|Year : 2016 | Volume
| Issue : 1 | Page : 40-47
Application of weight bearing biplanar stereoradiography in assessment of lower limb deformity
Saba Pasha1, Michelle Ho1, Victor Ho-Fung2, Richard S Davidson1
1 Division of Orthopedic Surgery, The Children's Hospital of Philadelphia, Philadelphia, PA 19104, USA
2 Department of Radiology, The Children's Hospital of Philadelphia, Philadelphia, PA 19104, USA
|Date of Submission||24-Feb-2016|
|Date of Acceptance||12-Apr-2016|
|Date of Web Publication||17-May-2016|
Division of Orthopedic Surgery, The Children's Hospital of Philadelphia, 2nd Floor, Wood Center, 34th Street and Civic Center Blvd., Philadelphia, PA 19104
Source of Support: None, Conflict of Interest: None
Background: The purpose of this study was to investigate the agreement between the three-dimensional (3D) weight-bearing radiological measurements of leg deformities in the presence of leg rotation and knee flexion in a biplanar stereoradiography system (EOS imaging) and measurements of computed tomography (CT) scans.
Methods: Upright biplanar X-rays of six Sawbones; with deformity were registered in no flexion/rotation angle, in 10° and 20° of axial rotation, and 10° and 20° of knee flexion. A CT scan of each bone was registered in supine position, and the 3D reconstruction of each bone was generated. Two-dimensional (2D) lengths and deformity angles were measured on the plain X-rays by two independent observers. Two independent observers generated the 3D reconstructions of the biplanar X-rays and the leg deformity parameters were measured in 3D using custom-developed software. 2D and 3D measurements were compared to CT measurements performed by two observers and repeated three times. The intraclass correlation and limit of agreement between the three measurement techniques were evaluated using Bland-Altman plots.
Results: The intraclass correlations were good to excellent for the three imaging modalities (intraclass correlation coefficients = 0.71-0.95). Frontal deformity angles and lengths were significantly different in the 2D X-rays and CT (P < 0.05) whereas all the length and deformity measurements were comparable between CT and 3D X-rays (P > 0.05).
Conclusions: The 3D measurements of the weight-bearing biplanar X-rays were comparable to 3D CT in assessment of the lower limb deformity.
Keywords: Computed tomography, low dose radiography, lower limb deformity, stereoradiography, biplanar imaging, three-dimensional reconstruction
|How to cite this article:|
Pasha S, Ho M, Ho-Fung V, Davidson RS. Application of weight bearing biplanar stereoradiography in assessment of lower limb deformity. J Limb Lengthen Reconstr 2016;2:40-7
|How to cite this URL:|
Pasha S, Ho M, Ho-Fung V, Davidson RS. Application of weight bearing biplanar stereoradiography in assessment of lower limb deformity. J Limb Lengthen Reconstr [serial online] 2016 [cited 2019 May 20];2:40-7. Available from: http://www.jlimblengthrecon.org/text.asp?2016/2/1/40/182575
| Introduction|| |
Normal pediatric lower limb alignment is not well-defined. ,,, However, an excessive deviation from the normal range that interferes with the child's normal gait requires medical intervention. , Pediatric lower limb abnormalities include femorotibial torsion, hip and knee contractures, knee varus/valgus angulation, metatarsus varus, postfracture malunion, and tibia vara and valga. ,,,,, Regardless of the pathology, proper surgical management of the lower limb deformity relies on accurate radiological assessment.  Ensuring precise alignment of the lower limb's mechanical axis is an important factor in surgical planning of the patients with leg deformities. ,, Accurate measurement of the deformity provides a dependable reference for future comparisons of growth and a reliable tool for treatment planning.  In addition to measurement concerns, the ionized radiation exposure and its associated risk of radiation-related cancer in pediatric patients  make it crucial to keep the radiation dose as low as reasonably achievable (ALARA principle) in pediatric imaging. 
In patients with severe leg deformity, computed tomography (CT) imaging in the supine position can provide accurate radiological measurements but does not characterize the true alignment of the leg in weight-bearing standing position.  Time-consuming positioning and potential higher dose of radiation to the patient are important considerations with CT imaging. Traditional weight-bearing two-dimensional (2D) X-ray measurements are subject to error in the presence of the tilt and rotation of the bone ,,, or significant valgus/varus deformity. ,, 2D weight-bearing X-ray methods and full leg computed radiography (CR) were comparable only if the deviation of leg mechanical axis (MAD) was <2 cm. , Teleoroentgenogram and orthoroentgenogram can be affected by cone-shaped radiography, movement artifacts, and stitching. ,
Recently, the EOS Imaging system (EOS Imaging, Paris, France) has been introduced as a low-dose, biplanar stereoradiography system. This imaging system allows upright standing X-rays, reduces the radiation dose, and eliminates image magnification using linear detectors. ,,,,,, While reliability of EOS in clinical assessment of the leg length discrepancy and lower extremity torsion has been tested ,,,, the application of this system in measuring leg deformity in three-dimensional (3D) is still to be investigated.
The objective of this study was to compare lower limb deformity parameters, i.e., length and deformity angles measured on 2D X-rays, 3D reconstructions of biplanar X-rays, and CT scans. We compared these measurements in several degrees of leg axial rotation or knee flexion angles. We hypothesized that 3D measurements of the leg deformity in standing position are not adversely affected by changes in the leg alignment with respect to the X-ray scanners and is not statistically different from 3D measurements of the CT scans.
| Materials and Methods|| |
A total number of six Sawbones® (Sawbones® ; Pacific Research Laboratories, Vashon, Washington, USA) models of lower limb with deformities, three models of tibia and three models of femur, were selected. The bone specifics described by the manufacturer are summarized in [Table 1]. The deformity angles are described as 2D deformities in the frontal plane.
The Sawbones models of the femur and tibia were randomly paired, attached at the knee joint, and assembled into three leg models. The three leg models were mounted in a Plexiglas scaffold that permits up to 55° axial rotation and 45° flexion of the models [Figure 1]. The Plexiglas scaffold was placed in EOS gantry and a total number of 5 biplanar scans, mimicking rotation, or flexion angles, of the 3 leg model were obtained. These scans include (I) baseline scans: The posterior aspects of the femoral condyles were parallel to the posterior-anterior (PA) scanners (0° rotation, 0° flexion). (II) Rotation test 1: Both femur and tibia models were rotated 10° internally, 0° flexion. (III) Rotation test 2: Both femur and tibia models were rotated 20° internally, 0° flexion. (IV) Flexion test 1: 0° axial rotation, 10° knee flexion, and (V) Flexion test 2: 0° axial rotation, 20° knee flexion. The degree of axial rotation and knee flexion were selected above the system maximum rotational measurement error (4°)  to imitate not only the pathological rotation and flexion in upright standing but also possible patient rotation with respect to the X-ray scanners. A total number of 30 X-rays, i.e. 15 pairs of PA and lateral leg scans were registered, resulting in 15 biplanar X-rays of femur bone and 15 bi-planar images of tibia bone. All scans were uploaded to a picture archiving and communication system in our institution directly from the EOS station.
|Figure 1: Plexiglas scaffold designed for lower limb Sawbones models X-ray scans|
Click here to view
This sample size was selected for a test of one-way ANOVA to provide a statistical power of 0.9 (P < 0.05). It was determined using the variance of the length and deformity measurements of the Sawbones provided by the manufacturer [Table 1] to detect at least a 5° and 1 cm difference respectively between the angle and length measurement techniques, i.e., 2D X-ray, 3D X-rays, and CT. This also provides the statistical power to determine good intraclass correlation coefficient (ICC > 0.7).
X-ray two-dimensional measurements
PA X-rays were viewed in Philips iSite Enterprise® (Philips Healthcare Informatics, Inc., Philips, 2011, N.V.) for 2D length measurements. Femur length was measured between the femoral head center and the intercondylar notch apex. Tibia length was measured between the center of the tibia plateau and the mid-point between the medial and lateral malleolus [Figure 2]a. These measurements were based on the SterEOS software 1.6 definition of length measurement, and used for all measurement techniques in this study to enable comparison between the different techniques. Two trained independent observers performed the 2D length measurements on the PA X-ray images independently. The magnitude of MAD, measured on the PA X-ray in iSite® Enterprise, was 6.9 ± 5.4 mm medial.
|Figure 2: Two-dimensional and three-dimensional length measurements. (a) The position of anatomical landmarks of femur (femoral head center and condyle notch) and tibia (center of tibial plateau, and midpoint of the medial and lateral malleolus) used in two-dimensional length measurements. (b) The three-dimensional position of the femoral head center, patellar notch, center of tibial plateau (proximal tibia) and position of the center of malleolus (distal tibia) used in three-dimensional length measurements of Femur and tibia identified in the SterEOS two-dimensional/three-dimensional software|
Click here to view
X-ray three-dimensional measurements
The 3D reconstructions of the 15 biplanar leg scans were generated in SterEOS 1.6 resulting in 30 3D models of femur and tibia bones. Two independent observers generated the 3D reconstruction models. 3D anatomical landmarks, i.e., femoral head center, medial and lateral femoral condyles, posterior, anterior, medial and lateral points on tibia plateau and center of malleolus were identified using SterEOS 1.6. Distance between 3D positions of the center of femoral head and the intercondylar notch apex defines 3D femoral length. Distance between 3D positions of the center of the tibia plateau and the center of malleolus [Figure 2]b defines 3D tibia length.
Two-dimensional deformity measurements
The deformity angles of the femur and tibia models were measured on PA and lateral X-ray images in iSite by two independent observers using the method described in literature  and summarized below.
Femoral frontal deformity angle
Femoral frontal deformity was calculated as the angle between the proximal femur anatomical axis passing through the midline of the proximal femur shaft and distal anatomical axis that passes through the midline of distal shaft and makes an 81° angle with the mediolateral bi-condyle axis of femur in frontal view [Figure 3]a. 
|Figure 3: Two-dimensional deformity measurements.(a) Femoral frontal deformity angle (α) (b) femoral lateral deformity angle (β) (c) tibia frontal deformity angle (α') (d) Tibia lateral deformity angle (β')|
Click here to view
Femoral lateral deformity angle
Femoral lateral deformity was calculated as the angle between the proximal femur anatomical axis passing through the midline of the proximal femur shaft and the midline of the distal shaft that makes an angle of 83° with the line connecting the anterior and posterior notches at the distal femur in sagittal view [Figure 3]b. 
Tibia frontal deformity angle
Tibia frontal deformity was calculated as the angle between the tibial proximal anatomical axis that makes an 87° angle with the axis connecting the medial and lateral tibial condyles and the line passing through the midline of the distal tibial shaft. This line makes a 89° angle with the tangent line to the ankle joint at the distal end of tibia in frontal view [Figure 3]c. 
Tibia lateral deformity angle
Tibia lateral deformity was calculated as the angle between the line passing through one-fifth of the anterior proximal tibial shaft which makes an 81° angle with the anterior-posterior axis tangent to the tibia plateau and the distal tibial midline which makes an 80° angle with the tangent to distal end in sagittal view [Figure 3]d. 
Three-dimensional deformity measurements
Since the 3D deformity measurements are not available in the SterEOS software 1.6, custom-developed software was generated in MATLAB (MATLAB 8.3, The MathWorks Inc., Natick, MA, USA, 2014a) to calculate the bone deformity angles using the 3D reconstruction of the femur and tibia generated in SterEOS software. First, a 2D rotation matrix was used to rotate the femur and tibia models in the transverse plane and place the bi-condyle axes of the femur and tibia in the frontal plane to correct for axial rotation of the bone with respect to the PA X-ray detector [Figure 4]a. The centerlines of the femur and tibia were defined using point cloud of the 3D reconstruction images; a set of center points was defined at the center of the circumferential circles along the tibia and femur shaft [Figure 4]b. Using these center points, a interpolation technique was used in MATLAB to generate the femur and tibia centerlines [Figure 4]b. Using a 2D rotation matrix, the femur and tibia then were oriented in the sagittal plane such that the distal femur and proximal tibia centerlines were placed in the sagittal plane to correct for the sagittal tilt of the bone (knee flexion) [Figure 4]c. The de-rotated 3D reconstruction of the bones then was projected on the frontal and sagittal views, and the deformity angles were calculated using the method explained in the 2D deformity angle measurement section for the femur and tibia in frontal and sagittal planes separately [Figure 3].
|Figure 4: Three-dimensional deformity measurement procedures: (a) The femur and tibia models were rotated in the transverse plane to correct for axial rotation of the bone. (b) The centerline of the models was determined by the barycenter of the circumferential circles along the bone shaft. (c) The models were tilted in the sagittal plane such that the center lines (black line) proximal to the knee joint were placed in the frontal plane to correct for the sagittal tilt of the bone (knee flexion). Deformity angles were calculated from the centerlines and tangent lines to the femoral condyles and tibial plateau (dashed lines)|
Click here to view
Computed tomography scans
The three leg models were scanned in a CT system in the supine position. The 3D reconstruction of the CT scans was generated in Mimics 10.01 (Materialise, Leuven, Belgium). Femur and tibia lengths and deformity angles were measured using the 3D anatomical landmarks [Figure 2] and the method explained in [Figure 3]. One observer generated the 3D reconstructions and two observers, blinded to each other measurements, measured the length and deformity angles. Each observer repeated the measurements of each scan three times, blinded to their previous measurements.
The number of the X-ray and CT studies and the number of repeated measurements are summarized in [Table 2]. Length measurements were calculated as the average of the femoral and tibia lengths. Deformity angles, i.e., frontal and lateral deformity angles were calculated as the average of femoral and tibia deformity angles.
The ICC for 2D X-ray, 3D X-ray, and CT techniques were determined between the two observers. The intraobserver reliability of the CT scan measurements was determined for repeated measurements by each observer. Bland-Altman plots were used to determine the agreement between the three techniques, i.e., 2D EOS, 3D EOS, and CT as the axial rotation or flexion angles changed. One-way ANOVA (level of significance P < 0.05), followed by post hoc Tukey's HSD test, was used to compare the length and deformity angles between the CT and 2D and 3D EOS X-rays measurements. The statistical analysis was performed in R 3.2 (R Foundation for Statistical Computing, Vienna, Austria). 
| Results|| |
All measurement techniques showed excellent intraclass reliability. The ICC for the length measurements were 0.78 (95% confidence interval [CI]: 0.72-0.85), 0.87 (95% CI: 0.85-0.90), and 0.95 (95% CI: 0.92-0.98) for EOS 2D, EOS 3D, and CT, respectively. For frontal deformity measurements, the ICC were 0.73 (95% CI: 0.68-0.77), 0.83 (95% CI: 0.80-0.85), and 0.90 (95% CI: 0.89-0.92) while the ICC for lateral deformity measurements were 0.71 (95% CI: 0.67-0.75), 0.85 (95% CI: 0.81-0.88), and 0.90 (95% CI: 0.87-0.93) for EOS 2D, EOS 3D, and CT, respectively. The intra-observer reliability for CT measurements was 0.98 at (95% CI: 0.96-0.99).
The average length measurement was significantly different between the CT and EOS 2D measurements (P < 0.05) [Table 3]. Bland-Altman plots showed a bias of 2.5 cm and an underestimation of 2D measurement (negative differences) between EOS 2D and EOS 3D measurements [Figure 5]. The 95% CI of limits of agreement (±1.96 standard deviation) was at –10 cm and 4.8 cm for lower and upper limits of agreement, respectively. The distribution of the differences between the two measurement techniques as a function of the average lengths suggest that the variation of at least one of the techniques is strongly affected by length [Figure 5].
|Figure 5: Bland-Altman graphs for statistical evaluation of the limits of agreement between the EOS two-dimensional and three-dimensional length measurements. 95% limits of agreement are shown as solid lines. The mean bias is 2.5 cm|
Click here to view
|Table 3: Average of EOS two-dimensional, EOS three-dimensional, and computed tomography scan measurements |
Click here to view
Frontal deformity angle was significantly different between the EOS 2D measurement and CT scans and EOS 2D and EOS 3D (P < 0.05) but the differences were not significant when the EOS 3D technique and CT scans were compared (P > 0.05). The lateral deformity angles were not significantly different between the three techniques (P > 0.05) [Table 3].
Bland-Altman plot showed a proportional error in the bone frontal deformity angle measurement between the EOS 2D and EOS 3D techniques in frontal plane [Figure 6]. As the average bone deformity angle increased, the disagreement between the EOS 2D and EOS 3D measurements increased. Lower and an upper limit of agreement were –15° and 12° respectively with a measurement bias at 2°.
|Figure 6: Bland-Altman graphs for statistical evaluation of the limits of agreement between the two-dimensional and three-dimensional measurements of tibia frontal deformity angle. 95% limits of agreement are shown as solid lines. The mean bias is 2°|
Click here to view
| Discussion|| |
Precise lower limb radiological measurements are critical in planning lower limb deformity correction. ,, Accurate longitudinal radiological measurements provide information about the rate of progression and optimal time for surgical intervention. , While CT scans provide accurate radiological measurements, the supine positioning limits the ability to apply the findings to the upright alignment of the legs. 2D conventional X-rays can be used to capture true weight-bearing femorotibial alignment; however, this technique is prone to measurement error due to patient positioning, , even when no leg deformity is present. ,,
Considering the limitations of current imaging modalities for clinical assessment of the lower limb deformity, this study focused on evaluating the applicability of a low dose stereoradiography system in upright radiological assessment of the leg deformity in the presence of axial rotation of the leg and knee flexion. The clinical measurements of the lower limb using this new imaging modality were compared to CT scans, the gold standard in 3D musculoskeletal imaging. This study was proof of the concept of clinical application of low dose EOS stereoradiography in evaluating pediatric lower limb deformity.
Significant differences were shown between the 2D and 3D measurements of the leg deformities. While the reliability of the 2D measurements in upright radiography of lower limb has been previously studied, , the impact of shifting the lower limb orientation with respect to the X-ray scanner on leg length and deformity measurements has not. Our results showed that the 3D measurements of the bone did not change as a function of bone orientation and were comparable to CT measurements. The ICC of the 2D X-ray measurements in our study was lower than the ones reported in the literature for leg length measurement using the EOS system. ,,, This difference can be attributed to the fact that previous studies did not consider any variability in the leg alignment with respect to the X-ray scanner. This eliminates the measurement error linked to the axial rotation of the femur and tibia or knee flexion, which can explain a higher ICC in those studies.
Considering the clinical advantages of weight-bearing scans in EOS, we focused on developing a technique that standardizes the patients position and corrects for the positioning differences across the patients. Unlike scanograms or CT scans, where the radiology technician can align the patient's lower limb to the CT beams in supine position, upright weight-bearing imaging in the EOS system can introduce errors associated with patient's natural standing position. The proposed 3D method herein includes a postprocessing platform that permits adjusting the leg orientation after the X-ray acquisition by rotating or tilting the bone as required. However, the accuracy of these 3D measurements still depends on the accuracy of the 3D reconstruction model, which is generated semi-automatically in sterEOS software. Since the sterEOS software uses a library of asymptomatic lower limbs along with a statistical model to generate the 3D shape of the bone, the initial shape of the bone, i.e., how closely it can be matched to the existing models can affect the accuracy of the 3D reconstructed model. This can explain a greater ICC of the CT scan measurements due to differences between the 3D reconstruction techniques in CT and biplanar radiography. Nevertheless, no statistically significant difference between the CT and EOS 3D measurements exists. The good to excellent ICC of the 3D EOS measurements suggests the feasibility of the application of SterEOS software and the postprocessing platform in 3D clinical measurements of the lower limb deformity.
Although the results showed an agreement between the 3D measurements of biplanar X-ray images and CT, 3D stereoradiography appears to be more clinically appropriate as it offers reduced radiation, decreased imaging and patient preparation time, and information regarding the 3D femorotibial alignment in upright standing position. The radiation dose of EOS stereoradiography was 2-18.8 times less than CR and between 2 and 23 times less than CT scans, depending on the radiography system configuration, material composition, and anatomic region scanned. ,,,, This is particularly important in pediatric radiography due to higher radiosensitivity in children.  The cost-effectiveness analysis of the EOS imaging system varied based on the countries' health care system but suggested a doubled number of stereoradiography imaging compared with CR is required to financially break-even. , Nevertheless, the association between exposure to medical imaging and risk of radiation-related cancer, especially in pediatric patients, is an important consideration in selecting an imaging modality in everyday clinical care. 
This study has several limitations. First, a limited number of CT scans were included in the analysis. This study did not investigate the impact of the limb positioning on the CT measurements and only one scan per bone in standard position - limbs parallel to the ruler and orthogonal to the CT beam - was registered in the CT scanner. As the study did not aim to evaluate the sensitivity of the CT measurements to changing bone alignment, the CT scans were used as our gold standard measurement to be compared against the 2D and 3D X-ray measurements. Second, the limited number of rotation and flexion tests (0°, 10°, and 20° of axial rotation and knee flexion) prohibits us from characterizing the impact of the limb rotation and flexion on the length and deformity parameters separately. As [Table 4] shows, the difference between the 2D and 3D X-ray length measurements are more prominent in the flexion test than rotation test, 3.4 cm versus 1.1 cm. Rotation and flexion appeared to have a similar impact on the deformity angle measurements, a difference of 3.4° versus 3.2° between the 2D and 3D measurement in rotation and flexion tests, respectively. However, the sample size was not adequate for statistical comparison between these two tests separately. Finally, we only included Sawbones models to develop and validate the application of the EOS imaging and the postprocessing platform in lower limb deformity measurement. The applicability of this method in clinical measurement of pediatric and adult lower limb deformity with different pathologies is still to be investigated through a clinical study.
|Table 4: Summary of EOS two-dimensional, EOS three - dimensional and computed tomography scan measurements. rotation and flexion tests conducted at 0°, 10°, and 20° with respect to the EOS scanner |
Click here to view
| Conclusion|| |
In the presence of 3D deformity of lower limbs and pathological axial rotation or knee flexion in natural standing as seen in clinical evaluation of the patients, a 3D evaluation of the frontal and lateral deformity angles is required. 2D measurements on the projections of a 3D deformity can result in error which can adversely affect vigilant patient follow-ups and surgical planning. The proposed 3D stereoradiography technique accompanied by the postprocessing platform provided a reliable imaging protocol in the evaluation of lower limb deformity.
Financial support and sponsorship
This study was funded by the Division of Orthopaedics at The Children's Hospital of Philadelphia.
Conflicts of interest
There are no conflicts of interest.
| References|| |
Alsancak S, Guner S, Kinik H. Orthotic variations in the management of infantile tibia vara and the results of treatment. Prosthet Orthot Int 2013;37:375-83.
Lincoln TL, Suen PW. Common rotational variations in children. J Am Acad Orthop Surg 2003;11:312-20.
Mooney JF 3 rd
. Lower extremity rotational and angular issues in children. Pediatr Clin North Am 2014;61:1175-83.
Sass P, Hassan G. Lower extremity abnormalities in children. Am Fam Physician 2003;68:461-8.
Ferguson J, Wainwright A. Tibial bowing in children. Orthop Trauma 2013;27:30-41.
Sabharwal S. Blount disease: An update. Orthop Clin North Am 2015;46:37-47.
Dahl MT. Preoperative planning in deformity correction and limb lengthening surgery. Instr Course Lect 2000;49:503-9.
Paley D. Principles of Deformity Correction. 3 rd
ed. Berlin, Heidelberg: Springer-Verlag; 2002.
Dexel J, Kirschner S, Günther KP, Lützner J. Agreement between radiological and computer navigation measurement of lower limb alignment. Knee Surg Sports Traumatol Arthrosc 2014;22:2721-7.
Linet MS, Slovis TL, Miller DL, Kleinerman R, Lee C, Rajaraman P, et al.
Cancer risks associated with external radiation from diagnostic imaging procedures. CA Cancer J Clin 2012;62:75-100.
Strauss KJ, Kaste SC. The ALARA (as low as reasonably achievable) concept in pediatric interventional and fluoroscopic imaging: Striving to keep radiation doses as low as possible during fluoroscopy of pediatric patients - A white paper executive summary. Pediatr Radiol 2006;36 Suppl 2:110-2.
Guggenberger R, Pfirrmann CW, Koch PP, Buck FM. Assessment of lower limb length and alignment by biplanar linear radiography: Comparison with supine CT and upright full-length radiography. AJR Am J Roentgenol 2014;202:W161-7.
Kawakami H, Sugano N, Yonenobu K, Yoshikawa H, Ochi T, Hattori A, et al.
Effects of rotation on measurement of lower limb alignment for knee osteotomy. J Orthop Res 2004;22:1248-53.
Lonner JH, Laird MT, Stuchin SA. Effect of rotation and knee flexion on radiographic alignment in total knee arthroplasties. Clin Orthop Relat Res 1996;331:102-6.
Radtke K, Becher C, Noll Y, Ostermeier S. Effect of limb rotation on radiographic alignment in total knee arthroplasties. Arch Orthop Trauma Surg 2010;130:451-7.
Sabharwal S, Zhao C, McKeon J, Melaghari T, Blacksin M, Wenekor C. Reliability analysis for radiographic measurement of limb length discrepancy: Full-length standing anteroposterior radiograph versus scanogram. J Pediatr Orthop 2007;27:46-50.
Sabharwal S, Zhao C, McKeon JJ, McClemens E, Edgar M, Behrens F. Computed radiographic measurement of limb-length discrepancy. Full-length standing anteroposterior radiograph compared with scanogram. J Bone Joint Surg Am 2006;88:2243-51.
Aaron A, Weinstein D, Thickman D, Eilert R. Comparison of orthoroentgenography and computed tomography in the measurement of limb-length discrepancy. J Bone Joint Surg Am 1992;74:897-902.
Sabharwal S, Kumar A. Methods for assessing leg length discrepancy. Clin Orthop Relat Res 2008;466:2910-22.
Escott BG, Ravi B, Weathermon AC, Acharya J, Gordon CL, Babyn PS, et al.
EOS low-dose radiography: A reliable and accurate upright assessment of lower-limb lengths. J Bone Joint Surg Am 2013;95:e1831-7.
Gaumétou E, Quijano S, Ilharreborde B, Presedo A, Thoreux P, Mazda K, et al.
EOS analysis of lower extremity segmental torsion in children and young adults. Orthop Traumatol Surg Res 2014;100:147-51.
Gheno R, Nectoux E, Herbaux B, Baldisserotto M, Glock L, Cotten A, et al.
Three-dimensional measurements of the lower extremity in children and adolescents using a low-dose biplanar X-ray device. Eur Radiol 2012;22:765-71.
Guenoun B, Zadegan F, Aim F, Hannouche D, Nizard R. Reliability of a new method for lower-extremity measurements based on stereoradiographic three-dimensional reconstruction. Orthop Traumatol Surg Res 2012;98:506-13.
Thelen P, Delin C, Folinais D, Radier C. Evaluation of a new low-dose biplanar system to assess lower-limb alignment in 3D: A phantom study. Skeletal Radiol 2012;41:1287-93.
Schlatterer B, Suedhoff I, Bonnet X, Catonne Y, Maestro M, Skalli W. Skeletal landmarks for TKR implantations: Evaluation of their accuracy using EOS imaging acquisition system. Orthop Traumatol Surg Res 2009;95:2-11.
R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2015.
Keppler P, Strecker W, Kinzl L. Analysis of leg geometry - Standard techniques and normal values. Chirurg 1998;69:1141-52.
Odenbring S, Berggren AM, Peil L. Roentgenographic assessment of the hip-knee-ankle axis in medial gonarthrosis. A study of reproducibility. Clin Orthop Relat Res 1993; 289:195-6.
Dietrich TJ, Pfirrmann CW, Schwab A, Pankalla K, Buck FM. Comparison of radiation dose, workflow, patient comfort and financial break-even of standard digital radiography and a novel biplanar low-dose X-ray system for upright full-length lower limb and whole spine radiography. Skeletal Radiol 2013;42:959-67.
Kalifa G, Charpak Y, Maccia C, Fery-Lemonnier E, Bloch J, Boussard JM, et al.
Evaluation of a new low-dose digital x-ray device: First dosimetric and clinical results in children. Pediatr Radiol 1998;28:557-61.
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;99:509-16.
Melhem E, Assi A, El Rachkidi R, Ghanem I. EOS(®) biplanar X-ray imaging: Concept, developments, benefits, and limitations. J Child Orthop 2016;10:1-14.
Applegate KE, Thomas K. Pediatric CT - The challenge of dose records. Pediatr Radiol 2011;41 Suppl 2:523-7.
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;16:1-188.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]
[Table 1], [Table 2], [Table 3], [Table 4]