Pe d i a t r i c I m a g i n g • O r i g i n a l R e s e a r c h Laor et al. MRI of Juvenile OCD

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

Pediatric Imaging Original Research

Juvenile Osteochondritis Dissecans: Is It a Growth Disturbance of the Secondary Physis of the Epiphysis? Tal Laor 1 Andrew M. Zbojniewicz1 Emily A. Eismann2 Eric J. Wall2 Laor T, Zbojniewicz AM, Eismann EA, Wall EJ

Keywords: children, juvenile osteochondritis dissecans, knee, MRI, osteochondritis dissecans DOI:10.2214/AJR.11.8085 Received October 14, 2011; accepted after revision June 5, 2012. 1 Department of Radiology, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, Cincinnati, OH 45229. Address correspondence to T. Laor ([email protected]). 2 Department of Orthopedic Surgery, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH.

CME This article is available for CME credit. AJR 2012; 199:1121–1128 0361–803X/12/1995–1121 © American Roentgen Ray Society

OBJECTIVE. The primary physis is responsible for longitudinal bone growth. Similarly, epiphyseal growth relies on endochondral ossification from the circumferential secondary physis. Physeal injury can result in disruption of normal ossification. The cause of juvenile osteochondritis dissecans (OCD) remains elusive. We hypothesized that juvenile OCD results from an insult affecting endochondral ossification from the secondary physis. The purpose of our study was to evaluate the MRI appearance of the distal femoral epiphysis—particularly the secondary physis—of children with juvenile OCD and to compare these findings with the MRI findings of unaffected children. MATERIALS AND METHODS. Knee MRI examinations of 30 children (age range, 8 years 8 months to 13 years 4 months) with OCD and 30 matched control patients were evaluated for skeletal maturity; location of the OCD lesion, if present; secondary physeal continuity; overlying chondroepiphyseal integrity, contour, and width; signal intensity of subchondral bone; and secondary physeal conspicuity. Variables were compared using chi-square tests. RESULTS. All children were skeletally immature. Condylar lesions were medial in 24 knees and lateral in six knees. All were in the middle one third, posterior one third, or middle and posterior thirds in the sagittal plane. The majority of lesions spanned the intercondylar and middle one third of the femoral condyle in the coronal plane (73%). There was a significant difference between secondary physeal disruption in juvenile OCD condyles compared with unaffected condyles (p < 0.001) and control condyles (p < 0.001). Compared with unaffected and control condyles, the OCD group showed chondroepiphyseal widening (p < 0.001) and subchondral bone edema (p < 0.001) on MRI. Neither chondroepiphyseal integrity nor chondroepiphyseal contour was significantly different between groups (p = 0.21, p = 0.31, respectively). CONCLUSION. MRI of children with OCD consistently showed secondary physis disruption, overlying chondroepiphyseal widening, and subchondral bone edema. We suggest that disruption of normal endochondral ossification may be associated with juvenile OCD.

J

uvenile osteochondritis dissecans (OCD), an acquired disorder of epiphyseal cartilage and adjacent subchondral bone, ultimately can result in partial joint destruction and early degenerative arthritis. It affects children with patent physes and continues to perplex clinicians as to its cause, prognosis, and the most appropriate treatment [1]. Theories about the cause of juvenile OCD include localized trauma, ischemia, and genetic predisposition [1]. Because the cause of juvenile OCD is unclear and is a subject of great debate, the most efficacious treatment approach is uncertain. However, most studies suggest that juvenile OCD lesions have a much greater chance of healing, particularly with nonsurgical therapy, than OCD lesions in adults [1].

Mechanical loading and biologic factors such as hormones influence endochondral growth and ossification [2]. Increased compressive stresses on bone are shown to correspond to decreased rates of bone growth [2]. A disruption in the normal process of physeal cartilage to new metaphyseal bone formation has been produced experimentally in rabbits [3, 4] and has been recognized in disorders affecting the primary (transverse or main) physis of various bones of the growing human skeleton. The effects of mechanical stress and resultant physeal dysfunction have been observed on MRI as apparent physeal widening in the wrists of child gymnasts [5] and around the knee in a variety of child athletes including football, basketball, and soccer players [6] (Fig. 1).

AJR:199, November 2012 1121

Laor et al.

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

Materials and Methods

Fig. 1—14-year-old male competitive soccer player who presented with right knee pain. Sagittal fatsuppressed proton density image (TR/TEeff, 2500/12) shows widening of posterior aspect of distal femoral physis. Hyperintense signal of physis extending into distal metaphyseal region (arrow) is consistent with disruption of normal endochondral ossification and persistent physeal cartilage.

Growth of the epiphyses at the ends of long bones and of epiphyseal equivalents, such as the tarsal bones, carpal bones, and apophyses, results from a process of endochondral ossification similar to that responsible for elongation of bone from the primary physis. A secondary physis mirroring the early spherical and subsequent hemispherical shape of the epiphysis is responsible for the circumferential growth of the secondary center of ossification, also by endochondral ossification [7]. This secondary physis is evident on MRI of growing children [8] (Figs. 2A and 2B) and is similar in appearance to but is much thinner than the primary physis. The younger the child, the more apparent is the secondary physis on fat-suppressed water-weighted sequences. In our clinical practice, we have anecdotally noted that the secondary physis overlying a juvenile OCD lesion routinely is not seen on imaging, whereas it is evident in the unaffected portions of the skeletally immature femoral condyle. We hypothesized that juvenile OCD is the result of an insult that affects the function and integrity of the secondary physis in growing children and that ultimately disrupts the normal process of endochondral ossification. Therefore, the purpose of this study was to evaluate the MRI appearance of the epiphysis—particularly the secondary physis—of children being treated for OCD and to compare these imaging findings with those of unaffected children.

1122

Institutional review board approval was obtained for this HIPAA-compliant study and informed consent was waived. MRI examinations of the knee of 30 children (18 boys and 12 girls; age range, 8 years 8 months to 13 years 4 months; mean age, 11.2 years) who were under the treatment of a single attending orthopedic surgeon for juvenile OCD were identified from the orthopedics database and were retrospectively evaluated. Inclusion criteria were patient age of between 8 and 14 years, clinical and radiographic findings of juvenile OCD, and referral for a clinical MRI examination by the same orthopedic surgeon. Seventeen children had juvenile OCD lesions of the right knee and 13 had lesions of the left knee. The MRI diagnosis of OCD included morphologic and signal abnormalities of the distal femoral chondroepiphysis with a focal area of underlying osseous irregularity of variable extent [9]. The OCD lesion had to be in situ on MRI, whether considered stable or unstable. A PACS query yielded a second cohort of sex-, age-, and side-matched control children (18 boys and 12 girls; age range, 8 years 6 months to 13 years 4 months; mean age, 11.2 years) who had normal findings on radiographs of the same-side knee and had undergone MRI for evaluation of acute knee trauma. The control group children were within 3 months of age of the matched OCD study patient at the time of MRI. These control children also had no clinical record documentation of OCD, systemic illness, arthropathy, steroid use, neoplasm, or musculoskeletal infection. An additional inclusion criterion for control children was no osteochondral injury detected on the knee MRI examination in the same location as the OCD lesion of the corresponding study patient. All MRI examinations were performed at the authors’ institution using a 1.5-T clinical system (n = 39) (Signal Horizon LX EchoSpeed, Excite HD EchoSpeed, or Signa LX EchoSpeed plus, GE Healthcare; or Symphony or Magnetom Espree, Siemens Healthcare) or a 3-T clinical system (n = 21) (Magnetom Trio, Siemens Healthcare; or Achieva, Philips Healthcare). Knees were imaged using a transmit-receive or receive-only extremity coil. All MRI examinations included at least a sagittal and coronal fat-suppressed proton density–, intermediate-, or T2-weighted fast spinecho or turbo spin-echo sequence with the following parameters at 1.5 T: TR range/TEeff range, 2500–4500/17–20, 2500–4500/20–58, or 2500– 4500/68–102; echo-train length, 6–8; and 2 signal averages. The following parameters were used at 3 T: TR/TEeff range, 3000/8–18 or TR range/TEeff range, 3000–3500/56–64; turbo factor, 5–7; and 1 signal average. The section thickness was 3–4

mm with an interslice gap of 0.5–1.5 mm and image matrix (256–512 × 192–256) that varied with patient size. Sagittal and coronal sequences were prescribed from an axial sequence. Each MRI examination, which included a fatsuppressed proton density–, intermediate-, or T2weighted sequence or a combination of those sequences, was evaluated in the region of the juvenile OCD lesion in the study group or in the equivalent location in the control group. Skeletal maturity was determined by assessing the patency of the distal femoral physis as fully open or as closing or closed based on the integrity of the primary physeal signal intensity. In the study group of patients with OCD, the location of the juvenile OCD lesion in the medial or lateral femoral condyle and in the sagittal and coronal planes, with each condyle divided into thirds [10], was noted. The continuity of the secondary physis overlying the juvenile OCD lesion and the continuity of the secondary physis of the unaffected condyle of the same femur were also recorded. In the control group, the secondary physis was assessed at the location corresponding to that of the OCD lesion in the OCD patient. The continuity of the secondary physis was recorded as continuous (uninterrupted around the epiphyseal ossification center, which occasionally was viewed on contiguous images), discontinuous (cannot be followed continuously overlying the OCD location or corresponding location in the unaffected condyle or control patient), or indeterminate (cannot be followed continuously at the site of the OCD lesion or corresponding area in the unaffected condyle or control patient and cannot be seen continuously elsewhere in the condyle). The integrity of the chondroepiphysis, which in children is composed of the not yet ossified epiphyseal cartilage and articular cartilage, overlying the juvenile OCD lesion was characterized as intact, disrupted, or indeterminate (abnormal heterogeneous signal but no definite disruption). The chondroepiphyseal contour was described as smooth, convex, or concave. The width of the chondroepiphysis overlying the juvenile OCD lesion site relative to the width of the adjacent uninvolved chondroepiphysis (perpendicular measurement from the articular surface to the subchondral bone) (Fig. 3) was determined to be the same, increased, or decreased. The water-weighted signal intensity of the epiphyseal bone marrow adjacent to the juvenile OCD lesion on fat-suppressed images was characterized as normal, hyperintense, or hypointense. The imaging plane or planes (coronal, sagittal, or both) and the imaging sequence that best showed the secondary physis in patients with juvenile OCD were noted. If a child had more than one OCD lesion, the more posterior lesion was considered the study lesion. Each MRI examination was evaluated

AJR:199, November 2012

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

MRI of Juvenile OCD

A

B

Fig. 2—MR appearance of normal secondary physis in children with healthy knees. A, Sagittal fat-suppressed T2-weighted image (TR/TEeff, 3000/70) of lateral femoral condyle of right knee of 10-year-old boy shows hyperintense primary or transverse physis (straight white arrow) and thinner secondary or hemispherical physis (black arrow) around ossified epiphysis. Unossified epiphyseal cartilage (asterisk) and hyperintense articular cartilage (curved arrow) also can be seen. B, Coronal fat-suppressed intermediate-weighted image (TR/TEeff, 3000/40) of right knee of 11-year-old girl shows hyperintense primary physis (black arrow) and hyperintense secondary physis (white arrows) of distal femoral epiphysis. Secondary physis of proximal tibial epiphysis (asterisk) also is seen.

Fig. 3—Sagittal fat-suppressed proton density– weighted image (TR/TEeff, 3000/12) of right knee of 10-year-old boy with juvenile osteochondritis dissecans (OCD). Width of chondroepiphysis (dashed line) overlying OCD lesion is measured perpendicular to articular surface and is greater than width of adjacent unaffected cartilage (solid lines).

by two radiologists: a pediatric fellowship–trained radiologist with 20 years’ experience and a musculoskeletal fellowship–trained radiologist with 3 years’ experience. The integrity of the secondary physis on coronal and sagittal images in the juvenile OCD, contralateral unaffected, and control condyles was rated independently by each radiologist. The remaining imaging features were scored by consensus. Because the OCD lesions were obvious on MRI, the

readers could not be blinded to which group a specific child’s MRI examination belonged. Interobserver variability for the scoring of the continuity of the secondary physis in both coronal and sagittal planes of the affected condyle, the unaffected condyle, and the control condyle was determined. For statistical comparisons involving the secondary physis between the groups—namely, the condyles with juvenile OCD lesions, unaffected condyles, and control condyles—only cases in which there was agreement between the two readers were used. Chondroepiphyseal cartilage integrity, contour, and width and bone marrow edema pattern were compared between OCD patients and control patients using the chi-square test. The chi-square test was used also to compare the classification of the secondary physis integrity between the affected and unaffected condyles of OCD patients and to compare the affected condyles of OCD patients and corresponding condyles of control patients. Interobserver reliability of the classification of the secondary physis integrity between the two raters for juvenile OCD, contralateral unaffected, and control condyles was analyzed using the Cohen kappa statistic, with larger kappa values indicating better agreement. Specifically, kappa values less than zero indicate poor agreement; 0–0.20, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement [11]. The reliability between classifications from coronal and sagittal views also was

measured using the Cohen kappa statistic. Percent agreement was calculated as the percentage of cases that were rated the same by both raters.

Results All 30 patients with juvenile OCD and all 30 control patients were considered skeletally immature with patent distal femoral primary physes. The femoral condylar location of the juvenile OCD lesion was medial in 24 knees (80%) and lateral in six knees (20%). On sagittal images, 11 lesions (37%) were in the middle third, five (17%) were in the posterior third, and 14 (47%) spanned the middle and posterior thirds. On coronal images, six (20%) were in the middle third, 22 (73%) spanned the intercondylar and middle thirds, one (3%) spanned the middle and outer thirds, and one (3%) spanned across all thirds. No lesions were isolated to the anterior third on sagittal images or to the lateral third on coronal images of either femoral condyle. Comparisons of the integrity of the secondary physis on both coronal and sagittal imaging planes between the condyles with juvenile OCD and the unaffected condyles and control group condyles are shown in Figure 4. There was a statistically significant difference in secondary physeal integrity between the juvenile OCD group and both the unaffected condyles and the condyles of the control group. The secondary physes in children with juvenile OCD

AJR:199, November 2012 1123

Laor et al.

Percentage of Condyles

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

p < 0.001

p < 0.001

p < 0.001

p < 0.001

100 80 60 40 20 0

OCD Condyles

Contralateral Condyles in OCD Patients

Control Condyles

OCD Condyles

Coronal Plane Indeterminate

Contralateral Condyles in OCD Patients

Control Condyles

Sagittal Plane Discontinuous

Continuous

Fig. 4—Integrity of secondary physis. Bar graph compares integrity of secondary physis of condyles with juvenile osteochondritis dissecans (OCD) and of unaffected contralateral condyles and control condyles as evaluated on coronal and sagittal imaging sequences. There is statistically significant difference between integrity of secondary physis in condyles with juvenile OCD when compared with each of other two groups and in both planes (p < 0.001). Indeterminate = cannot be followed continuously at site of OCD lesion or corresponding area in unaffected condyle or control patient and cannot be seen continuously elsewhere in condyle; discontinuous = cannot be followed continuously overlying OCD location or corresponding location in unaffected condyle or in control patient; continuous = uninterrupted around epiphyseal ossification center, which occasionally was viewed on contiguous images.

Fig. 5—Sagittal fat-suppressed proton density– weighted image (TR/TEeff, 4000/20) of right knee of 11-year-old boy with juvenile osteochondritis dissecans shows disruption (black arrow) of normal secondary physis (white arrow). Overlying cartilage (chondroepiphysis) is wider in region of physeal disruption. There is subtle marrow edema pattern in adjacent subchondral bone.

were significantly more frequently discontinuous (p < 0.001) in both comparisons. Of the cases in which there was agreement between raters (coronal plane, 23 children with OCD; sagittal plane, 24 children with OCD), no child with an OCD lesion showed a continuous secondary physis overlying the lesion in either the coronal or sagittal plane (Figs. 5–7), whereas 23 of the control group (77%) had a continu-

1124

ous secondary physis in the coronal plane and 22 (73%) had a continuous secondary physis in the sagittal plane. Twenty-seven children with OCD (90%) had a continuous secondary physis in the unaffected condyle in the coronal plane and 28 (93%) had a continuous secondary physis in the unaffected condyle in the sagittal plane. The chondroepiphysis overlying the juvenile OCD lesion was intact in 27 children, disrupted in two children, and indeterminate in one child. The chondroepiphysis was intact in all 30 control patients. The contour of the chondroepiphysis was smooth in 59 children from both groups. In one child with OCD, the chondroepiphysis overlying the lesion showed a concave contour. There was no statistically significant difference between the two groups with respect to chondroepiphyseal integrity (p = 0.21) or contour (p = 0.31). The width of the chondroepiphysis overlying the juvenile OCD lesion was increased compared with the width of the immediately adjacent chondroepiphysis in all 30 children and was not increased compared with the adjacent chondroepiphysis in any control child. This difference was statistically significant (p < 0.001). MRI of all children with juvenile OCD showed adjacent bone marrow edema pattern of varying degrees, whereas none of the MR images of the control children showed abnormal bone marrow

signal intensity (Fig. 8) in the corresponding portion of the femoral condyle (p < 0.001). In patients with juvenile OCD, neither imaging plane was significantly better for identifying the secondary physis, which was seen best in the coronal plane in 53%, sagittal plane in 43%, and both planes in 3%. The secondary physis was subjectively determined to be seen best on intermediate-weighted fat-suppressed sequences in 15 children (50%), with the remaining 15 cases divided almost equally between fat-suppressed proton density– and fatsuppressed T2-weighted sequences. There was substantial agreement between the two raters with respect to the integrity of the secondary physis of all condyles in the coronal plane (κ = 0.67; 95% CI, 0.52–0.83) and in the sagittal plane (κ = 0.66; 95% CI, 0.51–0.81). The two raters agreed about MR findings in both imaging planes in 80% of the patients. When there was consensus between raters, the integrity of the secondary physis was classified as the same in both the coronal and sagittal planes for 97% of patients (κ = 0.95; 95% CI, 0.86–1.04). Discussion Juvenile OCD is an abnormality of the chondroosseous zone of the growing skeleton in children that is most frequently seen in the distal femur. It can be a cause of knee pain and subsequent dysfunction [12] and is seen typically beginning in preadolescence. Because more children are participating in competitive sports at younger ages and MRI is used more frequently than previously utilized to assess knee injuries, recognition of juvenile OCD is increasing [12]. The cause of OCD is debated. König in 1887 [13] first used the term “osteochondritis dissecans” in his description of minimal trauma affecting an underlying lesion that resulted in loose bodies of the knee joint. Since that time, many theories regarding the cause of OCD have been postulated including inflammation, genetics, acute trauma resulting in subchondral fracture, ischemia, and repetitive trauma [12, 14]. Investigators have suggested that most OCD lesions in adults arise from incompletely healed juvenile OCD lesions that become clinically apparent after skeletal maturation as opposed to lesions that arise de novo after skeletal growth has stopped [12]. Despite the extensive controversy about the cause of juvenile OCD, most of the studies in the literature currently suggest that most juvenile OCD cases likely result from a repetitive stress injury in very active chil-

AJR:199, November 2012

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

MRI of Juvenile OCD

Fig. 6—Sagittal fat-suppressed T2-weighted image (TR/TEeff, 3000/64) of left knee of 10-year 6-monthold girl with juvenile osteochondritis dissecans (OCD) shows normal secondary physis (arrows) anterior and posterior to lesion but not overlying it. This case was scored as discontinuous secondary physis overlying OCD lesion.

dren, particularly high-level athletes [1, 15]. Affected children frequently present with poorly localized knee pain that has been present for more than 1 year before diagnosis [1, 15] and that often is exacerbated by exercise [12]. Occasionally, juvenile OCD lesions are discovered when imaging is acquired for other reasons. Conventional radiography performed during the evaluation for knee symptoms can show the juvenile OCD lesion, particularly if a tunnel view is obtained. This view optimizes visualization of the posterior half of the femoral condyles, a common location for a juvenile OCD lesion. Conventional radiography also helps to exclude other osseous lesions and to evaluate the degree of skeletal maturity [12]. If a lesion is identified on radiography, MRI frequently is then performed to evaluate the precise location and size of the lesion and the integrity of the overlying chondroepiphysis (both the unossified epiphysis and the articular cartilage) [15]. Like prior descriptions of symptomatic juvenile OCD [12], solitary lesions in our cohort were located within the middle and posterior portions of the femoral condyles. Stable lesions that are less likely to heal have been shown to involve a larger area on MRI than those that do heal [16]. Disruption of normal endochondral ossification from the primary physis as a result of metaphyseal vascular injury has been produced experimentally in rabbits [4, 17]. An appearance similar to that of the experimental models also has been described on MRI of children affected by a variety of insults including child abuse or Salter-Harris fractures,

radiation therapy, and chronic mechanical stress [6, 18, 19]. In each of these scenarios, compromise of the normal physeal endochondral ossification process results in persistent, disorganized physeal cartilage extension into the metaphysis. This change, in turn, results in apparent physeal widening, which can be seen on histologic specimens [17, 18]. Because chronic mechanical stress can affect endochondral ossification from the primary physis, a similar insult might also affect the integrity and function of the much thinner secondary physis. The cellular configuration of the secondary physis and process of endochondral ossification proceed in a direction from the chondroepiphyseal side toward the more central subchondral bone of the developing secondary center of ossification. The newest bone formation therefore occurs at the chondroosseous junction of the growing epiphyseal ossification center (Fig. 9A). Thus, the immediate subchondral bone can be considered a “metaphyseal equivalent.” As such, disruption in the vascular supply to the secondary physis would result in persistent cartilage extending deep and central to the chondroosseous border of the developing epiphysis (Fig. 9B). In this case, the abnormal cartilage extension would be deeper than or more central to the normal chondroosseous contour, resulting in a relative increased width of the overlying chondroepiphysis like that found in all of our juvenile OCD subjects. Because the secondary physis is much thinner than the main physis, the insult of chronic stress on the distal femur might disrupt not only the process of endochondral ossification, but also the architecture of the physis. These disruptions may result in loss of the normal hyperintense signal intensity of the physis compared with the signal of the remaining chondroepiphysis on fat-suppressed water-sensitive MRI sequences, as seen in children with OCD. The abnormal signal that replaces normal subchondral bone mirrors that of the remaining chondroepiphysis on all MRI sequences. The results of studies in the veterinary literature suggest that vascular injury to the epiphysis is responsible for the development of osteochondral lesions that progress to OCD in animals such as rabbits and birds [3, 20, 21]. The vascular injury may be from presumed decreased perfusion pressure through the subchondral capillaries because mechanical pressure on a joint is sustained [22]. Ytrehus et al. [20], suggest that it is unclear whether disruption affects vessels on

the unossified cartilaginous epiphyseal side or the metaphyseal-equivalent side of the secondary physis. We suggest that normal endochondral ossification of the secondary center of ossification is disrupted in children with OCD, but we cannot determine if the initial insult is located within the metaphyseal-equivalent portion of the developing secondary center of ossification or if it is within the overlying chondroepiphysis. Once an insult to the epiphysis, such as compression on a joint, was removed in a rabbit model, normal growth of the epiphysis resumed [3]. Similarly, cessation of an insult in a child with OCD, such as an interruption in chronic high-level athletic activity while the patient rests, allows healing to commence and normal endochondral ossification to resume. This restored growth process may then result in the renewed bone formation observed in healing juvenile OCD lesions. A similar pattern of healing has been observed with the treatment of a stress injury to the primary physis [6]. The process of resumed normal skeletal growth might then account for the improved rates of healing seen in children compared with adults with an OCD lesion. Once skeletal maturation is reached, endochondral bone formation from a physis can no longer occur and healing may be limited. Similar to the results of a prior study in the literature [23], we found a statistically significant difference in the presence of a bone marrow edema pattern in the epiphyses of children with juvenile OCD. The bone marrow edema pattern

Fig. 7—Coronal fat-suppressed proton density– weighted image (TR/TEeff, 3000/13) of right knee of 13-year-old boy with juvenile osteochondritis dissecans (OCD) shows secondary physis, which becomes discontinuous, overlying medial femoral condylar juvenile OCD with minimal adjacent bone marrow edema pattern. White arrow points to the still intact physis. Secondary physis of lateral femoral condyle (black arrow) is continuous along entire contour of ossified epiphysis, extending to intercondylar notch.

AJR:199, November 2012 1125

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

Laor et al. Fig. 8—Coronal fatsuppressed T2-weighted image (TR/TEeff, 3000/64) of right knee of 12-year-old boy with juvenile osteochondritis dissecans (OCD) shows bone marrow edema pattern adjacent to medial femoral condylar juvenile OCD lesion. Secondary physis (white arrow) abruptly stops at edge of OCD lesion but is continuous (black arrow) in unaffected lateral condyle.

may reflect the repetitive microtrauma caused by direct impact during high-stress sporting activity to bone that has lost its full buffering ability [24]—namely, abnormal diminished underlying subchondral bone. Bone marrow edema also has been shown histologically to accompany “osteochondrosis” changes in pigs [20]. In prior studies, investigators have suggested that a bone marrow edema pattern is associated with pain in patients with osteonecrosis of the femoral head and in patients with sickle cell disease [25, 26]. The bone marrow edema pattern seen on MRI of children with OCD similarly may contribute to the symptom of vague knee pain that often occurs in these patients.

Because most of the skeletally immature control condyles and contralateral unaffected condyles in our study showed a continuous secondary physis on MRI in the location corresponding to the OCD lesion and none of the juvenile OCD group did, we speculate that this structure might be helpful in differentiating normal ossification and variations from an OCD lesion. Numerous bones in the developing pediatric musculoskeletal system can develop from irregular centers of ossification, such as the trochlea at the distal end of the humerus and the navicular bone. These structures are considered “epiphyseal-equivalent bones” and, as such, undergo a process

A

of ossification that is analogous to that of the epiphysis of the distal femur [27, 28]. In asymptomatic children, an osseous irregularity on radiography [29] that corresponds to a chondroosseous irregularity on MRI has been deemed by some investigators to be an “ossification or developmental variant.” Differentiating normal development from early OCD is not always possible, particularly if the child has vague knee symptoms [15]. Attempts have been made to differentiate a chondroosseous irregularity of growth from a juvenile OCD lesion on MRI. Investigators have suggested that if the irregularity is located in the posterior femoral condyle, has intact overlying articular cartilage, or is associated with multiple subchondral ossification centers but lacks bone marrow edema pattern, then the irregularity likely represents a developmental variant [10]. Jans et al. [23] suggest that age is a differentiating feature because ossification variability did not occur in girls older than 10 years or boys older than 13 years in their study of 315 children and that OCD was not seen in children younger than 8 years old. They also found that if the physes were nearly fused and if the lesion extended into the intercondylar region, then ossification variability was not likely [23]. Similar to the results of our study and others [23], a perilesional edema pattern was very commonly seen in chil-

B

Fig. 9—Drawings show normal and disrupted endochondral bone formation of epiphysis in children. A, Longitudinal growth of distal femur results from endochondral ossification from distal femoral primary physis. This bone-forming process proceeds in direction from epiphysis to metaphysis as shown by straight arrow on left image. Most newly formed bone is in immediate juxtaphyseal portion of metaphysis. Metaphyseal blood supply is represented by red curvilinear structures. Curved arrows on sides of primary physis show how transverse physis can be arranged in spherical or hemispherical shape to compose secondary physis of epiphysis (shown on right side, drawn out of proportion for illustrative purposes). Endochondral ossification from secondary physis responsible for growth of epiphysis proceeds (arrow, right image) from chondroosseous border centrally toward epiphyseal subchondral bone. B, Disruption in normal vascular supply within metaphysis of distal femur has been shown to result in interference of normal endochondral ossification from primary physis. Consequently, persistent physeal cartilage extends into metaphysis resulting in relative physeal widening (arrow, left image). Similarly, disruption of vascular supply within “metaphyseal equivalent” of subchondral bone of epiphysis may result in cartilage extension and relative widening of secondary physis (arrow, middle image). This appearance of relative cartilage widening is similar to that seen on sagittal MR image of distal femur of child with juvenile osteochondritis dissecans (right image).

1126

AJR:199, November 2012

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

MRI of Juvenile OCD Fig. 10—Developmental ossification variant in 7-year-old boy. A, Sagittal proton density–weighted image (TR/ TEeff, 3000/27) of distal femur obtained after acute traumatic injury shows irregular ossification of posterior one third of lateral femoral condyle (arrow). B, Sagittal fat-suppressed T2-weighted image (TR/TEeff, 3000/64) of same location as A shows continuous secondary physis (arrow) overlying region of osseous irregularity. Lack of physeal disruption and lack of adjacent subchondral bone marrow edema pattern suggest that this appearance results from variation in normal development.

A

B

dren with OCD. Further investigative work with a longitudinal study would be useful to determine whether an intact physis associated with irregular ossification (Fig. 10) of an epiphysis or an epiphyseal-equivalent bone (such as a tarsus, carpus, or apophysis) might be a characteristic appearance of an ossification or developmental variant, particularly if there is no adjacent subchondral bone marrow edema pattern. The limitations of our study include its retrospective nature and the lack of follow-up MRI evaluations of all patients with juvenile OCD. Because the juvenile OCD lesions were obvious, blinding the interpreters as to which patient group an MRI examination belonged was not possible. Also, the level and type of any activity in the control group were not known. Patients were scanned at either 1.5 or 3 T depending on magnet availability according to department of radiology protocol. Field strength was not matched between juvenile OCD and control patients. However, during our clinical practice, we are able to see a normal secondary physis at both 1.5- and 3-T field strengths. The departmental protocol to evaluate the pediatric knee was updated over the study period, so the imaging parameters were not consistent over both groups. The secondary physis often is not well defined in the most posterior condylar location approaching the transverse physis, particularly as the child approaches skeletal maturity. Further evaluation with different imaging sequences based on newer isotropic datasets might help to define the secondary physis to better advantage. Histologic correlation is often not feasible in children with OCD because children with

stable juvenile OCD lesions do not routinely undergo surgery and biopsy is not justifiable. In their description of the histologic composition of OCD lesions, Uozumi et al. [14] did not detail the integrity of a secondary physis or the degree of skeletal maturation of the patients in the study group. However, extensive histologic reports of osteochondral lesions in the veterinary literature that illustrate that abnormal hypertrophic chondrocytes extend from the secondary physis into the adjacent subchondral bone [3, 20, 21, 30, 31] support the assumption that an insult similar to that suggested in animals might be responsible for our findings on MRI of children with OCD. Future study to determine whether a healing or healed juvenile OCD lesion is characterized by revisualization of the secondary physis in a skeletally immature child would support our current supposition and might ultimately be used to indicate a favorable prognosis. Trueta and Trias [3] suggested that discontinuing compression on a joint allows the transverse physis to return to normal function if the initial insult did not cause substantial damage to the cartilage. In summary, we found a statistically significant frequency of discontinuity of the secondary physis overlying an OCD lesion, relative widening of the adjoining chondroepiphysis, and adjacent subchondral bone marrow edema pattern in children with symptomatic lesions compared with the contralateral unaffected condyles and control group condyles. These findings suggest that the chondroosseous irregularity that characterizes these lesions in growing children might be the result of disruption of the nor-

mal process of endochondral ossification from the secondary physis, similar to that described extensively in animals. This abnormality of normal growth may ultimately compromise the integrity of the subchondral bone and result in a bone marrow edema pattern. Further longitudinal study is necessary to identify whether an intact secondary physis is uniformly present in children with an ossification or developmental variation, particularly in children being imaged for other indications. Furthermore, the lack of an intact secondary physis in patients being treated for OCD may explain why some children are resistant to reossification. It has not yet been determined whether the secondary physis can be revisualized over time and, if so, whether this MRI appearance may ultimately be used to indicate a good prognosis. References 1. Wall E, Von Stein D. Juvenile osteochondritis dissecans. Orthop Clin North Am 2003; 34:341–353 2. Chen G, Schuetz M, Pearcy MJ. Mechanobiology of bone development and computational simulations. In: Bronner F, ed. Bone development. London, UK: Springer, 2010:279–295 3. Trueta J, Trias A. The vascular contribution to osteogenesis. IV. The effect of pressure upon the epiphyseal cartilage of the rabbit. J Bone Joint Surg Br 1961; 43:800–813 4. Jaramillo D, Laor T, Zaleske DJ. Indirect trauma to the growth plate: results of MR imaging after epiphyseal and metaphyseal injury in rabbits. Radiology 1993; 187:171–178 5. Shih C, Chang CY, Penn IW, Tiu CM, Chang T, Wu JJ. Chronically stressed wrists in adolescent gymnasts: MR imaging appearance. Radiology

AJR:199, November 2012 1127

Downloaded from www.ajronline.org by 37.44.207.64 on 01/17/17 from IP address 37.44.207.64. Copyright ARRS. For personal use only; all rights reserved

Laor et al. 1995; 195:855–859 6. Laor T, Wall EJ, Vu LP. Physeal widening in the knee due to stress injury in child athletes. AJR 2006; 186:1260–1264 7. Rivas R, Shapiro F. Structural stages in the development of the long bones and epiphyses: a study in the New Zealand white rabbit. J Bone Joint Surg Am 2002; 84:85–100 8. Laor T, Jaramillo D. MR imaging insights into skeletal maturation: what is normal? Radiology 2009; 250:28–38 9. Bohndorf K. Osteochondritis (osteochondrosis) dissecans: a review and new MRI classification. Eur Radiol 1998; 8:103–112 10. Gebarski K, Hernandez RJ. Stage-I osteochondritis dissecans versus normal variants of ossification in the knee in children. Pediatr Radiol 2005; 35:880–886 11. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159–174 12. Kocher MS, Tucker R, Ganley TJ, Flynn JM. Management of osteochondritis dissecans of the knee: current concepts review. Am J Sports Med 2006; 34:1181–1191 13. Barrie HJ. Osteochondritis dissecans 1887–1987: a centennial look at König’s memorable phrase. J Bone Joint Surg Br 1987; 69:693–695 14. Uozumi H, Sugita T, Aizawa T, Takahashi A, Ohnuma M, Itoi E. Histologic findings and possible causes of osteochondritis dissecans of the knee. Am J Sports Med 2009; 37:2003–2008

15. Polousky JD. Juvenile osteochondritis dissecans. Sports Med Arthrosc 2011; 19:56–63 16. Wall EJ, Vourazeris J, Myer GD, et al. The healing potential of stable juvenile osteochondritis dissecans knee lesions. J Bone Joint Surg Am 2008; 90:2655–2664 17. Trueta J, Amato VP. The vascular contribution to osteogenesis. III. Changes in the growth cartilage caused by experimentally induced ischaemia. J Bone Joint Surg Br 1960; 42:571–587 18. Laor T, Hartman AL, Jaramillo D. Local physeal widening on MR imaging: an incidental finding suggesting prior metaphyseal insult. Pediatr Radiol 1997; 27:654–662 19. Kleinman PK, Marks SC Jr, Spevak MR, Belanger PL, Richmond JM. Extension of growthplate cartilage into the metaphysis: a sign of healing fracture in abused infants. AJR 1991; 156:775–779 20. Ytrehus B, Carlson CS, Ekman S. Etiology and pathogenesis of osteochondrosis. Vet Pathol 2007; 44:429–448 21. Riddell C. Studies on the pathogenesis of tibial dyschondroplasia in chickens. III. Effect of body weight. Avian Dis 1975; 19:497–505 22. Charles Louis Davis Foundation for the Advancement of Veterinary and Comparative Pathology Website. Pool R. Role of vasculature in bone and joint disease. www.cldavis.org/cgi-bin/download.gcgi?pid=316. Published 2007. Accessed January 20, 2010 23. Jans LB, Jaremko JL, Ditchfield M, Huysse WC, Verstraete KL. MRI differentiates femoral condylar

ossification evolution from osteochondritis dissecans: a new sign. Eur Radiol 2011; 21:1170–1179 24. Kornaat PR, de Jonge MC, Maas M. Bone marrow edema–like signal in the athlete. Eur J Radiol 2008; 67:49–53 25. Huang GS, Chan WP, Chang YC, Chang CY, Chen CY, Yu JS. MR imaging of bone marrow edema and joint effusion in patients with osteonecrosis of the femoral head: relationship to pain. AJR 2003; 181:545–549 26. Koo KH, Ahn IO, Kim R, et al. Bone marrow edema and associated pain in early stage osteonecrosis of the femoral head: prospective study with serial MR images. Radiology 1999; 213:715–722 27. McCarthy SM, Ogden JA. Radiology of postnatal skeletal development. V. Distal humerus. Skeletal Radiol 1982; 7:239–249 28. Williams GA, Cowell HR. Kohler’s disease of the tarsal navicular. Clin Orthop Relat Res 1981; 158:53–58 29. Caffey J, Madell SH, Royer C, Morales P. Ossification of the distal femoral epiphysis. J Bone Joint Surg Am 1958; 40:647–654 30. Olstad K, Ytrehus B, Ekman S, Carlson CS, Dolvik NI. Epiphyseal cartilage canal blood supply to the distal femur of foals. Equine Vet J 2008; 40:433–439 31. Ytrehus B, Carlson CS, Lundeheim N, et al. Vascularisation and osteochondrosis of the epiphyseal growth cartilage of the distal femur in pigs: development with age, growth rate, weight and joint shape. Bone 2004; 34:454–465

F O R YO U R I N F O R M AT I O N

This article is available for CME credit. Log onto www.arrs.org; click on AJR (in the blue Publications box); click on the article name; add the article to the cart; proceed through the checkout process.

1128

AJR:199, November 2012