Physeal widening observed on MRI of children has been described as the result of a variety of metaphyseal insults [1]. These insults have been produced by the disruption of the metaphyseal blood flow in experimental animal models [2, 3] and in skeletally immature children who participate in high-level sports and sustain repetitive trauma [4]. Apparent widening of the physis has also been described on radiographs in children who have sustained fractures from child abuse [5]. The widening may be broad or more focal, described as “tonguelike” [1].

We have observed broad areas of physeal widening on MRI of the knees of children who presented to the orthopedic clinic with chronic overuse knee pain. The knee pain was not associated with an acute traumatic event, but was specifically associated with an intense sport activity. These children were all high-intensity, competitive elite or subelite athletes who participated in sports on select or traveling teams, beyond the recreational level. The purpose of this retrospective study was to review the MRI characteristics of, the possible mechanism for, and the importance of recognizing broad areas of physeal widening in the knee on MRI of skeletally immature high-level athletes.

Our study group is a retrospective collection of a small number of child athletes who presented between January 1996 and October 2002 to the orthopedic clinic of a children's hospital where they were seen by a pediatric orthopedist for chronic overuse knee pain. Only children who underwent MRI of the knee that revealed physeal widening were included in the study. This collection included the knees of six children (three boys, three girls; age range, 8 years 1 month-15 years 7 months; mean age, 12.3 years) who were all skeletally immature at the time of the initial MRI examination. Each child was a high-level athlete who participated in football, basketball, gymnastics, soccer, or tennis. Five of the six children had conventional radiographs obtained within 3 weeks before the MRI examination. The radiology reports rendered at the time of the initial dictation and images were reviewed retrospectively.

All knees were evaluated with MRI at 1.5 T using routine clinical protocols. Each study included axial and coronal fast spin-echo proton density- and T2-weighted images with fat suppression, sagittal conventional spin-echo proton density- and T2-weighted images with fat suppression, and sagittal fast spin-echo T2-weighted images with fat suppression. Gradient-recalled echo, either 3D spoiled or 2D multiplanar, and T1-weighted images were obtained at the radiologist's discretion. Images were reviewed by a pediatric radiologist to evaluate the signal intensity of the physeal widening (hypointense, isointense, or hyperintense relative to adjacent normal physis) and the length of the largest transverse dimension of widening on coronal or sagittal images.

Clinical charts were reviewed for demographic data, type of sports activity, record of prior injury, treatment technique, duration of therapy, and time to resolution of clinical symptoms. Institutional review board approval for this study was obtained.

After evaluation by the orthopedic surgeon and the MRI interpretation, all children were prescribed rest without or with immobilization. Patients were seen in the orthopedic clinic for follow-up at 3- to 6-week intervals. Follow-up imaging studies included conventional radiographs in five children (obtained at 1-4 months in the compliant children and at 2 years in the noncompliant child) and MR images (obtained at 2 months) in one child.

As part of the inclusion criteria for this study, all children had a recognized area of physeal widening on MRI (Table 1). The six knees showed widening of the distal femoral physis (n = 4) (Figs. 1A, 1B, and 1C), the proximal tibial physis (n = 3) (Fig. 2), and the proximal fibular physis (n = 1) (Fig. 3A). The area of widening was isointense to the adjacent physeal cartilage in all children on all sequences except the fast spin-echo T2-weighted with fat suppression sequence and the conventional spin-echo T2-weighted sequence, where the physeal widening was iso- to hyperintense to physeal cartilage signal. The maximal transverse diameter of the widening ranged from 10 to 27 mm. In five of six MRI examinations, physeal widening was the only abnormality identified. The MRI examination of the basketball player also showed soft-tissue edema anterior to the tibial tubercle suggestive of Osgood-Schlatter disease.

Of the five children who underwent conventional radiography before the MRI examination, abnormal physes were seen in four, but only one report suggested chronic physeal stress as the cause. One child underwent imaging elsewhere, and the official report was not available. The radiographs of all five patients showed the areas of physeal abnormality seen on MRI in retrospective review.

Five children were compliant with the prescribed rest protocol (3-5 weeks). Three children were placed in knee immobilizers (Fig. 1C), one was placed in a cast, and one was prescribed strict rest without external immobilization. The time from the initial orthopedic clinic visit to resolution of clinical symptoms ranged from 23 to 87 days. Of the five compliant children, four had radiographs that showed the knee had returned to near normal or normal in 23-87 days from initial imaging (Fig. 1C). One boy had a follow-up MRI examination at 2 months that showed near complete resolution of the physeal widening (Fig. 3C).

One child was noncompliant and continued vigorous tennis training. At 50 months after her initial clinic visit, she had persistent knee pain even though she had reached skeletal maturity on conventional radiography. She had developed bilateral varus deformity of the knees, already visible at 24 months after her MRI examination (Figs. 4A, 4B, 4C, 4D, and 4E), in addition to tibial stress fractures and talar dome osteochondritis dissecans bilaterally.

Extension of physeal signal intensity into the adjacent metaphysis of bones has been described on MRI in both symptomatic and asymptomatic children [1]. Histologic evaluation of apparent physeal widening on both radiography and MRI shows extension of hypertrophic chondrocytes into the metaphysis [2, 5-7]. Chondrocyte extension into the metaphysis results from a variety of abnormalities including a disruption to the metaphyseal vascular supply, such as from trauma [1, 5], radiation therapy [8], or infection [9]; a failure of normal mineralization of the cartilaginous matrix, such as in rickets or hypophosphatasia [10]; or with forms of dysostosis multiplex [11]. Often, small tonguelike extensions are observed without a specific insult identified.

Broad physeal widening has been described in the upper extremities of athletes—namely, little league pitchers [12, 13] and gymnasts [4, 14-16]. Repetitive trauma has been implicated as the offending cause that produces the changes seen radiologically. Liebling et al. [15] presented a 13-year-old baseball catcher with widening of bilateral distal femoral and proximal tibial physes on conventional radiography, similar to the children in our series.

The metaphyseal vessels supply the physis with calcium, vitamin D, and phosphates needed for calcification of the matrix and play a role in the death of the hypertrophic chondrocytes and subsequent osteogenesis [2]. In the absence of normal blood flow, the normal process of endochondral bone formation is disrupted and long columns of hypertrophic cartilage cells from the physis extend into the metaphysis. This produces the cartilage signal intensity of apparent physeal widening seen on MRI. These areas of physeal widening differ from Salter-Harris type 1 injuries in that no discrete fracture is identified through the cartilage, the widening can be quite focal, and neither epiphyseal nor apophyseal displacement is seen. Salter-Harris fractures are often the result of an acute insult or injury in children, whereas we suggest that broad physeal widening is the result of chronic stress.

The newly formed metaphyseal bone immediately adjacent to the physis is relatively fragile and has poor resistance to compressive forces, such as those from the chronic stress of competitive sports activity [17]. We postulate that this form of repetitive microtrauma results in the equivalent of a stress injury with disruption of the microvascular blood supply to the physis, thereby interrupting normal endochondral bone formation. With strict rest, healing occurs and normal osteogenesis resumes. This has been shown on radiographs of gymnasts' wrists [17].

In our group, the children who were compliant with rest and immobilization improved both symptomatically and radiologically within 3 months. The one noncompliant child not only continued to be symptomatic, but also developed malalignment at her knees. It is important to recognize the widened physis both on radiographs, when apparent, and on MR images because these children should not undergo physical therapy and progressive rehabilitation, which often are prescribed for patients with overuse knee pain [18].

The radiologist may be the first to recognize the physeal widening seen on imaging and to suggest that this finding is a possible form of stress injury in competitive, skeletally immature athletes. MRI in children with overuse pain may be performed to confirm physeal widening detected on radiography and to exclude other injuries that may cause prolonged joint pain. Physeal widening and its significance must be communicated to the referring clinician, and the child should discontinue the offending sport and rest to allow rapid healing. Noncompliance with a rest regimen may result in subsequent malalignment.

Address correspondence to T. Laor.