Coxa magna quantification using MRI in Legg-Calve-Perthes disease

Jérôme Sales de Gauzy

a,b

, Jérôme Briot

b

, Pascal Swider

b,*

a

Children Hospital, Orthopedic Pediatric Department, Toulouse, France

b

Biomechanics Group, IMFT UMR CNRS 5502, University of Toulouse, Purpan University Hospital, 31059 Toulouse cedex, France

a r t i c l e

i n f o

Article history:
Received 21 July 2008
Accepted 10 September 2008

Keywords:
Computer assisted planning
LCP disease
MRI
3D reconstruction
Pediatric orthopaedics

a b s t r a c t

Background: The idiopathic avascular necrosis of the femoral epiphysis characterizes the Legg-Calve-Per-
thes disease in pediatric osteoarticular pathologies. The coxa magna, more frequently observed, corre-
sponds to an enlargement and deformation of the femoral head. The volume extension induces a
subluxation of the hip, which is a bad prognosis for the healthy function of the joint.

Methods: The aim of the study was to quantify the coxa magna in Legg-Calve-Perthes disease using

magnetic resonance imaging. Twenty-five patients with unilateral Legg-Calve-Perthes disease were
included in the clinical protocol and the volume properties of cartilaginous epiphyseal head were quan-
tified using custom-made image processing software.

Findings: Difference in cartilage volume between healthy hips and pathological ones were significant.

Excepting one patient, we observed a statistically significant volume increase for the pathological hip,
the mean value being +13%.

Interpretation: Our results confirmed the effective three-dimensional properties of the coxa magna,

which is clearly associated to a negative prognosis for the future of the joint. To our knowledge, the vol-
ume quantification of coxa magna has not been established before. The non-radiant MRI associated to
three-dimensional investigation could potentially improve the clinical follow up of children to adapt
the non-invasive treatment and to plan the surgery if necessary.

Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction

The idiopathic avascular necrosis of the femoral epiphysis char-

acterized the Legg-Calve-Perthes disease (LCPD) (

Catterall, 1971;

Herring, 1994

). Among the children population it concerns more

male patients (1.3‰) than female patients (0.3‰). Lateralization
with reduced containment and subsequent deformation resulting
from the antero-lateral flattening of the femoral head are clinical
signs that generally lead to an unsatisfying outcome with subse-
quent osteoarthritis (

Gershuni, 1980; Stulberg et al., 1981; Hall,

1986; Rosenfeld et al., 2007

). It appeared that prognosis is not as

sensitive to head size as it is to flattening of the head and the final
shape; large heads, which remain round, have a good prognosis
(

Stulberg et al., 1981

). Clinical studies established significant corre-

lations between the lateralization, the head subluxation, the carti-
lage hypertrophy and the early stage of the femoral head
deformation. The patient age at disease onset and the extent of
necrosis play a significant role on the prognosis.

LCPD involves four periods: necrosis, fragmentation, healing

and residual period. At residual period, the femoral head may
alternatively be spheric with an increase of volume (coxa magna),

flattened (coxa plana) or irregular. The coxa magna is more fre-
quently observed. This volume extension may induce a subluxation
of the hip, which is a bad prognosis for the healthy function of the
joint (

Schiller and Axer, 1972; Dickens et al., 1978; Grenn et al.,

1981; Yrjonen, 1992; Sales de Gauzy et al., 1997)

.

The clinical investigation of LCPD involved several imaging

modalities: X-rays, arthrography, tomodensitometry and magnetic
resonance imaging (

Heyman and Herndon, 1968; Meyer, 1977;

Toby et al., 1985; Rush et al., 1988; Pinto et al., 1989; Henderson
et al., 1990; Bos and Bloem, 1991; Sebag, 1998; De Sanctis et al.,
2000; Pouletaut et al., 2005

). Since being strongly correlated to

hydration and non-radiant, magnetic resonance imaging is partic-
ularly relevant to study LCPD in young patients. MRI provides pre-
cious information to investigate necrosis; vascular signs, femoral
head morphing, and lateral dislocation.

We previously highlighted the relevance of MRI to investigate

the active phase of the disease (

Sales de Gauzy et al., 1997

). MRI

more easily assessed containment of the femoral head than plain
radiographs. MRI shows not only the bone component but also
the cartilaginous area of the femoral head and acetabulum; it also
allowed detection of early signs of subluxation.

0268-0033/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.clinbiomech.2008.09.002

*

Corresponding author. Tel.: [phone]; fax: [phone].
E-mail address:

[e-mail]

(P. Swider).

Clinical Biomechanics 24 (2009) 43–46

Contents lists available at

ScienceDirect

Clinical Biomechanics

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / c l i n b i o m e c h

To our knowledge, the benefit of MRI to investigate soft-tissue in

LCPD remained limited to bi-dimensional approaches despite the
three-dimensional character of the pathology. It appeared that
acquisition of simultaneous multiplanar images could be used to de-
scribe the distortion of the pathological hip. We hypothesized that
3D reconstruction from MRI could provide an accurate quantifica-
tion of pathological and normal hip volume in Legg-Calve-Perthes

disease. A clinical study involving 25 patients has been imple-
mented, and volume properties of cartilaginous epiphyseal head
were quantified using custom-made image processing software.

2. Methods

2.1. Clinical investigation

This study was performed in 25 children comprising 20 male

and 5 female with unilateral LCPD. The mean age was 5 years
and 6 months ranging from two to eleven years old. All patients
underwent X-rays and MRI. MRI was performed at the necrosis
stage in 11 cases, fragmentation stage in 11 cases and reconstruc-
tion stage in three cases. Results are summarized in

Table 1

. The

imaging analysis was completed using the radiographic classifica-
tion of

Catterall, (1971) and Herring, (1994)

.

2.2. MRI protocol

The MRI was obtained according to a validated clinical infant

protocol: joined slices of 2 mm thickness were obtained. The aver-
age number of slices per hip joint was 30. Two acquisitions were
performed using a turbo spin echo T2-weighted sequence (Magne-
ton Vision, 1.5 T) in the frontal plane and two interleaved acquisi-
tions were required. The field was 400 SI and the pixel definition
was 0.45 mm. For youngest patients, anaesthesia was required.
The ethics committee of the Clinical Research Direction of the Uni-
versity Hospital approved this protocol.

2.3. 3D reconstruction process and statistical analysis

The MRI images were automatically decoded, the contours of

cartilaginous epiphyseal heads were semi-automatically seg-
mented, and the volumes reconstructed using custom-made image
processing software (Biomechlab

Ò

) developed with Matlab

Ò

. As

shown in

Fig. 1

, the global coordinate system was defined in agree-

ment with the Joint Coordinate System (JCS) (

Wu et al., 2002

)

where the e

1

-axis was perpendicular to the sagittal plane, the e

2

-

axis was perpendicular to the frontal plane and the e

3

-axis was

perpendicular to the axial plane. The post-treatment of data

Table 1
Volume properties of healthy and pathological hips

Patients

Stage

Classification

Volume (10

4

mm

3

)

Dv
(%)

Age

Gender

[1]

[2]

Pathological
hip

Healthy
hip

1

5

m

F

3

B

1.20 R

1.05

14.2

2

11

m

N

2.18 L

2.13

2.5

3

6

f

F

2

B

1.31 R

1.21

7.9

4

8

m

N

3

C

1.86 R

1.42

30.8

5

4

m

F

3

B

0.79 R

0.71

10.5

6

8

m

N

2–3

B

1.90 R

1.83

4.6

7

4

m

N

1.04 R

0.87

19.6

8

6

m

N

1.46 L

1.17

16.4

9

5

m

N

3

0.97 L

0.80

22.1

10

5

f

Re

1.12 R

1.03

8.9

11

4

m

F

2

B

1.44 L

1.33

7.7

12

3

m

N

0.75 L

0.63

18.7

13

6

m

F

1.37 R

1.13

21.1

14

5

m

N

0.82 R

0.79

2.9

15

5

m

N

1.06 L

0.98

7.6

16

10

m

F

2

B

1.29 L

1.20

8.6

17

5

f

F

3

C

0.60 L

0.64

7.9

18

7

m

F

3

B

1.13 L

0.99

14.0

19

7

m

F

3

B

1.70 L

1.41

19.8

20

8

m

Re

1.80 R

1.66

13.6

21

7

m

N

1.08 R

0.91

19.5

22

11

f

N

2.07 R

2.15

3.9

23

8

m

Re

1.40 R

1.40

1.9

24

4

m

F

3

B

0.79 R

0.71

10.5

25

4

f

F

4

C

1.22 L

0.98

23.9

Volume ratio was Dv = (pathological hip volume

healthy hip volume)/healthy hip

volume. Pathological hip: R, right hip and L, left hip. Stage disease: F, fragmentation;
N, necrosis; Re, reconstruction. [1] Caterall classification, index 2–4; [2] Herring
classification, group B and C.

Herring classification

Caterall classification: A: group 1, B: group 2, C: group 3, D: group 4

A

A

C

C

B

B

D

Fig. 1. Classification of LCPD according to Caterall [1] and Herring [2].

44

J. Sales de Gauzy et al. / Clinical Biomechanics 24 (2009) 43–46

concerned the volume of the cartilaginous epiphyseal heads. The
variation in percentage was obtained using the healthy head vol-
ume as reference (

Fig. 2

).

The accuracy of the method for volume quantification was eval-

uated using a reproducibility analysis since the contour detection
was semi-automatic and operator-dependent. The accuracy of the
3D reconstruction process has been quantified using the following
protocol: three operators performed six reconstructions of the
same epiphyseal heads and the outcome measures were the vol-
ume. It was verified that the data followed a Gaussian distribution
required for analysis of variance (ANOVA) techniques and the con-
fidence interval (CI) was fixed to 95%. We implemented a paired
Student’s T-test to compare the volumes of healthy hip and patho-
logical hip of each patient.

3. Results

The initial reproducibility analysis of the volume quantification

of the cartilaginous epiphyseal head showed average discrepancies

of 2% for the healthy heads and 4% for the pathological heads. We
estimated this result as satisfying considering the difficulties to de-
tect MRI signal gradient.

Results concerning the 25 patients are summarized in

Table 1

and plotted in

Fig. 3

. Differences in cartilage volume

D

v between

healthy and affected hips were significant. For all patients except-
ing one, we observed an increase in volume for the pathological
hip. The variation in volume are listed in

Table 1

and the mean va-

lue was 13%; these differences were statistically significant
(P = 1.5  10

6

). Coxa magna was observed whatever the stage of

the disease (necrosis, fragmentation and reconstruction).

4. Discussion

We initially hypothesized that MRI could provide an objective

quantification of the volume distortion of cartilaginous epiphyseal
head in LCPD. We initially evaluated the reproducibility of the
semi-automatic 3D reconstruction and admissible results were

a

c

e

3

e

1

e

2

b

e

3

e

1

e

2

Fig. 2. Segmentation and 3D reconstruction of the femoral head from MRI: (a) T

2

weighted image in the frontal plane; (b) volume reconstruction of cartilaginous epiphyseal

head; (c) X-ray.

0

1

2

1 2 3

4

5 6 7 8 9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

health
pathological fermoral head

y femoral head

Volume (10

4

mm

3

)

Fig. 3. Volume properties of healthy and pathological hips for the 25 patients.

J. Sales de Gauzy et al. / Clinical Biomechanics 24 (2009) 43–46

45

obtained considering the operator-dependant detection of signal
gradient inherent to MRI acquisition. Mean discrepancies of 5%
were obtained for volume evaluation.

The clinical study involved 25 patients, 20 boys and 5 girls

according to prevalence from epidemiological studies (ratio 4).
Age repartition was 5–11 with balanced pathological stages: 11
cartilaginous fragmentations, 11 epiphyseal necrosis. The coxa
magna was generally associated to LCPD and this derived from
clinical studies based upon planar data from X-ray arthrography
or MRI. Two-dimensional distortion of the femoral head profile
was depicted but this did not necessary imply a volume variation
of epiphyseal head. In our study and thank to the tri-dimensional
approach, we found that the volume of the pathological epiphyseal
head increased compared to the contra lateral healthy hip. Finally,
our results tended to confirm the effective three-dimensional
properties of the coxa magna. The mean volume increase was
13% and it was statistically significant. This result is closed to the
measurements performed with arthrography. In a recent study,

Rowe et al. (2005)

found a 9% increased of diameter using

arthrography.

The volume increase of the cartilaginous femoral head is clearly

associated to a negative prognosis for the future of the joint. In that
case, a clinical relevant question would be to establish an eventual
correlation between this volume increase and the evolution stage
of the pathology.

Finally, we showed that the non-radiant MRI associated to 3D

investigation could potentially improve the clinical follow up of
children to adapt the non-invasive treatment and to plan the sur-
gery if necessary. In our study, we investigated patients in various
phases of the disease. The aim of further studies would be to en-
large the database and repeat measurements over time. Knowing
at what stage the enlargement is first seen and if the early enlarge-
ment is predictive of enlargement at the outcome would be of high
interest for clinicians.

Conflict of interest

The authors hereby declare to have no conflict of interest.

Acknowledgment

We thank Dr. Christiane Baunin for her support in Pediatric

Imaging.

References

Bos, C.F., Bloem, R.M., 1991. Sequential magnetic resonance imaging in Perthes

disease. J. Bone Joint Surg. Br. 73, 219–224.

Catterall, A., 1971. The natural history of Perthes disease. J. Bone Surg. Br. 53, 37–53.
De Sanctis, N., Nunziata Rega, A., Rondinella, F., 2000. Prognostic evaluation of Legg-

Calve-Perthes disease by MRI, part I & II: the role of physeal involvement. J.
Pediatr. Orthop. 20, 455–470.

Dickens, D.R., Menelaus, M.B., 1978. The assessment of prognosis in Perthes’

disease. J. Bone Joint Surg. Br. 60, 189–194.

Gershuni, D.H., 1980. Preliminary evaluation and prognosis in Legg-Calve-Perthes

disease. Clin. Orthop. Relat. Res. 150, 16–22.

Grenn, N.E., Beauchamps, R.D., Griffin, P.P., 1981. Epiphyseal extrusion as a

prognostic index in Legg-Calve-Perthes disease. J. Bone Joint Surg. Am. 63,
900–905.

Hall, D.J., 1986. Genetic aspects of Perthes’ disease. A critical review. Clin. Orthop.

Relat. Res. 209, 100–114.

Henderson, R.C., Renner, L.B., Sturdivant, M.C., Greene, W.B., 1990. Evaluation of

magnetic resonance imaging in Legg Perthes Calve disease: a prospective,
blinded study. J. Pediatr. Orthop. 10, 289–297.

Herring, J.A., 1994. The treatment of Legg-Calve-Perthes disease. A critical review of

the literature. J. Bone Joint Surg. Am. 76, 448–458.

Heyman, C.H., Herndon, C.H., 1968. Legg-Perthes-Calve disease: a method for the

measurement of the roentgenographic result. J. Bone Joint Surg. Am. 32, 767–
778.

Meyer, J., 1977. Legg-Calve-Perthes’ disease. Radiological results of treatment and

their late clinical consequences. Acta Orthop. Scand. 167 (Suppl), 3–131.

Pinto, M.R., Peterson, H.A., Berquist, T.H., 1989. Magnetic resonance imaging of Legg

Perthes calve disease. J. Pediatr. Orthop. 9, 19–22.

Pouletaut, P., Claude, I., Winzenrieth, R., Hobatho, M.C., Sebag, G., 2005. Automated

analysis of MR image of hip: geometrical evaluation of the Legg-Calve-Perthes
disease. Med. Eng. Phys. 27, 415–424.

Rosenfeld, S.B., Herring, J.A., Chao, J.C., 2007. Legg-calve-perthes disease: a review of

cases with onset before six years of age. J. Bone Joint Surg. Am. 89, 2712–2722.

Rowe, S.M., Moon, E.S., Song, E.K., Seol, J.Y., Seon, J.K., Kim, S.S., 2005. The correlation

between Coxa Magna and final outcome in Legg-Calve-Perthes Disease. J.
Pediatr. Orthop. 25, 22–27.

Rush, B.H., Bramson, R.T., Ogden, J.A., 1988. Legg-Calve-Perthes disease: detection of

cartilaginous and synovial change with MR imaging. Radiology 167, 473–476.

Sales de Gauzy, J., Kerdiles, N., Baunin, C., Kany, J., Darodes, P., Cahuzac, J.P., 1997.

Imaging evaluation of sub-luxation in LCPD: magnetic resonance imaging
compared with the plain radiograph. J. Pediatr. Orthop. B 6, 235–238.

Schiller, M.G., Axer, A., 1972. Hypertrophy of the femoral head in Legg-Calve-

Perthes’ syndrome – L.C.P.S. A study of twenty-nine patients treated by femoral
osteotomy. Acta Orthop. Scand. 43, 45–55.

Sebag, G., 1998. Disorders of the hip, Pediatric musculoskeletal MR imaging. MRI

Clin. North Am. 3, 627–641.

Stulberg, S.D., Cooperman, D.R.T., Wallensten, R., 1981. The natural history of Legg-

Calve-Perthes disease. J. Bone Joint Surg. Am. 63, 1095–1108.

Toby, B.E., Koman, A.L., Bechtold, R.E., 1985. Magnetic resonance imaging of

pediatric hip disease. J. Pediatr. Orthop. 5, 665–671.

Yrjonen, T., 1992. Prognosis in Perthes disease after noncontainment treatment.

Acta Orthop. Scand. 63, 523–526.

Wu, G., Siegler, S., Allard, P., Kirtley, C., Leardini, A., Rosenbaum, D., Whittle, M.,

D’Lima, D.D., Cristofolini, L., Witte, H., Schmid, O., Stokes, I., 2002. ISB
recommendation on definitions of joint coordinate system of various joints
for the reporting of human joint motion – part I: ankle, hip, spine. J. Biomech.
35, 543–548.

46

J. Sales de Gauzy et al. / Clinical Biomechanics 24 (2009) 43–46