MRI of Legg-Calvé-Perthes Disease

MRI of Legg-Calvé-Perthes Diseasemore frequently affected by LCP disease than other ethnicities.
Although most cases of LCP disease are unilateral, approximately 15% of individu­ als are affected bilaterally [1]. When bilat­ eral, the femoral heads are most common­ ly asynchronously affected. LCP disease is generally considered to be idiopathic with­ out an identifiable inciting factor. No sea­ sonal variation or space–time clustering in the incidence of LCP disease has been rec­ ognized [1]. The incidence of this condition, however, has been correlated with lower so­ cioeconomic status and delayed skeletal age [1]. LCP disease is considered to be a diag­ nosis of exclusion, and other causes of avas­ cular necrosis (such as sickle cell disease, leukemia, corticosteroid administration, and Gaucher’s disease) as well as epiphyseal dys­ plasia must be ruled out. A recent study by Kenet et al. [2] failed to establish an asso­ ciation between LCP disease and a variety of genetic mutations responsible for certain thrombophilic states (including factor V Lei­ den), Gaucher’s disease, and familial os­ teonecrosis of the femoral head (due to mu­ tations affecting type 2 collagen).
MRI of Legg-Calvé-Perthes Disease Although radiography is the primary im­
aging technique used in the evaluation of pa­ tients with suspected or known LCP disease, MRI may play an important complimenta­ ry role. Using a combination of unenhanced and contrast­enhanced imaging sequences, this condition can be confidently diagnosed, even in the setting of normal or equivocal hip
MRI of Legg-Calvé-Perthes Disease
Dillman JR, Hernandez RJ
1Both authors: Department of Radiology, University of Michigan Health System, and Section of Pediatric Radiology, C. S. Mott Children’s Hospital, 1500 E Medical Center Dr., Ann Arbor, MI 48109. Address correspondence to J. R. Dillman ([email protected]).
CME This article is available for CME credit. See www.arrs.org for more information.
Pediatr ic Imaging • Review
AJR 2009; 193:1394–1407
© American Roentgen Ray Society
L egg­Calvé­Perthes (LCP) dis­ ease is a common cause of hip pain and limp in preadolescent children. Early in its course, this
condition, a form of idiopathic osteonecrosis (or osteochondrosis), may be difficult to di­ agnose both clinically and radiographically. MRI is a useful tool for the evaluation of LCP disease that may assist with prompt di­ agnosis, staging, and evaluation of associat­ ed complications. In addition, a variety of MRI findings may provide valuable prognos­ tic information. The MRI findings of LCP disease are quite variable depending on the different stages of the disease (avascular [or necrotic], revascularization, and healing [or reparative] phases). The purpose of this ar­ ticle is to illustrate the MRI appearances of the different stages of LCP disease as well as to present examples of various complications associated with this condition.
Epidemiology and Pathogenesis of Legg-Calvé-Perthes Disease
LCP disease is a relatively uncommon con­ dition that affects approximately 5.1–15.6 in 100,000 children (0.005–0.016%) [1]. Boys are affected approximately five times more commonly than girls [1]. Almost all affected individuals are diagnosed between 2 and 14 years of age, with a peak incidence around 5–6 years of age [1]. Those children diag­ nosed at a younger age typically experience a more benign disease course, whereas those diagnosed at an older age typically require increased rates of intervention and general­ ly experience poorer outcomes. Whites are
Keywords: complications, Legg-Calvé-Perthes disease, MRI, staging
DOI:10.2214/AJR.09.2444
Received January 23, 2009; accepted after revision May 3, 2009.
OBJECTIVE. Legg­Calvé­Perthes disease is a common cause of hip pain in children that may be initially clinically and radiographically difficult to diagnose and stage. The purpose of this article is to describe and illustrate the various MRI appearances of this condition.
CONCLUSION. MRI may show proximal femoral abnormalities before radiography in the setting of Legg­Calvé­Perthes disease, allowing appropriate diagnosis and prompt treat­ ment. MRI can also assess for revascularization, healing, and multiple complications.
Dillman and Hernandez MRI of Legg-Calvé-Perthes Disease
Pediatric Imaging Review
AJR:193, November 2009 1395
MRI of Legg-Calvé-Perthes Disease
radiographs. Early diagnosis of LCP disease is important because it allows prompt initia­ tion of potentially joint­preserving therapies [3]. In addition to providing early diagnosis, MRI also allows accurate staging of the dis­ ease process, evaluation of numerous asso­ ciated complications, and differentiation of LCP disease from other epiphyseal lesions, such as multiple epiphyseal, spondyloepiphy­ seal, and Meyer dysplasias (Fig. 1). Finally, MRI may reveal other unsuspected causes of hip pain, such as juvenile chronic arthritis, fracture, and apophyseal injury (JR Dillman, unpublished data).
MRI is both a sensitive and specific imag­ ing technique for the evaluation of LCP dis­ ease. Although MRI and bone scintigraphy findings in LCP disease correlate quite well [4, 5], MRI depicts the exact extent of femoral head involvement more precisely than pinhole scintigraphs [6] (Fig. 2). MRI also does not expose the pediatric patient to the potentially harmful effects of ionizing radiation.
MR images of the hips may be acquired using either a body or a surface coil [4, 7–9]. Surface coils are preferred because they allow increased signal­to­noise ratio and improved spatial resolution. A variety of pulse sequenc­ es can be used to evaluate for possible LCP disease. Unenhanced pulse sequences com­ monly used include T1­weighted spin­echo, T1­weighted spoiled gradient­recalled echo (SPGR) with or without fat saturation, T2­ weighted fast spin­echo (FSE) with or without fat saturation (or T2­weighted STIR), and pro­ ton density–weighted FSE with or without fat saturation. These sequences are typically ac­ quired in the coronal and sagittal planes. Sag­ ittal imaging has been shown to better reveal the extent of femoral head collapse as well as the angular span of involvement when com­ pared with coronal imaging [10] (Fig. 3). Stud­ ies have also shown that sagittal and coronal
images can be used to calculate the percent­ age of femoral head involvement [11], evaluate femoral head epiphyseal bone and cartilage heights [10], and predict risk of future collapse [12]. In a study by Ha et al. [10], femoral head collapse was seen only on sagittal imaging in 26% of cases reviewed. The axial plane is typi­ cally limited in the evaluation of LCP disease because the superior portion of the femoral head is often inadequately visualized because of partial volume averaging artifact.
Imaging of the hips after the IV admin­ istration of gadolinium­containing contrast material may also be helpful in the evalua­ tion of LCP disease [4, 5, 7]. Contrast­en­ hanced imaging is particularly useful in the assessment of proximal femoral epiphyseal perfusion. Hypoperfusion of the proximal femoral epiphysis may be one of the earliest indicators of LCP disease [4, 5, 7]. Standard unenhanced MR images may be normal in early LCP disease, with the proximal fem­ oral ossific nucleus showing normal signal characteristics [4, 13]. Dynamic multiphasic imaging of the hips in the coronal plane us­ ing either a T1­weighted SPGR or SE pulse sequence after the IV administration of gad­ olinium­containing contrast material (typi­ cally administered at 0.1 mmol/kg) can be used to establish whether proximal femoral epiphyseal vascularity is intact [4, 5, 7, 9]. Delayed T1­weighted imaging 2–5 minutes after contrast material administration may also be obtained to further evaluate the vas­ cular supply of the proximal femoral epiph­ ysis. Digital subtraction imaging techniques (on a pixel­by­pixel basis) should be used to identify subtle alterations in perfusion and likely increase both the sensitivity and specificity of MRI in the evaluation of ear­ ly LCP disease [4, 5]. Our institutional MRI protocol for the evaluation of suspected and known LCP disease is presented in Table 1.
Additional MRI techniques may also play an important role in the evaluation of LCP disease. Based on multiple animal studies, diffusion­weighted imaging appears to be quite sensitive for early proximal femoral epiphyseal ischemia [14–16]. Although dif­ fusion appears to be restricted immediately after the ischemic insult [14, 16], both dif­ fusion and apparent diffusion coefficients (ADCs) begin to increase hours to a few days later [14–16]. In a study by Menezes et al. [15], diffusion and ADC abnormalities were identified before evidence of reperfusion on contrast­enhanced (perfusion) imaging and persisted after blood flow was restored.
MR arthrography of the hip (MRI per­ formed after the intraarticular administration of gadolinium­containing contrast material) can also be used to evaluate patients with LCP disease. This imaging technique may be par­ ticularly beneficial in the evaluation of LCP disease­related complications. Intraarticular contrast material allows detailed evaluation of articular cartilage surfaces, the fibrocarti­ laginous labrum, and femoral head–acetabular congruency. After the intraarticular instillation of contrast material, a variety of T1­weighted spin­echo and SPGR imaging sequences with and without fat saturation should be acquired using a small field of view.
Imaging of the Avascular Phase The avascular or necrotic phase of LCP
disease typically lasts for several months and is commonly painful and associated with a limp. Affected children, on occasion, may be asymptomatic early in the avascular phase. If acquired early in the avascular phase, ra­ diographs can be entirely normal. It is during this time, however, that the earliest changes of LCP disease become evident on MRI [13].
A variety of MRI findings may be ob­ served in the avascular phase of LCP dis­
TABLE 1: Institutional MRI Protocol for Evaluation of Suspected and Known Legg-Calvé-Perthes Disease
Parameter STIR FSE T1-Weighted SE T2-Weighted FSE FS T1-Weighted SPGR FS 3D MRAa T1-Weighted SPGR FS Contrast-Enhanceda
Plane Coronal Coronal Axial Coronal Coronal Coronal/sagittal
No. of echo trains 7 3 17 NA NA NA
Inversion recovery time (ms) 150 NA NA NA NA NA
Flip angle (°) NA NA NA 35 40 35
TE (ms) 25 10 80 4.6 1.4 4.6
TR (ms) 3,000 500 3,000 32 4.4 32
No. of signal averages 2 4 6 2 2 2
Note—Coverage should extend from iliac crests through lesser trochanter of proximal femora. FSE = fast spin-echo, SE = spin echo, FS = fat-saturated, SPGR = spoiled gradient-recalled echo, MRA = MR angiography, NA = not applicable.
aAfter IV injection of 0.1 mmol/kg of gadolinium-based contrast material.
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 1
71 .2
43 .0
.1 61
o n
03 /1
1/ 23
f ro
m I
P ad
dr es
s 17
1. 24
3. 0.
16 1.
C op
yr ig
ht A
R R
S. F
or p
er so
Dillman and Hernandez
ease. All or only a portion of the proximal femoral ossific nucleus may become necrot­ ic [4, 17, 18]. Necrosis is most commonly subchondral and central in location (Figs. 3–5), and it is less often medial and lateral [9]. Proximal femoral epiphyseal MR signal abnormality may be observed early in the course of LCP disease on unenhanced im­ aging sequences. On T1­weighted imaging sequences, the proximal femoral ossific nu­ cleus typically contains focal or diffuse ab­ normally low or intermediate signal [7, 18, 19] (Figs. 4–8). T2­weighted/STIR imaging sequences can show variable signal intensity, including areas of increased signal thought to represent bone marrow edema. On occa­ sion, curvilinear subchondral T2­weighted signal hyperintensity and T1­weighted signal hypointensity may be observed in the antero­ superior aspect of the femoral head (Fig. 4). This finding, referred to as the crescent sign (or Caffey sign), is best seen on coronal and sagittal images and suggests the presence of a subchondral fracture. Epiphyseal signal hypointensity on all imaging sequences, in­ cluding contrast­enhanced sequences, sug­ gests more advanced necrosis as well as the presence of ossific nucleus microfractures [7, 9] (Figs. 5, 7, and 8).
The normal hip shows early rapid enhance­ ment after the IV administration of gadolini­ um­containing contrast material [4, 5]. Con­ trast­enhanced imaging may reveal partial (Figs. 1–4) or complete (Figs. 2, 6, and 7) nonenhancement of proximal femoral epiph­ ysis in the avascular phase of LCP disease [4]. Abnormal femoral head enhancement is best depicted approximately 2 minutes after the IV injection of contrast material using subtraction techniques [4, 5] (Figs. 2, 6, and 8–10). There is typically asymmetric antero­ posterior involvement of the femoral head; the anterior portion is frequently the earliest and most prominently involved [10].
A variety of other MRI findings may also be observed in early LCP disease. There is evidence to suggest that mild ischemia to the proximal femoral epiphysis may not result in frank necrosis but instead cause only de­ layed ossification and failure of ossification center growth [17]. Proximal femoral ossif­ ic nucleus morphologic changes, such as ar­ ticular surface flattening, may also suggest early LCP disease [19]. This flattening is likely due to a combination of necrotic tra­ becular bone collapse and resorption. Me­ chanical deformation of the femoral head may result in additional ischemic episodes
and increased necrosis [17]. Overlying femo­ ral head articular cartilage can be abnormal­ ly thickened [8, 17, 19] and show abnormal signal characteristics [8]. Articular cartilage may continue to proliferate despite ischemia due to nutrition derived from synovial fluid [20]. Adjacent acetabular articular cartilage and the fibrocartilaginous labrum may also hypertrophy [19]. Finally, periarticular T2­ weighted/STIR signal hyperintensity and as­ sociated contrast enhancement are common­ ly observed, suggesting the presence of joint effusion and synovitis (pannus formation) [19, 21] (Figs. 1 and 4–10). Such synovitis may be visualized greater than 60 months after the diagnosis of LCP disease [21].
Imaging of Revascularization and Reparative Phases
The revascularization and reparative or healing phases of LCP disease are charac­ teristically symptomatic and typically per­ sist for several years. During these phases, unenhanced and contrast­enhanced imaging sequences frequently reveal heterogeneous proximal femoral epiphyseal signal, likely reflecting a combination of necrotic changes, revascularization, and reparation (Figs. 5, 6, 8, and 9). After the avascular phase, necrot­ ic bone is eventually resorbed and replaced by granulation tissue. With time, granulation tissue is replaced by more mature fibrous tis­ sue, cartilage, and, eventually, mature trabec­ ular bone [7].
Revascularization of the proximal femo­ ral epiphysis likely occurs by one of two pro­ cesses [4, 22]. First, preexisting epiphyseal vessels may be recanalized. Alternatively, reperfusion may occur via neovasculariza­ tion. Revascularized areas of the proximal femoral epiphysis typically show T2­weight­ ed/STIR signal hyperintensity and contrast enhancement [4, 13] (Figs. 5, 6, 8, and 9). Areas of revascularized epiphysis may even hyperenhance after the IV administration of contrast material when compared with areas that were never avascular. Early reperfusion of the lateral column is associated with im­ proved prognosis [4]. Revascularized por­ tions of a femoral head may show persistent enhancement on delayed imaging [4].
A variety of epiphyseal abnormalities may be observed in the revascularization and re­ parative phases of LCP disease. Morpho­ logic changes related to epiphyseal necro­ sis may include articular surface flattening (coxa plana) and fragmentation (Figs. 8, 10, and 11). Epiphyseal fragments may show dis­
similar signal characteristics, suggesting dif­ ferent phases of the disease process [9] (Fig. 8). MRI can also be used to assess for lat­ eral subluxation of the femoral head or loss of containment (Figs. 7–14). This finding is thought most commonly to be due to a com­ bination of cartilage thickening, joint effu­ sion, and synovial thickening [8, 23]. Lat­ eral displacement of the femoral head may also result in abnormal horizontal position­ ing or upward eversion of the fibrocartilagi­ nous labrum [24] (Figs. 9–13). With healing, proximal femoral epiphyseal height is slow­ ly restored, ossific fragments coalesce, and mature trabecular bone again constitutes the entire ossific nucleus. After approximately 6 years, the epiphysis typically again shows normal MR signal characteristics [18].
Catterall et al. [17] also documented that the proximal femoral physis is abnormal in nearly all cases of LCP disease. Such physeal lesions may be the result of either the primary ischemic insult or secondary to abnormal me­ chanical loads. The normal proximal femoral physis is hypointense on T1­weighted imag­ ing and hyperintense on T2­weighted/STIR imaging [24]. Abnormalities of the physis are best depicted by MRI [8]. MRI findings that suggest possible physeal involvement by LCP disease include increased undulation of the growth plate (W- or M-shaped) (Figs. 7 and 14), deepening of the growth plate or “cup­ ping,” epiphyseal–metaphyseal osseous fu­ sion (bone bridge or bar formation across the physis) (Fig. 12), or physeal cystic change [8, 24]. Abnormal physeal signal and enhance­ ment may also be due to the presence of ab­ normal transphyseal blood vessels (Figs. 2, 5, 7, 9, and 10). Such physeal abnormalities may disturb the physis, leading to premature growth arrest of the femur [8].
An assortment of metaphyseal findings may also be observed in the revasculariza­ tion and reparative phases of LCP disease. At radiography, abnormal metaphyseal radi­ olucencies, or “cysts,” may be seen. At MRI, these areas of radiographic lucency common­ ly have signal characteristics identical to the adjacent physis, suggesting a cartilaginous cause, perhaps ectopic physeal cartilage [8]. Not all metaphyseal “cysts” are due to ectop­ ic cartilage, however [17, 25, 26] (Fig. 15). Kim et al. [26] documented that such radio­ lucencies may be due to extension of physeal cartilage, metaphyseal osseous resorption, or fibrovascular (granulation) tissue deposition. Catterall et al. [17] showed that such radio­ lucencies may be due to increased metaphy­
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 1
71 .2
43 .0
.1 61
o n
03 /1
1/ 23
f ro
m I
P ad
dr es
s 17
1. 24
3. 0.
16 1.
C op
yr ig
ht A
R R
S. F
or p
er so
AJR:193, November 2009 1397
MRI of Legg-Calvé-Perthes Disease
seal adipose tissue, the presence of fibrocar­ tilage, disorganized ossification, or, most commonly, metaphyseal extension of unos­ sified growth plate. In a study by Eckerwall et al. [27], histopathologic examination of 22 core biopsy specimens from the proximal femoral metaphysis of 22 LCP disease pa­ tients revealed fat necrosis, vascular prolifer­ ation, and focal fibrosis. Abnormal metaphy­ seal broadening (coxa magna) (Figs. 9–12) and shortening (coxa breva) (Figs. 9–11) may be observed with healing.
MRI and Prognosis of Legg-Calvé- Perthes Disease
Multiple MRI findings have prognostic value in the setting of LCP disease. First, the extent and distribution of epiphyseal necrosis have prognostic implications. As the extent of femoral head necrosis increases, overall prognosis worsens [9, 28]. Conversely, prog­ nosis is improved when only a small amount of femoral head is affected. Prognosis is also adversely affected by involvement of the lat­ eral column or pillar, the lateral most one third of the femoral head [10, 28, 29]. Con­ versely, preservation of femoral head later­ al column perfusion appears to be associated with an improved prognosis [5, 22] (Figs. 1, 2, 8, and 10).
Disturbance of the physis, including the presence of transphyseal neovascularity (Figs. 2, 5, 7, 9, and 10), is associated with overall worse prognosis, including increased risk of proximal femoral growth disturbance [4, 8, 24]. Up to 90% of femurs affected by LCP disease may show decreased growth and up to 25% of affected hips may have pre­ mature physeal closure at radiography [30]. The affected femur is on average 1.0–1.5 cm shorter than the contralateral normal femur [31]. De Sanctis et al. [24] suggested that ne­ crosis of less than 50% of the proximal fem­ oral epiphysis was never associated with se­ vere physeal involvement, increased lateral extrusion, or metaphyseal changes. They concluded that disturbance of the physis may be the greatest single predictor of poor out­ comes in LCP disease [24].
Subchondral ossific nucleus fracture also has prognostic significance because the ex­ tent of this finding appears to correlate with the extent of eventual necrosis at radiogra­ phy [32, 33]. This finding may also be visible at MRI (Fig. 4). Metaphyseal signal abnor­ malities (Figs. 2, 5, 9, and 15) in the setting of LCP disease, particularly physeal bridg­ ing (Fig. 12),…