bony acetabulum intersected by a perpendicular line relative to line AM through point M Percentage of uncovered femoral head in comparison to the total horizontal head diameter Morphologic acetabular variation where the anterior acetabular rim is projected more laterally than the posterior rim in the cranial part of the acetabulum Quotient between the length of overlap of the anterior rim in comparison to the entire length of the lateral acetabular opening (An index <5% was considered to be clinically irrelevant) The posterior wall sign is positive if the outline of the posterior acetabular rim is more medial than the center of the hip Angle formed by a vertical line and a line connecting the center of the femoral head with the anterior edge of the acetabulum in the false profile view

3 ANALYSIS OF FEMOROACETABULAR IMPINGEMENT 1201 Figure 2. Experimental setup. Figure 3. Software accuracy validation concept. A malrotated radiograph can be analyzed with the help of an additional lateral x- ray. The detected values with the Hip 2 Norm software were compared to the actual measurements of a radiograph in neutral orientation, a false profile for calculation of the ACE angle, and a CT reconstruction. wires. Each pelvis was mounted on a holding device with radiolucent brackets (Fig. 2). To obtain reference measurements, AP radiographs were obtained in neutral rotation utilizing the standard radiographic technique described above. To obtain the calibration distance a, one lateral x-ray of each pelvis was obtained. Additionally, to obtain the reference measurement of anterior center-edge (ACE) angle, a false profile of the hip was obtained with the pelvis in 258 of rotation around the longitudinal axis. The radiographs were then digitized at 300 dots per inch with an 8- bit (256 gray scale) whole-film scanner (Diagnostic PRO Plus, Vidar Corp, Herndon, VA). The reference measurements were performed by an independent observer using commercially available software (Adobe Photoshop, Version 7.0, Adobe Systems, CA). The reference femoral head coverage was measured utilizing CT scan data that were acquired for each pelvis (Siemens Somatom Sensation 16, Forchheim, Germany; 0.75 mm slice thickness, 0.75 mm slice distance, 300 mm field of view). A 3D reconstruction of the pelvis was created (GE Medical Systems, Advantage Workstation 4.0, Milwaukee, WI) and virtually placed in the neutral orientation of 608 of pelvic inclination. The amount of head coverage was measured on a transparent caudal-cranial view of the standardized 3D pelvic model according to the method by Klaue et al. (Fig. 3). 17 Since no femora were available for the cadavers, the head diameter was set be equal to the acetabular diameter, which does not jeopardize the quality of this validation subpart. For each pelvis, two sets of randomly maloriented study AP radiographs were obtained resulting in two sets of calculated measurements using the Hip 2 Norm program. The study radiographs were blinded, randomized, and analyzed twice at least 3 months apart by two independent observers with the Hip 2 Norm software. The calculated parameters from the study radiographs were finally compared to the reference measurements. Software Consistency This part of the study was designed to evaluate the consistency of the correction algorithm of our software for each individual calculated hip parameter. Therefore, 10 dried cadaver pelves without macroscopic abnormalities (20 hips) were mounted on the holding device. After calibration with a lateral radiograph, serial radiographs were taken in 38 steps for rotation and tilt. A range of 158 to 158 for tilt and 98 to 98 for pelvic rotation were chosen according to the range of pelvic orientation found in a pilot study. 6 This resulted in a 16 x-rays per pelvis (32 hips) of which 7 (14 hips) were chosen randomly, blinded, and analyzed by two independent observers. All calculated parameters of each radiograph of one pelvis were normalized to the neutral orientation and compared to test for consistency of the correction algorithm. The agreement for each calculated parameter was calculated for each observer independently. Intra-/Interobserver Analysis This part of the study was approved by the local institutional review board. The interobserver reliability and intraobserver reproducibility of the calculation of all (corrected and noncorrected) parameters with the developed software was investigated in a clinical series consisting of 51 consecutive patients from our outpatient clinic (102 hips), comprising 20 men and 31 women with a mean age of 32.7 years (SD ¼ 9.6 years, range years). Selection criteria were especially focused on femoral head coverage. Two hips with Legg Calvé Perthes disease were excluded due to head asphericity, leaving a total of 100 hips. There were 65 hips with femoroacetabular impingement, 19 with developmental dysplasia of the hip (DDH) and anterior focal acetabular overcoverage, and 16 asymptomatic hips. Fifty-eight hips were nonoperated; 42 hips had undergone previous surgery (30 surgical hip dislocations, 12 periacetabular osteotomies). A standard AP pelvic x-ray and a lateral radiograph were acquired from each patient and analyzed twice by two independent examiners blinded to the clinical data and details of radiological or surgical reports.

4 1202 TANNAST ET AL. Table 2. Validation Results Software Accuracy k/icc Software Consistency k/icc Inter-/Intraobserver Analysis Parameter Bias Precision Limits k/icc Observer 1 Observer 2 Intraobserver 1 Intraobserver 2 Interobserver 3 Total anterior coverage [%] (CI, ) Anterior craniomedial quadrant [%] ( ) (CI, 0.73 ) ( 0.96) (0.72 ) ( ) 0.73 (CI, ) 0.71 ( ) 0.76 (0.66 ) 0.75 (0.65 ) 0.66 ( ) 0.67 ( ) Anterior craniolateral quadrant [%] ( 0.95) ( ) (0.73 ) ( ) ( ) ( ) Anterior caudomedial quadrant [%] ( ) ( 0.98) ( 0.98) ( ) ( ) ( ) Anterior caudolateral quadrant [%] n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. Total posterior coverage [%] ( 0.97) Posterior craniomedial quadrant [%] (0.73 ) Posterior craniolateral quadrant [%] (0.74 ) Posterior caudomedial quadrant [%] ( ) Posterior caudolateral quadrant [%] (0.76 ) Total cranio-caudal coverage [%] ( ) Anterolateral quadrant [%] ( ) Posterolateral quadrant [%] ( ) 0.62 ( ) 0.79 (0.66 ) (0.72 ) (0.72 ) 0.66 ( ) 0.94 ( ) 0.79 (0.65 ) ( ) 0.93 ( 0.95) ( ) 0.87 (0.81 ) 0.88 ( ) n.d. n.d ( ) ( ) 0.86 ( ) 0.69 ( ) ( ) 0.87 ( ) ( ) ( ) ( ) ( ) Anteromedial quadrant [%] n.d. n.d. n.d. n.d. n.d. n.d ( ) Posteromedial quadrant [%] n.d. n.d. n.d. n.d. n.d. n.d. ( ) LCE [degrees] ( ) Acetabular index [degrees] ( ) ACM [degrees] ( ) 0.67 ( ) ( 0.95) 0.84 (0.74 ) 0.89 ( ) 0.64 ( ) ( ) 0.98 ( ) 0.89 ( 0.93) (0.78 ) 0.95 ( 0.97) 0.95 ( 0.96) ( ) ( ) 0.93 ( ) ( ) ( 0.93) 0.94 ( 0.96) 0.96 ( ) 0.99 ( ) 0.97 ( ) 0.74 (0.63 ) 0.98 ( ) ( ) ( ) ( ) ( ) ( ) 0.86 ( ) 0.84 ( ) ( ) 0.60 ( ) 0.51 (0.41 ) ( ) ( ) 0.63 ( )

5 ANALYSIS OF FEMOROACETABULAR IMPINGEMENT 1203 ( ) 0.94 ( 0.96) 0.97 ( ) 0.86 ( ) ( 0.96) Extrusion index [%] ( ) 0.60 ( ) 0.56 ( ) 0.62 ( ) 0.63 ( ) 0.73 ( ) 0.80 ( ) 0.68 ( ) 0.69 ( ) 0.77 (0.64 ) 0.73 ( ) 0.70 (0.50 ) 0.54 ( ) 0.78 (0.72 ) 0.74 (0.67 ) Cross-over sign [% agreement] n.d. n.d. n.d. ( ) 0.72 (0.53 ) 0.88 ( ) Retroversion index [%] ( 0.95) ( ) ( ) Posterior wall sign [% agreement] n.d. n.d. n.d ( ) (0.29 ) (0.41 ) ACE [degrees] ( ) ICC, intraclass correlation coefficient; CI, 95% confidence interval. Statistical Analysis Normal distribution was determined with the Kolmogorov Smirnov test. Paired Student s t-test was used for comparison of normally distributed data, the Wilcoxon rank sum test to compare paired data without normal distribution. Analysis of the software accuracy was determined with the Bland Altman analysis 18 for each parameter by plotting the difference between the two measurement techniques against their averages. Bias was defined as the average difference between the measurements of Hip 2 Norm and the gold standard, precision was defined as one standard deviation of differences including 65% of comparison points, and limits of agreement were defined as two standard deviations including 95% of comparison points. We used the Kappa-value k 19 for calculation of agreement between two categorical measurements, and the intraclass correlation coefficient (ICC) for calculation of agreement among two or more continuous variables: k/icc < 0.20 ¼ poor; ¼ fair; ¼ moderate; 0.80 ¼ good; and ¼ very good. 19 Significance was defined by p < RESULTS Software Accuracy The graphical Bland Altman analysis showed that the mean of the measurement pairs were spread evenly and randomly above and below the zero line (a clinically acceptable bias). Exemplary for the ACM angle, the results can be read as follows: if a radiograph is standardized with Hip 2 Norm, the corrected ACM angle can be calculated with a mean difference of 0.18 when compared to the reference radiograph. In 65% of all cases, the average difference to the real radiograph was within 3.38, and in 95% of all cases the difference to the gold standard was between 6.58 and Based on the ICC, the least accurate measures were LCE, ACE, and acetabular index (Table 2). All except two parameters showed good-to-very good accuracy for the comparison between study and reference radiographs (Table 2). A moderate accuracy was found for the acetabular index and the ACE angle. Referring to measurements based on CT scans, the total femoral cranio-caudal coverage could be detected with good accuracy. Due to the natural course of the acetabulum with the pelves not having osseous abnormalities, the ICC of the medial quadrants in the craniocaudal projection could not be calculated since all values reached almost 100% consistently. The anterolateral/posterolateral quadrant could be calculated with good accuracy. When looking at the two categorical parameters, a correct cross-over sign for the neutral orientation could be simulated with the software in 95%, and a correct posterior wall sign in 92%, of all cases, respectively. Software Consistency The transformation of each parameter back to the neutral orientation achieved a good-to-very good consistency for all computed values (Table 2). Again, the ICC of the anteromedial and -lateral quadrants in the craniocaudal direction as well as the caudolateral quadrant of the anterior and posterior coverage could not be calculated due to the natural course of the

6 1204 TANNAST ET AL. acetabulum, because these values consistently reached 100% or zero impairing a reliable calculation of the ICC. Intra-/Interobserver Analysis A good-to-very good intraobserver variability was found except for the reproducibility of the anteromedial quadrant in craniocaudal direction and the ACE angle for one observer (Table 2). The interobserver reliability was good-to-very good for all parameters except the posteromedial quadrant craniocaudal coverage and the retroversion index for which moderate ICCs were found. DISCUSSION The determination of acetabular coverage has regained increased attention. Not only was a deficient acetabular roof (dysplasia) recognized as a precursor of osteoarthritis, but also excessive femoral coverage could be identified as a cause of early degenerative hip joint disease due to femoroacetabular impingement. Our study validates a computer program Hip 2 Norm that is able to calculate specific radiographic parameters for FAI on AP pelvic radiographs with respect to individual pelvic tilt and rotation (Fig. 4). The program with its correction algorithm was shown to be an accurate, reproducible, and reliable method to restore acetabular morphology in a neutral orientation when compared to radiographs and CT-based measurements of acetabular coverage. The major limitation of the analysis remains the intra- and interobserver variance. Our study has limitations. The ACE angle and the acetabular index could only be calculated with a moderate accuracy. In addition, the external and internal validation was performed in an experimental setup with optimal conditions, including the maintenance of a nearly perfect centering of the x-ray beam. Less accuracy can be anticipated during clinical daily practice since centering of the x-ray beam might not be as easy as under experimental conditions. To reduce this potential error source, the standard radiographic technique with centering of the x-ray beam based on easily palpable anatomical landmarks for the radiology technician was introduced. In addition, the software is not applicable in hips with an aspherical acetabulum where the projected lines of the acetabulum cannot be assumed to lie on a sphere. Although our software allows adjustment of brightness and contrast to modify the image properties, there are low-quality x-rays (particularly older images) where the anterior and the posterior acetabular wall cannot be distinguished clearly, impairing a correct analysis. However, with digital imaging, image quality and processing is better and easier to handle allowing easy detection of the acetabular rim in almost all cases. Our approach always requires the existence of a lateral x-ray of the same patient. This x-ray is only needed once in the patient s lifetime. In cases for which such an image is unavailable and not indicated clinically, the additional radiation exposure and cost is required to apply Hip 2 Norm. To our knowledge, four computerized methods have been developed to analyze native hip morphology based on an AP pelvic radiograph None of them is feasible for the analysis of hips with pincer type FAI, mostly Figure 4. This 26-year-old female underwent bilateral two-stage surgical hip dislocation for FAI treatment with trimming of the acetabular rim. The AP pelvic radiograph revealed an acetabular retroversion on the right side and a more or less normal configuration of the rim on the left side. The calculated pelvic rotation was 6.58 to the right side based on the horizontal distance b between the sacrococcygeal joint (SCJ) and the upper end of the symphysis (SY) with a pelvic inclination of 768 on the lateral pelvic radiograph (indicating excessive pelvic tilt). After computerized correction to the neutral orientation of 608 inclination, the cross-over sign on the right side disappeared. (AW, anterior wall; PW, posterior wall).

7 ANALYSIS OF FEMOROACETABULAR IMPINGEMENT 1205 because they do not provide specific impingement parameters, such as femoral head coverage in the AP direction or the determination of the retroversion sign and index. This is mainly related to the fact that these computer methods were realized before the concept of FAI had gained interest in the orthopedic community. Most of these applications were created to calculate the amount of femoral head coverage before or after acetabular rotational osteotomies for DDH. Furthermore, only two methods include a correction for pelvic tilt, 10,12 and none of them offers a correction for pelvic rotation. The tilt indicators of the two methods with correction could be proven inferior for calculating absolute pelvic tilt in comparison with distance a used in this study. Our software provides the user (surgeon or radiologist) with a comprehensive, validated tool for radiographic analysis of hips with FAI or DDH with or without surgical treatment (including periacetabular osteotomy and surgical hip dislocation with trimming of the acetabular rim). Our calculation of the ACE angle differs from the real false profile and therefore has to be interpreted with caution. We calculate the ACE angle from the 3D reconstructed acetabular configuration based on an x- ray taken in supine position. Originally, the ACE angle is measured on a false profile view with the patient standing. 20 Considering that there is a backward rotation of the pelvis around the transverse axis when changing from the lying to the standing position, 10,21 Hip 2 Norm generally overestimates the ACE angle. In summary, based on the validation results, our software can be used during standardized analysis of hips with FAI independently from individual tilt and rotation, including all relevant hip parameters for this novel entity. Hip 2 Norm is applicable in prospective and retrospective studies, since only one single lateral radiograph over time is necessary to calibrate distance a. The program will be used to correlate clinical and intraoperative findings with radiographical predictors in FAI. Thus, it could become a valuable tool during the diagnostic phase of hip joint pathologies, and can be utilized wherever planar x-rays are available. ACKNOWLEDGMENTS One or more of the authors has received funding from a grant from the National Center for Competence in Research Computer Aided and Image Guided Medical Interventions (Co-Me) of the Swiss National Science Foundation. REFERENCES 1. Jacobsen S, Sonne-Holm S, Lund B, et al Pelvic orientation and assessment of hip dysplasia in adults. Acta Orthop Scand 75: Tannast M, Zheng G, Anderegg C, et al Tilt and rotation correction of acetabular version on pelvic radiographs. Clin Orthop Relat Res 438: DiGioia AM, Hafez MA, Jaramaz B, et al Functional pelvic orientation measured from lateral standing and sitting radiographs. Clin Orthop Relat Res 453: Tannast M, Murphy SB, Langlotz F, et al Estimation of pelvic tilt on anteroposterior x-rays a comparison of six parameters. Skeletal Radiol 35: Hammerberg EM, Wood KB Sagittal profile of the elderly. J Spinal Disord Tech 16: Siebenrock KA, Kalbermatten DF, Ganz R Effect of pelvic inclination on determination of acetabular retroversion. A study on cadaver pelves. Clin Orthop Relat Res 407: Watanabe W, Sato K, Itoi E, et al Posterior pelvic tilt in patients with decreased lumbar lordosis decreases acetabular femoral head covering. Orthopaedics 25: Tannast M, Siebenrock KA, Anderson SE Femoroacetabular impingement: radiographic diagnosis what the radiologist should know. AJR Am J Roentgenol 188: Dutoit M, Zambelli PY Simplified 3D-evaluation of periacetabular osteotomy. Acta Orthop Belg 65: Konishi N, Mieno T Determination of acetabular coverage of the femoral head with use of a single anteroposterior radiograph. A new computerized technique. J Bone Joint Surg 75A: Pedersen DR, Lamb CA, Dolan LA, et al Radiographic measurements in developmental dysplasia of the hip: reliability and validity of a digitizing program. J Pediatr Orthop 24: Kojima A, Nakagawa T, Tohkura A Simulation of acetabular coverage of femoral head using anteroposterior pelvic radiographs. Arch Orthop Trauma Surg 117: Ganz R, Parvizi J, Beck M, et al Femoroacetabular impingement: a cause for osteoarthritis of the hip. Clin Orthop Relat Res 417: Zheng G, Tannast M, Anderegg C, et al Hip 2 Norm: an object-oriented cross-platform program for 3D analysis of hip joint morphology using 2D pelvic radiographs. Comput Methods Programs Biomed 87: Williams PL The skeleton of the lower limb. In: Williams PL, Warkick R, Dyson M, et al., editors. Gray s Anatomy, 37th ed. Edinburgh, UK: Churchill Livingstone; p Drenckhahn D, Eckstein F Becken. In: Drenckhahn D, editor. Benninghoff Anatomie, Vol. 1, 16th ed., Munich, Germany: Urban & Fischer; p [in German]. 17. Klaue K, Wallin A, Ganz R CT evaluation of coverage and congruency of the hip prior to osteotomy. Clin Orthop Relat Res 232: Bland JM, Altman DG Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1: Landis JR, Koch GG The measurement of observer agreement for categorical data. Biometrics 33: Lequesne M, De Sèze S Le faux profil du bassin: nouvelle incidence radiographique pour l étude de la hanche: son utilité dans les dysplasies et les différentes coxopathies. Rev Rhum 12: [in French]. 21. Eddine TA, Migaud H, Chantelot C, et al Variations of pelvic anteversion in the lying and standing positions. Analysis of 24 control subjects and implications for CT measurement of position of a prosthetic cup. Surg Radiol Anat 23: