(Radiology. 1999;212:519-525.)
© RSNA, 1999
Legg-Calvé-Perthes Disease: MR Imaging Evaluation during Manual Positioning of the Hip-Comparison with Conventional Arthrography1
Diego Jaramillo, MD,
Tamara A. Galen, MS,
Carl S. Winalski, MD,
James DiCanzio, MS,
David Zurakowski, PhD,
Robert V. Mulkern, PhD,
Patricia A. McDougall, MD,
Olga L. Villegas-Medina, MD, MPH,
Ferenc A. Jolesz, MD and
James R. Kasser, MD
1 From the Departments of Radiology (D.J., T.A.G., D.Z., R.V.M., O.L.V.M.), Orthopaedic Surgery (P.A.M., J.R.K.), and Research Computing and Biostatistics (J.D.), Children's Hospital, 300 Longwood Ave, Boston, MA 02115; and the Department of Radiology, Brigham and Women's Hospital and Harvard Medical School, Boston, Mass (C.S.W., R.V.M., F.A.J.). Received May 14, 1998; revision requested July 13; final revision received December 2; accepted February 9, 1999. Supported in part by National Institutes of Health grant AR42396-04 (D.J., T.A.G., O.L.V.M.) and GE Medical Systems (all authors). Address reprint requests to D.J. (e-mail: jaramillo@a1.tch.harvard.edu).
 |
Abstract
|
|---|
PURPOSE: To evaluate the use of magnetic resonance (MR) imaging during manual positioning of the hip, or multipositional MR imaging, in an open-magnet configuration to study femoral head containment, articular congruency, and femoral head deformity in Legg-Calvé-Perthes disease.
MATERIALS AND METHODS: In 12 children with advanced Legg-Calvé-Perthes disease, multipositional MR imaging and conventional arthrography were compared in the assessment of containment, femoroacetabular congruency, and femoral head deformity. Images of the hips in several positions were compared subjectively and objectively.
RESULTS: MR imaging correlated well with arthrography for overall subjective assessment of severity of disease (r = 0.71, P = .01), with good interobserver agreement (
= 0.65, P < .001). MR images demonstrated all cases of hinge abduction shown arthrographically. However, MR imaging failed to depict one case of femoral head flattening. MR imaging correlated well with arthrography in the objective evaluation of joint fluid and lateral subluxation (r = 0.80, P < .01). MR imaging correlated poorly with arthrography in the measurement of sphericity of the femoral head.
CONCLUSION: Multipositional MR imaging with an open-magnet configuration was comparable to arthrography for demonstration of femoral head containment and congruency of the articular surfaces of the hip. In the evaluation of deformity, it performed less well.
Index terms: Bones, necrosis, 442.443 • Hip, arthrography, 442.122 • Hip, MR, 442.121411, 442.121412 • Hip, necrosis, 442.443 • Magnetic resonance (MR), comparative studies, 442.121411, 442.121412 • Magnetic resonance (MR), in infants and children, 442.121411, 442.121412, 442.443
|
Introduction
|
|---|
Legg-Calvé-Perthes disease is an idiopathic avascular necrosis of the growing femoral epiphysis, which results in progressive deformity and outward displacement of the femoral head (1,2). One of the main goals of therapy is to maximize containment so that the femoral head can heal adequately within the acetabulum (3). It is also important to position the articular surfaces so that they remain congruent to decrease the risk of osteoarthritis. Patients ultimately may require osteotomies of the femur or the ilium to optimize femoroacetabular relationships.
Arthrography is often performed in severe Legg-Calvé-Perthes disease to decide whether the hips require bracing or surgery and to assess the best position for immobilization (4–7). Arthrography helps to show containment of the femoral head by the acetabulum, congruency of the femoral head and the acetabulum, and deformity of the cartilaginous femoral head. Arthrography of the hip in children usually requires general anesthesia, placing a needle into the joint, injecting contrast material in the joint space, and exposing the pelvis to radiation (8,9). Although magnetic resonance (MR) imaging has been used increasingly to evaluate Legg-Calvé-Perthes disease, it has not yet been used to show the positional variation of femoral head containment and congruency within the acetabulum.
We have performed multipositional MR imaging in a magnet system with an open configuration and compared the findings with those from arthrography. Our goal was to evaluate MR imaging as a possible substitute for arthrography in children with advanced Legg-Calvé-Perthes disease.
|
MATERIALS AND METHODS
|
|---|
Experimental Design
The study was a prospective comparison of multipositional MR imaging and conventional arthrography in children with Legg-Calvé-Perthes disease. Containment of the femoral head by the acetabulum, congruency between the femoral head and acetabulum, and deformity of the femoral head were each evaluated objectively and subjectively in many positions.
Subjects
From April 1996 to October 1997, we performed multipositional MR studies in 17 children with advanced Legg-Calvé-Perthes disease. Five subjects were not included in the analysis. The first three children were examined solely with fast gradient-recalled-echo sequences that did not allow adequate differentiation between joint fluid and epiphyseal cartilage. We subsequently used fast spin-echo T2-weighted sequences that provided excellent contrast between fluid and cartilage and therefore became the standard. One 12-year-old girl refused to undergo arthrography after the MR imaging study. One 3-year-old boy did not complete the study because it was impossible to perform imaging adequately without sedation.
The remaining 12 children (11 boys, one girl; age range, 4.3–11.5 years; median, 6.8 years) were the subjects of the study. All but one of the children underwent intraoperative arthrography within 1 week of MR imaging. This child, by parental decision, delayed the arthrography for 3 months after the MR imaging study. Because the radiographs had not changed substantially in the interval, we included this child in the results. The study was approved by the hospitals' institutional review boards, and informed consent was obtained from the parents for all patients.
Imaging
All MR imaging was performed with a 0.5-T open configuration MR system (Signa SP; GE Medical Systems, Milwaukee, Wis) (10). The imaging field of the system is within a 56-cm wide vertically oriented space; this allows ample space for manipulation of the hips by the examiner and close contact with one or both parents (Fig 1). A circular 25-cm rigid transmit-receive coil was placed around the affected hip from the groin to the anterior superior iliac spine.
View larger version (108K):
[in this window]
[in a new window]
|
Figure 1. Photograph obtained during abduction and flexion. The child is in the lateral decubitus position. The knee is supported by a foam pad (straight arrow). The curved arrow points to the patient's foot.
|
|
An initial coronal fast gradient-recalled-echo sequence of the involved hip (16.9/7.9 [repetition time msec/echo time msec]; flip angle, 40°) was used as a localizer. Fast spin-echo T2-weighted images (2,500/108 [effective]; echo train length, eight; bandwidth, 16 kHz) were then obtained in the coronal plane with the hip in several positions and in the sagittal plane in the neutral position. Imaging parameters for the multipositional sequence were as follows: field of view, 24 x 24 cm; section thickness, 5 mm; intersection gap, 1 mm; matrix, 256 x 192; two signals acquired. Imaging time in each position was 1 minute 58 seconds. Fat saturation, although available, was not used in the open-magnet configuration owing to the low field strength and unavoidable magnetic field inhomogeneities over the field of view.
The patients were examined in the lateral decubitus position with the hip and knee supported with foam pads (Fig 1). Six hip positions in the multipositional series were studied: neutral, 5° of adduction, 20° of abduction, 40° of abduction, 30° of abduction with 45° of flexion, and 45° of abduction with 90° of flexion and maximal external rotation. Different foam pads were used for each position to maintain reproducible degrees of abduction. We standardized the plane of the coronal section by placing a marker on the patient, which was kept in alignment with the laser light during the imaging. A coronal T1-weighted (500/17) image of both hips was then obtained to show any contralateral avascular necrosis.
The spot radiographs from the hip arthrogram were obtained in positions closely matching those of the multipositional MR study. The arthrograms were obtained with use of general anesthesia in the operating room. With fluoroscopic guidance, a 22-gauge needle was placed into the joint at the level of the lower aspect of the femoral neck. Injection of 10–15 mL of iopamidol (Isovue 300; Bracco Diagnostics, Princeton, NJ) was performed with fluoroscopic control until full distention of the joint was achieved. Images were obtained by using a digital C-arm unit (BV29; Phillips Medical Systems, Shelton, Conn).
Image Analysis
Three main parameters were analyzed subjectively and objectively: containment of the femoral head by the acetabulum, congruency of the articular surfaces of the femoral head and acetabulum, and deformity of the femoral head.
Subjective Evaluation
The subjective evaluation was performed by two observers, one pediatric radiologist (D.J.) and one pediatric orthopedic surgeon (J.R.K.), who analyzed the images independently and then by consensus. The observers were blinded to the identity of the patients. The scoring followed two practice sessions during which the observers agreed on the grading systems by using the images of the five patients who were not included in the final comparison. Evaluation of the overall severity of the disease is outlined in the Table. In addition, several parameters were graded separately.
Containment was analyzed by assessing whether with abduction the medial joint space was increased (graded 0 or 1) and whether abduction resulted in improved containment or hinging. Congruency of the articular surfaces was determined by assessing the degree of medial pooling of joint fluid or contrast material between the femoral head and acetabulum and grading it from 0 (absent) to 3 (severe). Femoral head deformity was evaluated by grading the degree of flattening of the femoral head from 0 to 3, depending on whether flattening was absent, mild, moderate, or severe. In addition, we determined whether flattening remained the same (grade 0) or became more apparent with abduction (grade 1).
Objective Evaluation
The objective evaluation also was based on containment, congruency, and femoral head deformity. Two observers (D.J., T.A.G.) independently performed measurements on the MR images and the arthrograms. MR imaging measurements were obtained with use of a workstation with version 2.0 software (Advantage Windows; GE Medical Systems). The arthrograms were digitized and analyzed on a computer (Power Macintosh 8500; Apple Computer, Cupertino, Calif) with NIH Image version 1.60 software (National Institutes of Health, Bethesda, Md). Each of the three parameters was measured by using the arthrograms and MR images from each subject for all available positions.
For evaluation of containment, we calculated a subluxation ratio (6) (Fig 2a), defined as the quotient of the gap between the acetabulum and femoral head, which was defined as the distance between the most medial surface of the femoral head and the lateral border of the acetabular teardrop, and the greatest acetabular diameter. Acetabular diameter was measured from the tip of the teardrop, on arthrograms, or from the corresponding inferior tip of the bony acetabulum, on MR images, to the most lateral portion of the bony acetabular rim. The subluxation ratio is independent of image magnification.
View larger version (146K):
[in this window]
[in a new window]
|
Figure 2a. (a) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows a technique for measuring containment of the femoral head by the acetabulum. Subluxation is determined by dividing the femoroacetabular gap, B, by the bony acetabular diameter, A. (b) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows a technique for measuring congruency of the femoral head and the acetabulum. The area, A, of fluid between the femoral head and the acetabulum is measured and divided by the square of the acetabular diameter (see a and the text). (c) Anteroposterior arthrogram obtained with the hip in the neutral position shows a technique for measuring femoral head sphericity. Sphericity is determined by dividing the height, H, by half of the maximal transverse diameter, D. The sphericity of the weight-bearing aspect of the head is measured by dividing the height, h, by the maximal transverse diameter, D. The height is measured at 135° from the maximal diameter; the 135° arc is identified by the arrow.
|
|
View larger version (161K):
[in this window]
[in a new window]
|
Figure 2b. (a) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows a technique for measuring containment of the femoral head by the acetabulum. Subluxation is determined by dividing the femoroacetabular gap, B, by the bony acetabular diameter, A. (b) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows a technique for measuring congruency of the femoral head and the acetabulum. The area, A, of fluid between the femoral head and the acetabulum is measured and divided by the square of the acetabular diameter (see a and the text). (c) Anteroposterior arthrogram obtained with the hip in the neutral position shows a technique for measuring femoral head sphericity. Sphericity is determined by dividing the height, H, by half of the maximal transverse diameter, D. The sphericity of the weight-bearing aspect of the head is measured by dividing the height, h, by the maximal transverse diameter, D. The height is measured at 135° from the maximal diameter; the 135° arc is identified by the arrow.
|
|
View larger version (114K):
[in this window]
[in a new window]
|
Figure 2c. (a) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows a technique for measuring containment of the femoral head by the acetabulum. Subluxation is determined by dividing the femoroacetabular gap, B, by the bony acetabular diameter, A. (b) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows a technique for measuring congruency of the femoral head and the acetabulum. The area, A, of fluid between the femoral head and the acetabulum is measured and divided by the square of the acetabular diameter (see a and the text). (c) Anteroposterior arthrogram obtained with the hip in the neutral position shows a technique for measuring femoral head sphericity. Sphericity is determined by dividing the height, H, by half of the maximal transverse diameter, D. The sphericity of the weight-bearing aspect of the head is measured by dividing the height, h, by the maximal transverse diameter, D. The height is measured at 135° from the maximal diameter; the 135° arc is identified by the arrow.
|
|
The evaluation of congruency was based on the measurement of the area of joint fluid or contrast material that was detectable between the femoral head and acetabulum (Fig 2b). This area was divided by the square of the greatest acetabular diameter defined earlier, again to make the measurement independent of image magnification.
Assessment of deformity of the femoral head was based on the determination of the sphericity ratio. Sphericity of the femoral head was calculated as the ratio of the height to half of the greatest transverse diameter as defined by Jonsäter (11) (Fig 2c). To evaluate whether the MR imaging–arthrography correlation varied with location of the height measurement of the femoral head, one observer (T.A.G.) also obtained measurements of femoral head height at an angle of 135° with respect to the greatest transverse femoral diameter (Fig 2c).
Statistical Analysis
For the subjective ratings, the statistic was used to measure chance-corrected agreement between readers for both MR imaging and arthrography and intermethod agreement for each reader: poor agreement, less than 0.40; fair agreement, = 0.40–0.59; good agreement, = 0.60–0.74; excellent agreement, = 0.75–1.00 (12). Spearman rank correlation coefficients (rs) were calculated to determine the strength of association between MR imaging and arthrography for grade of disease, medial space widening, hinging with abduction, medial pooling of joint fluid, flattening in the neutral position, and flattening with abduction.
For the objective measurements, linear association between methods and readers for the three parameters (containment, congruency, and deformity) was analyzed by least squares regression analysis and the Pearson correlation coefficient (r). A comparison between arthrography and MR imaging was performed for the neutral position, positions of simple abduction (including 20° and 40°), and positions of flexion-abduction; this comparison between imaging modalities also was performed for all positions combined. For each parameter, agreement between MR imaging and arthrography was investigated by using the approach of Bland and Altman (13). Arthrography was considered to be the standard, and differences between the methods were calculated and plotted. A slope test was used to assess whether these differences varied systematically over the range of measurements (14). An 
level of .05 was used to indicate statistical significance. SAS version 6.11 (SAS Institute, Cary, NC) and SPSS version 7.0 (SPSS, Chicago, Ill) software packages were used. All statistical tests were two tailed.
|
RESULTS
|
|---|
MR Imaging Appearance
T2-weighted fast spin-echo images allowed clear differentiation between the low-signal-intensity epiphyseal and articular cartilage and the high-signal-intensity articular fluid (Fig 3). The physis was seen as a low-signal-intensity area between the high signal intensities of the epiphyseal and metaphyseal marrow. Osteonecrotic marrow was generally of low signal intensity and often difficult to differentiate from the adjacent epiphyseal cartilage (Figs 4, 5). This was particularly noticeable in the younger children. In all cases, the contour of the femoral head changed with changes in position (Figs 3, 4). The ligamentum teres femoris, the transverse acetabular ligament, the fibrocartilaginous labrum, and the joint capsule were readily identified (Fig 3).
View larger version (167K):
[in this window]
[in a new window]
|
Figure 3a. Images obtained in an 11-year-old boy with severe limitation of abduction and rotation of the left hip. The series of images shows that femoral head flattening may be demonstrated only in abduction with MR imaging. (a) Anteroposterior arthrogram obtained with the hip in the neutral position shows flattening of the femoral head (arrows). The femoral head is displaced laterally. (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a shows that the medial joint space contains tissue of high signal intensity, presumably representing thick synovium (open white arrow). The image shows joint fluid (solid white arrow) with the ligamentum teres femoris (black arrow) lying between the fluid and the acetabular tissues. The femoral epiphysis has abnormal signal intensity but is not far from round, and there is good congruency between the femoral head and the acetabulum. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows medial pooling (large arrow) of the contrast material. The femoral head (small arrows) is flattened. (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm that the femoral head (small arrows) is flat. The large arrow shows joint fluid pooled between the femoral head and the acetabulum.
|
|
View larger version (190K):
[in this window]
[in a new window]
|
Figure 3b. Images obtained in an 11-year-old boy with severe limitation of abduction and rotation of the left hip. The series of images shows that femoral head flattening may be demonstrated only in abduction with MR imaging. (a) Anteroposterior arthrogram obtained with the hip in the neutral position shows flattening of the femoral head (arrows). The femoral head is displaced laterally. (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a shows that the medial joint space contains tissue of high signal intensity, presumably representing thick synovium (open white arrow). The image shows joint fluid (solid white arrow) with the ligamentum teres femoris (black arrow) lying between the fluid and the acetabular tissues. The femoral epiphysis has abnormal signal intensity but is not far from round, and there is good congruency between the femoral head and the acetabulum. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows medial pooling (large arrow) of the contrast material. The femoral head (small arrows) is flattened. (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm that the femoral head (small arrows) is flat. The large arrow shows joint fluid pooled between the femoral head and the acetabulum.
|
|
View larger version (160K):
[in this window]
[in a new window]
|
Figure 3c. Images obtained in an 11-year-old boy with severe limitation of abduction and rotation of the left hip. The series of images shows that femoral head flattening may be demonstrated only in abduction with MR imaging. (a) Anteroposterior arthrogram obtained with the hip in the neutral position shows flattening of the femoral head (arrows). The femoral head is displaced laterally. (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a shows that the medial joint space contains tissue of high signal intensity, presumably representing thick synovium (open white arrow). The image shows joint fluid (solid white arrow) with the ligamentum teres femoris (black arrow) lying between the fluid and the acetabular tissues. The femoral epiphysis has abnormal signal intensity but is not far from round, and there is good congruency between the femoral head and the acetabulum. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows medial pooling (large arrow) of the contrast material. The femoral head (small arrows) is flattened. (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm that the femoral head (small arrows) is flat. The large arrow shows joint fluid pooled between the femoral head and the acetabulum.
|
|
View larger version (181K):
[in this window]
[in a new window]
|
Figure 3d. Images obtained in an 11-year-old boy with severe limitation of abduction and rotation of the left hip. The series of images shows that femoral head flattening may be demonstrated only in abduction with MR imaging. (a) Anteroposterior arthrogram obtained with the hip in the neutral position shows flattening of the femoral head (arrows). The femoral head is displaced laterally. (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a shows that the medial joint space contains tissue of high signal intensity, presumably representing thick synovium (open white arrow). The image shows joint fluid (solid white arrow) with the ligamentum teres femoris (black arrow) lying between the fluid and the acetabular tissues. The femoral epiphysis has abnormal signal intensity but is not far from round, and there is good congruency between the femoral head and the acetabulum. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows medial pooling (large arrow) of the contrast material. The femoral head (small arrows) is flattened. (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm that the femoral head (small arrows) is flat. The large arrow shows joint fluid pooled between the femoral head and the acetabulum.
|
|
View larger version (136K):
[in this window]
[in a new window]
|
Figure 4a. Images obtained in a 6-year-old boy show poor containment of the femoral head and improvement with abduction. (a) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained with the hip in the neutral position shows that the most lateral part of the flattened femoral head (solid arrow) is lateral to the hypointense acetabular labrum (open white arrow) and helps confirm the poor containment. Most of the physis (open black arrow) is prematurely fused. (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained with the hip in 40° of abduction shows improved containment and also demonstrates the mild accumulation of fluid (arrow) in the joint space.
|
|
View larger version (138K):
[in this window]
[in a new window]
|
Figure 4b. Images obtained in a 6-year-old boy show poor containment of the femoral head and improvement with abduction. (a) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained with the hip in the neutral position shows that the most lateral part of the flattened femoral head (solid arrow) is lateral to the hypointense acetabular labrum (open white arrow) and helps confirm the poor containment. Most of the physis (open black arrow) is prematurely fused. (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained with the hip in 40° of abduction shows improved containment and also demonstrates the mild accumulation of fluid (arrow) in the joint space.
|
|
View larger version (153K):
[in this window]
[in a new window]
|
Figure 5a. Images obtained in a 9-year-old boy prior to Petrie abduction casting shows the failure of multipositional MR imaging to demonstrate severe femoral head deformity. (a) Anteroposterior arthrogram obtained with the hip in 20° of abduction shows that there is substantial pooling of fluid (arrow) between the femoral head and the acetabulum. This indicates lack of congruency. (b) Corresponding coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) shows only minimal accumulation of fluid (arrow) between what appear to be relatively congruent surfaces. Although a and b were obtained with comparable abduction maneuvers, greater abduction was achieved during arthrography, which may explain the differences between the images.
|
|
View larger version (156K):
[in this window]
[in a new window]
|
Figure 5b. Images obtained in a 9-year-old boy prior to Petrie abduction casting shows the failure of multipositional MR imaging to demonstrate severe femoral head deformity. (a) Anteroposterior arthrogram obtained with the hip in 20° of abduction shows that there is substantial pooling of fluid (arrow) between the femoral head and the acetabulum. This indicates lack of congruency. (b) Corresponding coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) shows only minimal accumulation of fluid (arrow) between what appear to be relatively congruent surfaces. Although a and b were obtained with comparable abduction maneuvers, greater abduction was achieved during arthrography, which may explain the differences between the images.
|
|
Eight patients showed interruption of the physis on coronal MR images (Fig 4a), three showed physeal irregularity, and one had an apparently normal physis. Sagittal images revealed that the deformity of epiphyseal contour was predominantly anterior in 11 patients and central in one. The marrow of the contralateral epiphysis was normal in all patients.
Subjective Analysis
MR imaging correlated well with arthrography for the determination of the overall grade of disease as evidenced by the scoring of the consensus evaluation (rs = 0.71, P = .01) and of that of each independent reader (rs = 0.75 and rs = 0.86, P < .01). There was fair interobserver agreement for the two observers for the scoring of the overall grade of disease as determined by means of arthrography ( = 0.44, P < .01), and interobserver agreement was good with MR imaging ( = 0.65, P < .001).
The evaluation of containment showed a high correlation for hinging with abduction (rs = 0.78, P < .01) (Fig 6) but no correlation between modalities for medial space widening (rs = 0.00, P = .99). The evaluation of congruency showed a high correlation between modalities for medial pooling of joint fluid (rs = 0.80, P < .01 for the consensus data). There was better agreement between observers for the evaluation of pooling with arthrography ( = 0.87, P < .001) than with MR imaging ( = 0.31, P = .12). The evaluation of deformity showed marginally significant correlation for flattening in the neutral position (rs = 0.59, P = .04) and no significant correlation for improved detection of flattening with abduction (rs = 0.43, P = .17) (Fig 5).
View larger version (162K):
[in this window]
[in a new window]
|
Figure 6a. Images obtained in a 9-year-old boy show hinge abduction with marked deformity of the right hip. (a) Anteroposterior arthrogram obtained with the hip in a neutral position shows flattening of the femoral head. The contours of the femoral head and the acetabulum (black arrows) are nearly parallel. The white arrow shows the end of the acetabular labrum defined by the "rose thorn" (capsuloacetabular recess). (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a helps confirm the severe flattening of the femoral head (arrow). The articular surfaces appear congruent. The epiphyseal marrow is nearly normal, which indicates that healing has occurred. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows that the joint opens by hinging on the acetabular margin (curved arrow). There is medial joint pooling between the articular surfaces (straight arrows). (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm severe medial pooling of fluid (straight arrow) and hinging of the femoral head on the acetabular margin (curved arrow).
|
|
View larger version (160K):
[in this window]
[in a new window]
|
Figure 6b. Images obtained in a 9-year-old boy show hinge abduction with marked deformity of the right hip. (a) Anteroposterior arthrogram obtained with the hip in a neutral position shows flattening of the femoral head. The contours of the femoral head and the acetabulum (black arrows) are nearly parallel. The white arrow shows the end of the acetabular labrum defined by the "rose thorn" (capsuloacetabular recess). (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a helps confirm the severe flattening of the femoral head (arrow). The articular surfaces appear congruent. The epiphyseal marrow is nearly normal, which indicates that healing has occurred. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows that the joint opens by hinging on the acetabular margin (curved arrow). There is medial joint pooling between the articular surfaces (straight arrows). (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm severe medial pooling of fluid (straight arrow) and hinging of the femoral head on the acetabular margin (curved arrow).
|
|
View larger version (153K):
[in this window]
[in a new window]
|
Figure 6c. Images obtained in a 9-year-old boy show hinge abduction with marked deformity of the right hip. (a) Anteroposterior arthrogram obtained with the hip in a neutral position shows flattening of the femoral head. The contours of the femoral head and the acetabulum (black arrows) are nearly parallel. The white arrow shows the end of the acetabular labrum defined by the "rose thorn" (capsuloacetabular recess). (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a helps confirm the severe flattening of the femoral head (arrow). The articular surfaces appear congruent. The epiphyseal marrow is nearly normal, which indicates that healing has occurred. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows that the joint opens by hinging on the acetabular margin (curved arrow). There is medial joint pooling between the articular surfaces (straight arrows). (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm severe medial pooling of fluid (straight arrow) and hinging of the femoral head on the acetabular margin (curved arrow).
|
|
View larger version (155K):
[in this window]
[in a new window]
|
Figure 6d. Images obtained in a 9-year-old boy show hinge abduction with marked deformity of the right hip. (a) Anteroposterior arthrogram obtained with the hip in a neutral position shows flattening of the femoral head. The contours of the femoral head and the acetabulum (black arrows) are nearly parallel. The white arrow shows the end of the acetabular labrum defined by the "rose thorn" (capsuloacetabular recess). (b) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as a helps confirm the severe flattening of the femoral head (arrow). The articular surfaces appear congruent. The epiphyseal marrow is nearly normal, which indicates that healing has occurred. (c) Anteroposterior arthrogram obtained with the hip in 40° of abduction shows that the joint opens by hinging on the acetabular margin (curved arrow). There is medial joint pooling between the articular surfaces (straight arrows). (d) Coronal fast spin-echo T2-weighted MR image (2,500/108; echo train length, eight) obtained in the same position as c helps confirm severe medial pooling of fluid (straight arrow) and hinging of the femoral head on the acetabular margin (curved arrow).
|
|
Objective Analysis
The evaluation of containment by means of the subluxation ratio showed good correlation between modalities (r = 0.64, P < .001 for observer 1; r = 0.50, P < .001 for observer 2). The evaluation of congruency revealed good correlation for measurements of joint fluid area (r = 0.45, P < .001; r = 0.37, P = .004). Correlation of joint fluid area was the best for positions of simple abduction (r = 0.60, P < .001) and the worst for the neutral position (r = -0.17, P = .32). For the evaluation of deformity, MR imaging and arthrography correlated poorly. The sphericity index correlations were r = 0.07 (P = 0.62) and r = 0.28 (P = .03) for the two observers. The correlation improved when measurement was calculated by using the height of the femoral head along a line at 135° from the femoral head diameter (r = 0.39, P = .003).
The correlation between the two observers was statistically significant for all parameters and for both modalities. The interobserver correlation coefficients for arthrograms and MR images were 0.56 and 0.74 for subluxation ratio, 0.82 and 0.63 for joint fluid area, and 0.55 and 0.41 for sphericity ratio.
For all parameters, the difference between the measurements of the two modalities depended on the magnitude of the measurements. By using arthrography as the standard, a slope test indicated overestimation with MR imaging at small measurements and underestimation at large measurements (all P < .001).
|
DISCUSSION
|
|---|
The study shows that multipositional MR imaging can provide useful information about containment and congruency of the femoral head in children with Legg-Calvé-Perthes disease. It can reveal whether abduction causes the femoral head to come into the acetabulum or hinge out of it. MR images depict the contour of the femoral head in the coronal and sagittal planes, but our data suggest that there is poor correlation between the MR imaging and the arthrographic depiction of deformity.
The role of MR imaging in the evaluation of Legg-Calvé-Perthes disease varies according to the severity and duration of the disease (15,16). In early Legg-Calvé-Perthes disease, when radiographs are normal, the role of MR imaging is to detect the presence of ischemic marrow (17). In more advanced Legg-Calvé-Perthes disease, MR imaging can demonstrate the extent of marrow involvement (18), the degree of cartilaginous or synovial hypertrophy (19,20), and the extent of physeal and metaphyseal cartilaginous abnormalities (21). The main limitation of static MR imaging is its inability to demonstrate how the deformed femur relates to the acetabulum in various positions. For this reason, arthrography is still widely used prior to bracing or surgery (15).
The main imaging questions in advanced Legg-Calvé-Perthes disease are (a) what is the containment of the femoral head and what are the causes of decreased containment (22), (b) how congruent are the femoral head and the acetabulum, and (c) how severe is the femoral head deformity. Under most circumstances, containment of the femoral head improves with abduction (Fig 4). Abduction bracing and redirecting osteotomies of the ilium or proximal femur result in a greater portion of the femoral head lying within the acetabulum, which in turn facilitates the remodeling of the femoral head (22). A notable exception is hinge abduction. In this situation, the deformed femoral head hinges against the lateral margin of the acetabulum (Fig 6). With greater abduction, the femoral head tends to come out of the acetabulum (23). Therefore, hinge abduction is a contraindication to conventional bracing or surgical therapy.
In advanced Legg-Calvé-Perthes disease, there is little correspondence between the contour of the ossified epiphysis and that of the cartilaginous femoral head, especially when the ossified femoral head undergoes fragmentation and resorption. Assessment of femoroacetabular relationships requires imaging that can define the contour of the cartilaginous femoral head, such as arthrography or MR imaging.
MR imaging offers some important advantages over arthrography. It is not invasive and allows simultaneous imaging of the joint space, the epiphyseal and physeal cartilage, and the marrow of both the affected side and the contralateral epiphysis. Furthermore, it allows a more physiologic evaluation of the range of motion without anesthetic relaxation and without added fluid. With multipositional MR imaging, however, femoral head contour is sometimes difficult to evaluate (Fig 5), and extreme positions are less well tolerated. Motion degradation of the MR images may require repetition of some imaging sequences. It is possible that with improvements in imaging sequences (such as the use of more rapid fast spin-echo acquisitions) and improvements in coil design, the imaging can be improved to the point where multipositional MR imaging is superior to arthrography in all aspects.
MR imaging was successfully performed without sedation and with, at most, mild discomfort. Most children complained about the duration of the study but otherwise tolerated the examination well. By comparison, arthrography is much more invasive and requires general anesthesia. The use of anesthesia allows for the evaluation of more extreme positions of the hip. We believe that multipositional MR imaging should be used instead of arthrography in patients who are at risk with anesthesia, arthrocentesis, or administration of contrast material.
The data suggest that the only substantial discrepancy between the modalities was in the evaluation of femoral head deformity. This was particularly evident in one of the patients (Fig 5) for whom the MR imaging evaluation suggested only mild deformity whereas the arthrography showed clear flattening. When there is little joint fluid, the contrast material injected during arthrography allows better differentiation between joint fluid and cartilage than that provided by MR imaging. In addition, when the deformity involves predominantly the anterior femoral head, the deformity is more apparent in projectional images, such as those obtained with arthrography, than in sectional images, such as coronal MR images.
The actual information about deformity provided by the MR images is better than what is suggested by our comparative study, as the anterior flattening could be seen best on sagittal images. However, we could not evaluate these images in the comparative analysis because it is difficult to obtain good arthrograms in a similar plane. We performed the objective comparison by using the sphericity ratio, widely used in the literature (16), which can be used to assess flattening only perpendicular to the greatest diameter of the femoral head. This measurement also varies owing to subjective differences in assessment of the greatest diameter of the femoral head. The statistically significant interobserver correlation, however, suggests that this is a reproducible way of assessing deformity.
There were differences in the assessment of medial joint fluid pooling. This is not surprising because the amount of joint fluid is much greater during the arthrographic study. Arthrography optimizes the demonstration of fluid distribution because of the injection of additional fluid; MR imaging, however, shows the kinematics of the joint with its usual amount of fluid. Although arthrography was taken as the standard, both observers agreed that the MR images allowed better depiction of what occupied that space by showing the edematous synovial membrane, the ligamentum teres femoris, and acetabular fat as causes of lateral femoral head displacement. MR imaging and arthrography were comparable for the objective evaluation of lateral displacement of the femoral head. The differences in the assessment of the fluid pooled between the femur and acetabulum can be explained in the same fashion as the subjective differences already discussed.
The main limitation of this study was the small number of patients examined. The cases were collected from a very busy practice (Children's Hospital, Boston, Mass) with a large number of patients with Legg-Calvé-Perthes disease, yet it took 2 years to obtain 12 volunteers who would undergo both procedures sequentially. Further evaluation may require a multicenter trial. Great efforts were made to obtain the images in comparable positions by keeping the abduction foam pads constant and marking the plane of the coronal section on the patients. Nonetheless, the images obtained differed between modalities (Fig 5) and between patients.
Our data suggest that the more complex positions that involved flexion and abduction were more difficult to reproduce, although the images were clinically informative. The study was performed in a 0.5-T magnet with a vertically oriented open configuration that allows ample access to the patients; it is likely that similar results can be obtained in other open magnets with different configurations.
Our data show that MR imaging with use of manual positioning of the hip in an open-configuration magnet system in children with Legg-Calvé-Perthes disease is comparable to conventional arthrography for overall assessment of femoroacetabular relationships (containment and congruency) and detection of hinge abduction. Fast spin-echo T2-weighted MR imaging was necessary to show adequate contrast between fluid and cartilage. We hope further improvements such as faster imaging and better resolution will allow widespread use of MR imaging and may resolve some of the challenges in the evaluation of femoral head deformity. Future studies should evaluate a larger population and address cost and benefit. Improved multipositional MR imaging, a noninvasive, possibly less expensive modality that allows concurrent assessment of marrow, physis, and contralateral femoral head, clearly would be the modality of choice over conventional arthrography.
|
Footnotes
|
|---|
Author contributions: Guarantor of integrity of entire study, D.J.; study concepts, D.J., T.A.G., C.S.W., R.V.M., F.A.J., J.R.K.; study design, D.J., T.A.G., C.S.W., J.D., D.Z., R.V.M., O.L.V.M., J.R.K.; definition of intellectual content, D.J., T.A.G., C.S.W., J.R.K.; literature research, D.J., T.A.G.; clinical studies, D.J., C.S.W., P.A.M., O.L.V.M., J.R.K.; experimental studies, D.J., C.S.W., R.V.M., O.L.V.M., F.A.J.; data acquisition, D.J., T.A.G., C.S.W., R.V.M., P.A.M., O.L.V.M., J.R.K.; data analysis, D.J., T.A.G., C.S.W., J.R.K.; statistical analysis, J.D., D.Z.; manuscript preparation, D.J., T.A.G.; manuscript editing, all authors; manuscript review, D.J., T.A.G.
|
References
|
|---|
-
Catterall A. The natural history of Perthes' disease. J Bone Joint Surg [Br] 1971; 53:37-53.
-
Herring JA, Williams JJ, Neustadt JN, Early JS. Evolution of femoral head deformity during the healing phase of Legg-Calvé-Perthes disease. J Pediatr Orthop 1993; 13:41-45.[Medline]
-
Herring JA. The treatment of Legg-Calvé-Perthes disease: a critical review of the literature. J Bone Joint Surg [Am] 1994; 76:448-458.[Free Full Text]
-
Gallagher JM, Weiner DS, Cook AJ. When is arthrography indicated in Legg-Calvé-Perthes Disease?. J Bone Joint Surg [Am] 1983; 65:900-905.[Abstract/Free Full Text]
-
Hochbergs P, Eckerwall G, Egund N, Jonsson K, Wingstrand H. Femoral head shape in Legg-Calvé-Perthes disease: correlation between conventional radiography, arthrography and MR imaging. Acta Radiol 1994; 35:545-548.[Medline]
-
Kamegaya M, Moriya H, Tsuchiya K, Akita T, Ogata S, Someya M. Arthrography of early Perthes' disease: swelling of the ligamentum teres as a cause of subluxation. J Bone Joint Surg [Br] 1989; 71:413-417.
-
Kaniklides C, Lonnerholm T, Moberg A, Sahlstedt B. Legg-Calvé-Perthes disease: comparison of conventional radiography, MR imaging, bone scintigraphy and arthrography. Acta Radiol 1995; 36:434-439.[Medline]
-
Wilkinson R. Hip arthrography in children. In: Dalinka M, eds. Arthrography. New York, NY: Springer-Verlag, 1980; 119-126.
-
Lonnerholm T. Arthrography of the hip in children: technique, normal anatomy and findings in unstable hip joints. Acta Radiol 1980; 21:279-292.
-
Schenck JF, Jolesz FA, Roemer PB, et al. Superconducting open-configuration MR imaging system for image-guided therapy. Radiology 1995; 195:805-814.[Abstract/Free Full Text]
-
Jonsäter S. Coxa plana: a histo-pathologic and arthrographic study. Acta Orthop Scand 1953; (suppl 12):58-88.
-
Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159-174.[Medline]
-
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307-310.[Medline]
-
Zar JH. Biostatistical analysis 3rd ed. Upper Saddle River, NJ: Prentice-Hall, 1996.
-
Hubbard AM, Dormans JP. Evaluation of developmental dysplasia, Perthes disease, and neuromuscular dysplasia of the hip in children before and after surgery: an imaging update. AJR 1995; 164:1067-1073.[Abstract/Free Full Text]
-
Kaniklides C. Diagnostic radiology in Legg-Calvé-Perthes disease. Acta Radiol 1996; 406:1-28.
-
Sebag G, Ducou le Pointe H, Klein I, et al. Dynamic gadolinium-enhanced subtraction MR imaging: a simple technique for the early diagnosis of Legg-Calvé-Perthes disease—preliminary results. Pediatr Radiol 1997; 27:216-220.[Medline]
-
Bos CF, Bloem JL, Bloem RM. Sequential magnetic resonance imaging in Perthes' disease. J Bone Joint Surg [Br] 1991; 73:219-224.
-
Rush BH, Bramson RT, Ogden JA. Legg-Calvé-Perthes disease: detection of cartilaginous and synovial changes with MR imaging. Radiology 1988; 167:473-476.[Abstract/Free Full Text]
-
Ducou le Pointe H, Haddad S, Silberman B, Filipe G, Monroc M, Montagne JP. Legg-Perthes-Calvé disease: staging by MRI using gadolinium. Pediatr Radiol 1994; 24:88-91.[Medline]
-
Jaramillo D, Kasser JR, Villegas-Medina OL, Gaary E, Zurakowski D. Cartilaginous abnormalities and growth disturbances in Legg-Calvé-Perthes disease: evaluation with MR imaging. Radiology 1995; 197:767-773.[Abstract/Free Full Text]
-
Weinstein SL. Legg-Calvé-Perthes syndrome. In: Morrissy RT, Weinstein SL, eds. Pediatric orthopedics. Vol 2. Philadelphia, Pa: Lippincott-Raven, 1996; 951-991.
-
Reinker KA. Early diagnosis and treatment of hinge abduction in Legg-Perthes disease. J Pediatr Orthop 1996; 163:3-9.
This article has been cited by other articles:
|
|
|
|
 
A Gough-Palmer and K McHugh
Investigating hip pain in a well child
BMJ,
June 9, 2007;
334(7605):
1216 - 1217.
[Full Text]
[PDF]
|
|
|
|
|
|
|
N. M. Menezes, S. A. Connolly, F. Shapiro, E. A. Olear, R. M. Jimenez, D. Zurakowski, and D. Jaramillo
Early Ischemia in Growing Piglet Skeleton: MR Diffusion and Perfusion Imaging
Radiology,
January 1, 2007;
242(1):
129 - 136.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
Abstract
Introduction
