Skeletal Radiol DOI 10.1007/s00256-015-2298-y
SCIENTIFIC ARTICLE
Measuring physiological and pathological femoral anteversion using a biplanar low-dose X-ray system: validity, reliability, and discriminative ability in cerebral palsy Matthias Thépaut 1,2 & Sylvain Brochard 1,4,6 & Julien Leboucher 1 & Mathieu Lempereur 1,4 & Eric Stindel 1,5,6 & Valentin Tissot 3 & Bhushan S. Borotikar 1,7
Received: 20 July 2015 / Revised: 2 November 2015 / Accepted: 15 November 2015 # ISS 2015
Abstract Objective The aims of this study were to evaluate the concurrent validity and reliability of a low-dose biplanar X-ray system (Ld-BPR) for the measurement of femoral anteversion (FA) by comparing Ld-BPR-based three-dimensional measures with CT-scan-based measures and to assess the discriminative ability of this method in children with cerebral palsy. Materials and methods Fifty dry femora were scanned using both a CT scan and the Ld-BPR system. Ten femora were artificially modified to mimic a range of anteversion from −30° to +60° and scanned by both modalities. FA was quantified using the images from both modalities and statistically compared for concurrent validity. Intra- and inter-observer reliability of the Ld-BPR system was also determined. Further, Ld-BPR data from 16 hemiplegic and 22 diplegic children were analyzed for its discriminative ability. * Bhushan S. Borotikar
[email protected] 1
Laboratoire de Traitement de l’Information Médicale, INSERM U1101, Batiment 1, Premier etage, CHRU Morvan - 2 Ave. Foch, Brest, France 29609
2
CHRU de Brest, Hôpital Morvan, Service de chirurgie pédiatrique, Brest, France
3
CHRU de Brest, Service de radiologie, Brest, France
4
CHRU de Brest, Hôpital Morvan, Service de Médecine Physique et de Réadaptation, Brest, France
5
CHRU de Brest, La Cavale Blanche, Service d’Orthopédie-Traumatologie, Brest, France
6
Université de Bretagne Occidentale, Brest, France
7
Institut Mines-Telecom, Telecom Bretagne, Brest, France
Results The concurrent validity between the Ld-BPR and CTscan measures was excellent (R2 = 0.83-0.84) and no significant differences were found. The intra- and inter-trial reliability were excellent (ICCs = 0.98 and 0.97) with limits of agreement of (−2.28°; +2.65°) and (−2.76°; +3.38°) respectively. Further, no significant effects of angle or method were found in the sample of modified femora. Ld-BPR measures for FA were significantly different between healthy and impaired femora. Conclusions The excellent concurrent validity with the CT scan modality, the excellent reliability, and the ability to discriminate pathological conditions evaluated by this study make this radiological method suitable for a validated use across hospitals and research institutes. Keywords Biplanar X-rays . Computed tomography . Reliability . Validity . Discriminative ability . EOS
Introduction Femoral anteversion (FA) is defined as the angle made by the femoral neck axis and bicondylar axis of the human femoral bone [1]. Being a direct measure of functional hip joint stability, FA is an important clinical parameter in many pathological conditions related to hip and lower extremity in children [2–4]. Specifically, children with cerebral palsy (CP) exhibit increased FA [5–8], which is typically addressed by surgical correction using rotational osteotomies [9–11]. Two-dimensional CT scan has been the most evaluated technique [12–14] and is considered as the reference for driving treatment decision-making [8, 13, 15–20]. However, accurately and reliably measuring FA
Skeletal Radiol
as a three-dimensional (3D) parameter in children with CP using such 2D measures remains a radiological challenge [8]. Thus, 3D measures have to be developed and validated to determine FA in clinical routines. The radiography system (EOS®, EOS Imaging, Paris, France) is a low-dose [21–23] biplanar X-ray (Ld-BPR) technology [24] for measuring bone morphology. Using this system, Chaibi and associates [25] introduced a fast (≤5 min) and accurate 3D reconstruction method for measuring lower limb geometry parameters in a clinical environment. Buck and colleagues [26] reported excellent accuracy and inter-rater reliability of the Ld-BPR measurements for physiological lower limb torsions in comparison with the CT scan. A recent study aimed to validate the Ld-BPR system for quantifying FA on two cadaveric femurs and a ±10° range of modified femoral torsion [27], whereas another study [28] reported the Ld-BPR system’s feasibility for lower limb measurements in five children with CP. Thus, the feasibility and discriminative ability as a main part of metrological properties of FA measurement still remain to be demonstrated in a greater number of children. This study aimed to establish three main components of validity (concurrent validity, reliability, and discriminative ability) of the Ld-BPR system in measuring the FA, especially in case of pathological FA. In order to avoid dual exposure of radiation, the following aims were targeted: (1) to evaluate the concurrent validity and the intra- and interobserver reliability of the Ld-BPR system on a set of dry healthy femora in comparison with the CT scan, (2) to evaluate the concurrent validity of the Ld-BPR system on a subset of dry femora by mimicking pathological FA, and (3) to illustrate the discriminative ability of the Ld-BPR system in measuring FA in children with hemiplegic and diplegic CP.
Materials and methods To achieve the aims of this study, experiments were conducted in three stages. The first stage comprised of acquiring data on intact femoral bones. For this, 50 dry intact femoral bones (32 right) were acquired from the anatomical department at our hospital and scanned using both the Ld-BPR (EOS, EOS imaging, Paris, France) and CT-scan (Somatom Open Sensation 40, Siemens AG, Erlangen, Germany) modalities. A radiologist confirmed the anatomical integrity of the dry bones and excluded any bones with signs of trauma. The second stage involved mimicking the FA on dry bones and acquiring the images using both the Ld-BPR and CT scans. For this, ten out of 50 femoral bones were artificially modified in order to mimic excessive FA and retroversion as could be found in pathological conditions [6]. Data from the first two stages was used to determine concurrent validity of the Ld-BPR system as well as inter- and intra-rater reliability. The third stage
consisted of data acquisition on 38 children with consecutive hemiplegic or diplegic CP using the Ld-BPR system only. To accomplish this, 16 children (mean age, 10.4 years; range, 5.415.6 years) with hemiplegic CP and 22 children (mean age, 9.8 years; range, 5.7-17.7 years) with diplegic CP were analyzed retrospectively from our hospital’s clinical gait database. This data was used to test the discriminative ability of the LdBPR system. The Ld-BPR measurement was performed as a part of standard care either as a pre-operative assessment or as a clinical monitoring process. Parents and children provided consent for the use of the Ld-BPR data and this study was approved by a local ethical committee. Data acquisition on intact femoral bones Each femur was set in a vertical position with its mechanical axis aligned at the center of the Ld-BPR. This system, through its two vertically translating X-ray sources, allowed simultaneous acquisition of two orthogonal images in the anteroposterior and medio-lateral planes of each femur. The sources were coupled to linear detectors that were based on micromesh gaseous structure technology [29]. Scan time lasted for approximately 10 s. For the CT scan, each femur was placed in a cranio-caudal direction in the scanner with femoral mechanical axis aligned to that of the scanner’s main axis. Then, 3D scans were acquired using 3-mm-thick slices and with no slice gap. Data acquisition on artificially modified femoral bones To mimic the extreme FA, ten dry bones were randomly selected and marked in the physiological position (=0°). An osteotomy was performed to cut the femoral bone in its proximal third location making sure that the mark for physiological position was also split. Then, a metallic tube and a goniometric device were inserted in the diaphysis (Fig. 1). This setup allowed the bone to be rotated gradually to six desired positions viz. -30°, 0°, 15°, 30°, 45°, and 60°, in order to cover the range typically observed in individuals with cerebral palsy [6] (Fig. 2). To block the femoral torsion at each position, setscrews were added. All the modified femoral bones were scanned consecutively in the CT and Ld-BPR modalities in each position, without changing the angle of anteversion between the two data-captures. All the scanning parameters and setup within each scanner were kept the same as used for 50 intact bones. Data acquisition on children Ld-BPR scans were acquired while children were standing in a shifted-feet position [25] in which the pelvis was rotated by approximately 10° and one limb was misaligned with another in the sagittal plane. This position allowed a clear
Skeletal Radiol
Fig. 1 Photographic representation of the modified femora after osteotomy (left) showing the goniometric device and setscrew fixations. An Ld-BPR view of a modified femur (right) showing the intra-medullar metallic device with the axis inserted in the diaphysis, the goniometric system, and the setscrews for fixation
differentiation between left and right femur in the sagittal (medio-lateral) plane during the data processing. Scans were performed by certified technicians from the radiology department of the hospital. Data processing and measurements All the Ld-BPR imaging data were processed using a semiautomatic 3D modeling technique based on the two orthogonal radiographs. First, a semi-automatic segmentation of 2D
Fig. 2 Illustration of the range of angles of the modified femora, by a photographic superimposition of a modified femur in all the different selected positions of rotation from −30° to +60°
bone boundaries using an non-stereo corresponding contours (NSCC) algorithm available in the SterEOS workstation was conducted. The NSCC algorithm was based on parametric models of the femur as described by Le Bras et al. [30] and Laporte et al. [31] for the proximal and distal femur, respectively. The model reconstruction process using the NSCC algorithm was a step-by-step software-guided procedure as described by Chaibi et al. [25] and briefly explained here for clarity. The first step required identification of anatomical landmarks for layout and adjustment of all the geometric elements of the femur model on two orthogonal radiographs. The second step consisted of applying a generic 3D pattern generated from the geometric elements on the 2D radiological contours, resulting in a subject-specific 3D model of the femur. The third and final step comprised of manually adjusting the 3D contours of the model in order to obtain an accurate 3D reconstruction of the femur (Fig. 3). This process generated an adapted 3D model representing the dry bone for one of the subjects. After 3D modeling, the sterEOS® software automatically calculated the FA as an angle between the femoral neck axis and the bicondylar axis. For CT scan measurements, one of the most clinically acceptable methods [32] was used to create a CT scan-based 2D model of the femur. This method has been proposed for measuring FA in case of excessive coxa-valga (>140°), which often occurs in children with cerebral palsy [6]. The method involved superimposing three slices (Fig. 4) viz., a head slice, a neck slice, and a distal slice. All the CT scan data were processed with the appropriate post-reconstruction software on the Leonardo workstation (Siemens Medical Solutions, Erlangen, Germany). Evaluation of concurrent validity, reliability, and discriminative ability For determining the reliability of the measurements, all the dry bones were numbered from F1 to F50 for blind reading by the observers. All the observers started the procedure with original radiographs. It was determined that in order to reach a power of 80 % at the 5 % significance level, two raters measuring at least 40 femoral bones were required [33]. Thus, for the assessment of inter-rater reliability, two observers performed Ld-BPR measurements (EOS A and EOS B) and for intra-rater reliability one observer carried out two measurements (EOS A1 and EOS A2). For intra-rater reliability, an interval of 3 weeks was set between the two readings of the same femur bone. The first observer was a pediatric orthopedic surgeon trainee with 3 years of experience and the second observer was an engineer scientist with 2 years of experience. Both observers were trained for several months in Ld-BPR 3D reconstruction of the lower limbs with sterEOS software before this study. For the assessment of concurrent validity, a third observer independently performed the CT scan
Skeletal Radiol Fig. 3 Full 3D reconstruction of a femur obtained on the SterEOS workstation using the NSCC algorithm. Front and profile view of the femur (left) with adjusted 2D boundaries on the orthogonal radiographs. Front and profile view (right) of the personalized 3D model after semi-automatic reconstruction
measurements (CT). The third observer was a radiologist with 3 years of experience. All the validity and reliability measures were determined on both intact and artificial femur bones. For data on children, a one-time measurement of FA was performed by one of the two certified technicians of the radiology department, depending on their availability.
The concurrent validity between CT-scan measurement and Ld-BPR measurement of FA was carried out using a coefficient of determination R2 and a one-way analysis of variance (ANOVA). The intra- and inter-rater reliability was tested using an intraclass correlation coefficient (ICC 3,1) [34],
standard error of measurement (SEM = SD*√(1-ICC), where SD is the standard deviation of whole sample of values), and the limits of agreement as defined by Bland and Altman [35]. In order to evaluate the interaction effects of the method (CT scan, Ld-BPR) and the angle for the artificial anteversion part of the study, a two-way ANOVA (method*angle) was carried out. Finally, a one-way ANOVA was conducted to compare FA of the healthy side in hemiplegic population with the opposite side of hemiplegic children and both sides of diplegic children. All the statistical analysis was computed using Statistica V6 (StatSoft, Inc., Paris, France) except for the limits of agreement which was computed using a custommade program in Matlab v7.0 (MathWorks Inc., Natick, MA, USA). For ANOVA statistics, if significant differences were
Fig. 4 Reikeras’ method for CT scan assessment of femoral neck anteversion. From left to right: Separated views of the Bhead slice^, Bneck slice^, Bdistal slice^, and superimposition of the three slices for measuring femoral anteversion angle. The head slice was the scan image that passed through the center of the femoral head where it had the largest diameter. The neck slice was the scan image in which the
anterior and posterior cortices of the neck were parallel. The distal slice on the condyles was the one with the greatest antero-posterior width. The neck axis was defined on two superimposed slices of the femoral head and neck. The FA angle was then measured between the neck axis and the line between the posterior condyles which was drawn from a third distal slice
Statistical analysis
Skeletal Radiol Table 1 Evaluation of the concurrent validity between CT scan and Ld-BPR
Concurrent validity CT-SCAN - Ld-BPR CT
EOS A1
EOS A2
EOS B
Mean (SD) (°)
11.18 (7.37)
13.64 (6.99)
13.46 (6.88)
13.15 (6.77)
Mean difference (SD) (°)
-
−2.46 (3.04)
−2.28 (2.93)
−1.97 (2.90)
p value R2 coefficient
-
0.09 0.83
0.11 0.84
0.17 0.84
SD standard Deviation, R2 coefficients of determination
found, a post-hoc analysis with Bonferroni correction was performed. A p < 0.05 was considered significant for all the statistical measures.
Results Reconstruction time for one complete femur was approximately 3 min in the Ld-BPR system and 2 min in the CT-scan’s workstation for dry femur. For data on children, the reconstruction time was 5 min per femur. The mean FA measured using the CT-scan on 50 intact femoral bones was 11.18° (SD = 7.37, min = −2°, max = 25°), whereas the mean neckshaft angle was 123.50° (SD = 4.30°, min = 113°, max = 133°) confirming the healthy anatomical range of the femoral set. The R2 coefficients between the Ld-BPR and CT scan measures of FA were excellent (0.83 to 0.84) for all the Ld-BPR measures (Table 1, Fig. 5). Although it was not significant, the Ld-BPR system tended to over-estimate the CT scan measure by a mean of 2.46°, 2.28°, and 1.97° for EOS A1, EOS A2, and EOS B, respectively (Table 1). The intra and inter-rater reliability were excellent (ICC ≥ 0.97) with a mean difference of 0.18° and 0.31° and the limits of agreement of (−2.28°; 2.65°) and (−2.76°; 3.38°), respectively (Table 2, Fig. 5). In the analysis of artificially modified femora, ANOVA did not show any significant differences between the Ld-BPR and the CT scan measurements when analyzing the six positions together (Table 3). However, the mean difference between the
Ld-BPR and CT scan measurements was highest (7.82°) at +60° (Table 3). Ld-BPR measures for FA of healthy and pathologic limbs of children were significantly different (Table 4). A post hoc analysis revealed that significance was found between healthy and impaired femora, but not between hemiplegic and diplegic limbs (Table 4).
Discussion This study illustrates three main components of the validity of physiological and pathological FA measures using the LdBPR system. In doing so, this study demonstrated the excellent concurrent validity of the Ld-BPR system-based 3D measures with the CT scan-based 2D measures as well as established its ability to differentiate pathological measures of FA in children with CP. The potential of this new technique for the evaluation of 3D parameters of the lower limbs coupled with its very low irradiation can set the Ld-BPR system as a primary technique of radiological diagnosis across hospitals and research institutes. However, this study also raises some technical limitations and clinical perspectives. Previous studies have reported concurrent validity assessments between the Ld-BPR system and CT scan modalities for various lower-limb measurements including FA [23, 26, 27, 36–39]. Buck and associates [26] conducted CT and Ld-BPR scans in 35 individuals with knee osteoarthritis and reported
Fig. 5 Concurrent validity (on the left, an example of the concurrent validity of EOS A2 and CT scan) and reliability results (Bland and Altmann plots for the intra-rater reliability are shown in the middle and inter-rater on the right)
Skeletal Radiol Table 2 Evaluation of the reliability of the Ld-BPR system
Intra-observer reliability
Inter-observer reliability
EOS A1
EOS A2
EOS A1
EOS B
Mean (SD) (°)
13.64 (6.99)
13.46 (6.88)
13.64 (6.99)
13.15 (6.77)
Mean difference (SD) (°)
0.18 (1.26)
0.31 (1.56)
ICC (3.1)/ SEM (°) LOA (°)
0.98/0.88 −2.28; 2.65
0.97/1.11 −2.76; 3.38
ICC intraclass correlation coefficient, SEM standard error of measurement, LOA limits of agreement
no difference (min -5° max + 7°) between two methods for femoral torsion measurement. A recently published study by Rosskopf and colleagues [38] compared CT scan and Ld-BPR data and reported high concurrent validity (ICC ≥ 0.90) when measuring FA on a cohort of 50 children with unknown pathologies. Our study confirms this excellent concurrent validity in another set of 50 physiological FA measures. A slightly superior concurrent validity achieved in Buck et al.’s study was probably due to the different CT scan measurement method they used [12] for measuring femoral torsion. In the current study, although the measurement comparisons did not show any significant differences between the two modalities, LdBPR measurements were found to over-estimate the CT scan measurements. This over-estimation using Reikeras’ method for CT scan measurements is in line with the results observed by Sugano and colleagues who reported under-estimation of 3D CT scan models by 2° in a set of 30 dry femoral bones [40]. They also reported that none of the existing methods performed on 2D CT scan without reconstruction were able to predict FA with a 95 % confidence of better than ±10° [40]. It could then be speculated that the Ld-BPR measurement method, being a 3D measurement method, might be a superior method over the routinely used 2D CT scan measurements; however, further studies including a third gold standard measure are warranted to prove this. FA in pathologies such as CP can reach up to 40° [6]. In the current study, pathological range selected for mimicking the FA was much higher (60°) than average reported FA of 40°, Table 3
and yet no significant differences were found between the CT and Ld-BPR measures. In a similar study, Pomerantz and colleagues [27] reported no differences between CT and Ld-BPR comparison as well with the measures of two artificially modified cadaveric femur (n = 2) from −10° retroversion to +10° anteversion. So the current study extended the validity of the Ld-BPR-based measurements beyond the scope of Pomerantz’s study and encompassed higher pathological ranges of the FA. However, this study also revealed some clinically significant differences (mean error > 7°) for extreme anteversion angle of 60°, exhibiting the limits of both modalities. The highest mean difference between the measures was found at this angle (60°) and at the same time, measurement errors of both the measures (from 45° setup) were highest at this angle. The standard deviation of CT measures was more than twice that of the Ld-BPR measures at 60°, which suggested that the CT method might have deteriorated more than the Ld-BPR method at this higher angle but this phenomenon should be confirmed with further scientific evidence and any FA measurements above 45° should be viewed with caution for treatment planning by the surgeons. In terms of reliability, the current study re-affirms the conclusion of various other studies that have confirmed the excellent inter- and intra-rater reliability of the Ld-BPR system [26, 27, 36, 39]. However, the sample size of 40 needed to have a statistical power of 80 % was not reached in the previous studies and at the same time, these studies did not report the limits of agreement and SEM as reported here. The excellent
Concurrent validity between CT scan and Ld-BPR for ten femurs with artificially modified femoral anteversion
Modified angle
−30°
0°
+15°
+30°
+45°
+60°
Method
Ld-BPR CT
Ld-BPR CT
Ld-BPR CT
Ld-BPR CT
Ld-BPR CT
Ld-BPR CT
Mean (°) SD (°) Mean difference from previous angle (°) Mean difference (°) ANOVA method ANOVA angle ANOVA method*angle
−26 −27.9 4.72 6.71 +1.9 p = 0.7649 p = 10−6 p = 0.5080
13.68 7.82 39.68 +3.78
9.9 28.3 10.66 8.67 37.8 14.62 +2.97
25.06 42.99 10.77 7.39 15.16 14.69 +1.09
41.9 56.73 9.35 7.53 16.84 13.74 −0.87
57.6 77.48 8.42 8.77 15.7 20.75 −7.82
85.3 18.13 27.7
Skeletal Radiol Table 4 Discriminative ability of EOS to differentiate between healthy and pathologic measurements of femoral anteversion Discriminative ability of EOS®
Mean (SD) (°)
Healthy side
Hemiplegic side
Diplegic data
19.12 (10.9)
ability to discriminate pathological conditions evaluated by this study make this radiological method suitable for a validated use across hospitals and research institutes. Acknowledgments We would like to sincerely thank the radiology department team for their assistance in collecting the data.
28.75 (8.62)
33.68 (10.96)
Mean difference (°)
−9.62
−14.55
Compliance with ethical standards
p value Bonferroni post hoc
0.00036 0.033
0.0002
Conflict of interest The authors declare that they have no conflicts of interest.
inter-rater reliability reported by Buck et al. (ICC ≥ 0.943) [26] and the limits of agreement we found for intra and inter-rater reliability confirm the excellent metrological properties of FA measures using the Ld-BPR system. Based on these results, an observer should keep in mind an SEM of approximately 1° when assessing comparative or longitudinal measurements. In children with CP, femoral derotational osteotomy is one of the frequently performed procedures to correct increased FA [10, 41, 42]. Although the quantification of correction needed is still a debated research topic, surgeons still rely on the femoral derotational osteotomy as both short-term and long-term follow-up studies have shown satisfactory outcomes [11, 41, 42]. Previous reports of FA in children with CP use typical 2D CT techniques [13, 20, 43] that have been shown to underestimate this angle from 2.96° to 4.8° [8]. A recent study reported the in-vivo feasibility of measuring FA using the Ld-BPR system in five children with CP [28]. By carrying out the in vivo discriminative analysis in a group of children with CP, the current study illustrated the feasibility of the Ld-BPR system to quantify FA in 38 children with CP. More importantly, the current study originally reported that the Ld-BPR system was able to differentiate between healthy and pathological FA of the hemiplegic and diplegic children, which is one of the key elements of the metrological properties. The Ld-BPR system is increasingly used for assessing 3D bone morphology such as measurement of 3D limb segment parameters, analysis of patella-femoral disorders, and 3D assessment of scoliosis. One of the potential clinical applications of the Ld-BPR system would be to get a complete 3D model of both lower limbs in individuals with cerebral palsy in order to study pre- and post-surgical parameters during rehabilitation and follow-up periods. Indeed, the Ld-BPR system provides a complete three-dimensional goniometry of the entire lower limbs and allow for instant follow-up of the lower-limb 3D morphology during childhood and adulthood. In association with the results of previous studies, the low time required to build the femoral models, the excellent concurrent validity with the CT scan modality, the excellent reliability, and the
References 1.
2. 3. 4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Dunlap K, Shands Jr AR, Hollister Jr LC, Gaul Jr JS, Streit HA. A new method for determination of torsion of the femur. J Bone Joint Surg (Am Vol). 1953;35-A(2):289–311. Cibulka MT. Determination and significance of femoral neck anteversion. Phys Ther. 2004;84(6):550–8. Alvik I. Increased anteversion of the femur as the only manifestation of dysplasia of the hip. Clin Orthop. 1962;22:16–20. Tonnis D, Heinecke A. Diminished femoral antetorsion syndrome: a cause of pain and osteoarthritis. J Pediatr Orthop. 1991;11(4): 419–31. Gordon GS, Simkiss DE. A systematic review of the evidence for hip surveillance in children with cerebral palsy. J Bone Joint Surg Br Vol. 2006;88(11):1492–6. Robin J, Graham HK, Selber P, Dobson F, Smith K, Baker R. Proximal femoral geometry in cerebral palsy: a population-based cross-sectional study. J Bone Joint Surg Br Vol. 2008;90(10):1372– 9. Shore BJ, Yu X, Desai S, Selber P, Wolfe R, Graham HK. Adductor surgery to prevent hip displacement in children with cerebral palsy: the predictive role of the gross motor function classification system. J Bone Joint Surg (Am Vol). 2012;94(4):326–34. Pons C, Remy-Neris O, Medee B, Brochard S. Validity and reliability of radiological methods to assess proximal hip geometry in children with cerebral palsy: a systematic review. Dev Med Child Neurol. 2013;55(12):1089–102. Tylkowski CM, Rosenthal RK, Simon SR. Proximal femoral osteotomy in cerebral palsy. Clin Orthop Relat Res. 1980;151: 183–92. Ounpuu S, DeLuca P, Davis R, Romness M. Long-term effects of femoral derotation osteotomies: an evaluation using threedimensional gait analysis. J Pediatr Orthop. 2002;22(2):139–45. Pirpiris M, Trivett A, Baker R, Rodda J, Nattrass GR, Graham HK. Femoral derotation osteotomy in spastic diplegia. Proximal or distal? J Bone Joint Surg Br Vol. 2003;85(2):265–72. Hernandez RJ, Tachdjian MO, Poznanski AK, Dias LS. CT determination of femoral torsion. AJR Am J Roentgenol. 1981;137(1): 97–101. Ruwe PA, Gage JR, Ozonoff MB, DeLuca PA. Clinical determination of femoral anteversion. A comparison with established techniques. J Bone Joint Surg (Am Vol). 1992;74(6):820–30. Abel MF, Wenger DR, Mubarak SJ, Sutherland DH. Quantitative analysis of hip dysplasia in cerebral palsy: a study of radiographs and 3-D reformatted images. J Pediatr Orthop. 1994;14(3):283–9. Henriksson L. Measurement of femoral neck anteversion and inclination. A radiographic study in children. Acta Orthop Scand Suppl. 1980;186:1–59.
Skeletal Radiol 16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Phillips HO, Greene WB, Guilford WB, Mittelstaedt CA, Gaisie G, Vincent LM, et al. Measurement of femoral torsion: comparison of standard roentgenographic techniques with ultrasound. J Pediatr Orthop. 1985;5(5):546–9. Miller F, Liang Y, Merlo M, Harcke HT. Measuring anteversion and femoral neck-shaft angle in cerebral palsy. Dev Med Child Neurol. 1997;39(2):113–8. Tomczak RJ, Guenther KP, Rieber A, Mergo P, Ros PR, Brambs HJ. MR imaging measurement of the femoral antetorsional angle as a new technique: comparison with CT in children and adults. AJR Am J Roentgenol. 1997;168(3):791–4. Hudson D, Royer T, Richards J. Ultrasound measurements of torsions in the tibia and femur. J Bone Joint Surg (Am Vol). 2006;88(1):138–43. Chung CY, Lee KM, Park MS, Lee SH, Choi IH, Cho TJ. Validity and reliability of measuring femoral anteversion and neck-shaft angle in patients with cerebral palsy. J Bone Joint Surg (Am Vol). 2010;92(5):1195–205. Deschenes S, Charron G, Beaudoin G, Labelle H, Dubois J, Miron MC, et al. Diagnostic imaging of spinal deformities: reducing patients radiation dose with a new slot-scanning X-ray imager. Spine. 2010;35(9):989–94. McKenna C, Wade R, Faria R, Yang H, Stirk L, Gummerson N, et al. EOS 2D/3D X-ray imaging system: a systematic review and economic evaluation. Health Technol Assess. 2012;16(14):1–188. Wade R, Yang H, McKenna C, Faria R, Gummerson N, Woolacott N. A systematic review of the clinical effectiveness of EOS 2D/3D X-ray imaging system. Eur Spine J. 2013;22(2):296–304. Dubousset J, Charpak G, Dorion I, Skalli W, Lavaste F, Deguise J, et al. A new 2D and 3D imaging approach to musculoskeletal physiology and pathology with low-dose radiation and the standing position: the EOS system. Bull Acad Natl Med. 2005;189(2):287– 97. discussion 297–300. Chaibi Y, Cresson T, Aubert B, Hausselle J, Neyret P, Hauger O, et al. Fast 3D reconstruction of the lower limb using a parametric model and statistical inferences and clinical measurements calculation from biplanar X-rays. Comput Methods Biomech Biomed Engin. 2012;15(5):457–66. Buck FM, Guggenberger R, Koch PP, Pfirrmann CW. Femoral and tibial torsion measurements with 3D models based on low-dose biplanar radiographs in comparison with standard CT measurements. AJR Am J Roentgenol. 2012;199(5):W607–12. Pomerantz ML, Glaser D, Doan J, Kumar S, Edmonds EW. Threedimensional biplanar radiography as a new means of accessing femoral version: a comparative study of EOS three-dimensional radiography versus computed tomography. Skelet Radiol. 2015;44(2):255–60. Assi A, Chaibi Y, Presedo A, Dubousset J, Ghanem I, Skalli W. Three-dimensional reconstructions for asymptomatic and cerebral
palsy children's lower limbs using a biplanar X-ray system: a feasibility study. Eur J Radiol. 2013;82(12):2359–64. 29. Despres P, Beaudoin G, Gravel P, de Guise JA. Physical characteristics of a low-dose gas microstrip detector for orthopedic x-ray imaging. Med Phys. 2005;32(4):1193–204. 30. Le Bras A, Laporte S, Bousson V, Mitton D, De Guise JA, Laredo JD, et al. 3D reconstruction of the proximal femur with low-dose digital stereoradiography. Comput Aided Surg. 2004;9(3):51–7. 31. Laporte S, Skalli W, de Guise JA, Lavaste F, Mitton D. A biplanar reconstruction method based on 2D and 3D contours: application to the distal femur. Comput Methods Biomech Biomed Engin. 2003;6(1):1–6. 32. Reikeras O, Hoiseth A. Torsion of the leg determined by computed tomography. Acta Orthop Scand. 1989;60(3):330–3. 33. Donner A, Eliasziw M. Sample size requirements for reliability studies. Stat Med. 1987;6(4):441–8. 34. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86(2):420–8. 35. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307–10. 36. Folinais D, Thelen P, Delin C, Radier C, Catonne Y, Lazennec JY. Measuring femoral and rotational alignment: EOS system versus computed tomography. Orthop Traumatol Surg Res. 2013;99(5): 509–16. 37. Lazennec JY, Rousseau MA, Rangel A, Gorin M, Belicourt C, Brusson A, et al. Pelvis and total hip arthroplasty acetabular component orientations in sitting and standing positions: measurements reproducibility with EOS imaging system versus conventional radiographies. Orthop Traumatol Surg Res. 2011;97(4):373–80. 38. Rosskopf AB, Ramseier LE, Sutter R, Pfirrmann CW, Buck FM. Femoral and tibial torsion measurement in children and adolescents: comparison of 3D models based on low-dose biplanar radiography and low-dose CT. AJR Am J Roentgenol. 2014;202(3): W285–91. 39. Wybier M, Bossard P. Musculoskeletal imaging in progress: the EOS imaging system. Joint Bone Spine: revue du rhumatisme. 2013;80(3):238–43. 40. Sugano N, Noble PC, Kamaric E. A comparison of alternative methods of measuring femoral anteversion. J Comput Assist Tomogr. 1998;22(4):610–4. 41. Dreher T, Wolf S, Braatz F, Patikas D, Doderlein L. Internal rotation gait in spastic diplegia–critical considerations for the femoral derotation osteotomy. Gait Posture. 2007;26(1):25–31. 42. Dreher T, Wolf SI, Heitzmann D, Swartman B, Schuster W, Gantz S, et al. Long-term outcome of femoral derotation osteotomy in children with spastic diplegia. Gait Posture. 2012;36(3):467–70. 43. Murphy SB, Simon SR, Kijewski PK, Wilkinson RH, Griscom NT. Femoral anteversion. J Bone Joint Surg (Am Vol). 1987;69(8): 1169–76.