Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, cont

Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights. ORIGINAL RESEARCH n PEDIATRIC IMAGING Lennart B. O. Jans, MD Jacob L. Jaremko, MD, PhD Michael Ditchfi eld, MD, PhD Koenraad L. Verstraete, MD, PhD Evolution of Femoral Condylar Ossification at MR Imaging: Frequency and Patient Age Distribution1 Purpose: To determine how the magnetic resonance (MR) signal intensity seen with variability in distal femoral epiphyseal ossification in children varies with (a) age, (b) sex, (c) distribution to the medial or lateral condyles, and (d) residual physeal cartilage. 1 From the Department of Radiology and Medical Imaging, Ghent University Hospital, De Pintelaan 185, 9000 Gent, Belgium (L.B.O.J., K.L.V.); and Department of Medical Imaging, University of Melbourne, Royal Children s Hospital, Melbourne, Victoria, Australia (J.L.J., M.D.). Received June 1, 2010; revision requested July 9; revision received September 10; accepted September 20; fi nal version accepted September 28. Address correspondence to L.B.O.J. (e-mail: lennartjans@hotmail.com ). Materials and Methods: Results: Conclusion: Ethics committee approval was obtained, and informed patient consent was waived. Two pediatric radiologists retrospectively reviewed the consecutive knee MR imaging studies of 910 children (457 boys, 453 girls; age range, 0.7 16.9 years) for variability in ossification and categorized the variability as preossification center, early ossification center, puzzle piece, incomplete puzzle piece, spiculation, or accessory ossification center. Patient age and sex, ossification variability site, residual physeal cartilage, and associated findings were analyzed. Basic descriptive statistical analysis, Student t tests for comparison of continuous variables, and k statistics analysis of interobserver agreement were performed where appropriate. In 202 (22.2%) patients (278 condyles), ossification variability was present. In the 910 patients, early ossification center ( n = 172, 18.9%) and spiculated configuration of the secondary ossification center ( n = 151, 16.6%) were the most common variants. Preossification center (50 [5.5%] patients), puzzle piece (26 [2.9%] patients), accessory ossification center (nine [1.0%] patients), and incomplete puzzle piece (two [0.2%] patients) were seen less often. Ossification variability was more common in the medial condyles (169 [18.6%] of 910 cases) than in the lateral condyles (109 [12.0%] of 910 cases), nearly always posteriorly located (277 [99.6%] of 278 condyles), and more common in boys (153/457 [33.5%]) than in girls (49/453 [10.8%]). Ossification variability was less common with decreasing residual physeal cartilage. Peak patient age ranges for ossification variability were 2 12 years for boys and 2 10 years for girls. Ossification variability in the femoral condyles is common in children and should not be confused with abnormal processes. q RSNA, 2010 q RSNA, 2010 880 radiology.rsna.org n Radiology: Volume 258: Number 3 March 2011

Skeletal growth and maturation in children are dynamic processes that can be documented by using magnetic resonance (MR) imaging. While epiphyseal cartilage is homogeneous during early infancy, with advancing age the signal intensity in the posterior femoral condyles increases and becomes progressively more focal ( 1 ). This is a normal age-related variation and should not be confused with disease ( 2 ). Although the cartilaginous distal femoral epiphysis has been recognized to have heterogeneous signal intensity ( 2,3 ), there is a paucity of literature addressing how MR findings of the distal femoral epiphysis change during infancy and childhood. With increasing reliance on MR imaging for evaluation of knee disease in children, an understanding of the normal MR imaging appearance of the developing distal femoral condyle is important for distinguishing normal developmental variations in this structure from disease ( 1 ). The aims of this study were to determine the prevalence of variability in femoral condyle ossification in children and to demonstrate how the signal intensity in the distal femoral epiphysis of children varies with (a) age, (b) sex, (c) position in the medial or lateral condyles, and (d) residual physeal cartilage on MR images. Materials and Methods Study Group This retrospective study was performed in a pediatric tertiary care center. Ap- Advances in Knowledge n Age-related variability in ossifica- tion of the femoral condyles in children is common and should not be confused with epiphyseal abnormality. n A high-signal-intensity early ossi- fication center and a spiculated configuration of the secondary ossification center are the most common variant findings. n Variability in ossification of the femoral condyles is more common in boys than in girls. proval for the study was obtained from the Melbourne Royal Children s Hospital ethics committee, and informed patient consent was waived. Subjects younger than 17 years who had undergone MR imaging of the knee were identified by means of a radiology information system database search of final radiology reports for the key words MRI and knee, covering a period of 10 years (2000 2010). If the patient had undergone more than one MR examination, only the first set of MR images was reviewed. The MR images were analyzed retrospectively in consensus by two fellowship-trained pediatric radiologists (L.B.O.J., J.L.J.), each with 7 years of experience. Patient age groups were defined according to chronologic age, in years (eg, 0 1 years, 1 2 years). If an age group included more than 50 boys and 50 girls, we selected from among the available images the images of the first 50 boys and 50 girls in chronologic order for assessment. This also applied to our data set of older children those aged 11 16 years. The indications for MR imaging and the inclusion and exclusion criteria are presented in Figure 1. Although it was not possible to blind the reviewing radiologists to the original reported findings, they used the original report only as a means of identifying patients for inclusion in the study. The consecutive MR studies of 1971 patients were reviewed. Of these patients, 910 children (mean age, 11.5 years; median age, 12.0 years; age range, 0.7 16.9 years) met the criteria for inclusion in the study. This final study group consisted of 457 boys (50.2%) and 453 girls (49.8%). The mean ages of the boys (11.51 years) and girls (11.49 years) included in the study were comparable. The right knee was imaged in 499 patients (54.8%), and the left knee was imaged in 411 patients (45.2%). MR Imaging Examinations MR imaging was performed by using one of two 1.5-T MR systems: The GE Implication for Patient Care n Recognizing variability in femoral condyle ossification as a normal finding in children may prevent unnecessary treatment. EchoSpeed LX system (GE Medical Systems, Milwaukee, Wis) was used from January 2000 through December 2005, and the TIM Avanto system (Siemens Medical Systems, Erlangen, Germany) was used from January 2006 through December 2009. The following MR sequences were performed: coronal and sagittal T1-weighted imaging, axial and sagittal fat-saturated T2-weighted imaging, coronal and sagittal fat-saturated proton-density weighted imaging, and three-dimensional volumetric spoiled gradient-echo imaging. The MR imaging parameters are summarized in Table 1. The affected knee was imaged by using a four-channel transmit-receive phasedarray coil (Medrad, Warrendale, Pa) with the GE EchoSpeed LX unit and by using an eight-channel phased-array coil or 15-channel transmit-receive coil (TxRx; Siemens Medical Systems) with the TIM Avanto unit. In 132 patients, axial, sagittal, and coronal fat-saturated contrast material enhanced T1-weighted imaging was performed with intravenous administration of 0.2 ml of diluted gadopentetate dimeglumine (Magnevist; Bayer Healthcare Pharmaceuticals, Berlin, Germany) per kilogram of body weight. Review of MR Studies We retrospectively reviewed and recorded the demographic and imaging features of variability in femoral condyle ossification. These features included patient age and sex, ossification variability Published online before print 10.1148/radiol.10101103 Radiology 2011; 258:880 888 Abbreviation: CI = confi dence interval Author contributions: Guarantors of integrity of entire study, L.B.O.J., K.L.V.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript fi nal version approval, all authors; literature research, L.B.O.J., K.L.V.; clinical studies, L.B.O.J., M.D.; statistical analysis, L.B.O.J., J.L.J., K.L.V.; and manuscript editing, all authors Potential confl icts of interest are listed at the end of this article. Radiology: Volume 258: Number 3 March 2011 n radiology.rsna.org 881

site, lesion configuration, residual physeal cartilage, and presence or absence of bone marrow edema. The presence or absence of associated findings such as loose bodies and joint effusion also was recorded. Joint effusion was diagnosed when the suprapatellar bursa was distended by more than 4 mm. The midline suprapatellar pouch was measured by determining the anteroposterior distance of the synovial fluid at its widest point in the midline of the suprapatellar pouch ( 4 ). Site of variability was classified as right or left knee; medial or lateral femoral condyle; external, central, or internal third of the condyle; and anterior, middle, or posterior third Figure 1 Figure 1: imaging. of the condyle in accordance with the classifications used by Harding ( 5 ) and Hughston et al ( 6 ) ( Fig 2 ). The preossification center represents the initial structure of the secondary ossification center. This is followed by invasion of the preossification center by vessels that bring osteoblast precursors, resulting in an early ossification center ( 7,8 ). Variability in ossification of the residual physeal cartilage was categorized as preossification center or early ossification center. Preossification center was defined as a well-defined lobulated focus of high signal intensity on T2-weighted MR images ( 7,9 ). Early ossification center was defined as a focal ill-defined Flowchart of study population, with inclusion and exclusion criteria and indications for MR zone of intermediate to high signal intensity in the cartilage on T2-weighted images. We categorized the configuration of ossification variability in the secondary ossification center (or centers) as puzzle piece, incomplete puzzle piece, spiculated configuration, and/or accessory ossification center in accordance with the categorization used by Gebarski and Hernandez ( 10 ). Puzzle piece referred to an osseous defect filled with an ossification that fit the defect, and incomplete puzzle piece referred to a defect with incomplete ossification or no ossification filling the defect ( 10 ). Measurements were performed by using 2009 Fuji Synapse, version 3.2 (Fujifilm Medical Systems, Stamford, Conn), image-viewing software. The relative size of the nonossified physeal cartilage was calculated as the ratio (expressed as a percentage) of the thickness of the remaining cartilage to the distance between the center of the secondary ossification center and the outer rim of articular cartilage, measured at the junction of the central and posterior thirds of the femoral condyle in the sagittal plane. Relative sizes were categorized into 0% 9%, 10% 19%, 20% 29%, 30% 39%, 40% 49%, and 50% or greater groups in accordance with the categorization used by Gebarski and Hernandez ( 10 ). Bone marrow edema was deemed to be present if the signal intensity of Table 1 MR Imaging Sequences and Parameters GE EchoSpeed LX System TIM Avanto System Parameter T1-weighted T2-weighted Fast Recovery Proton-density weighted T1-weighted T2-weighted RESTORE * Proton-density weighted T2-weighted Spoiled GRE TR/TE 400/12 6000/85 3600/48 580/14 6080/72 3000/33 37/20 Echo train length 3 15 5 128 35 31 1 Field of view (mm) 140 x 140 140 x 140 140 x 140 120 x 75 180 x 75 180 x 100 140 x 140 Section thickness (mm) 3.5 3.5 3.5 3 3 3 0.7 Matrix 320 x 256 256 x 224 320 x 256 512/512 512/512 512/512 256/256 Bandwidth (Hz/pixel) 16 16 25 130 65 195 60 Flip angle (degrees) 145 90 90 145 180 180 12 No. of signals acquired 1 1 1 1 1 1 1 Note. With the exception of three-dimensional volumetric spoiled gradient-echo (GRE) imaging (T2-weighted spoiled GRE), all sequences were fast spin-echo examinations and all T2-weighted and proton-density weighted sequences were performed with fat saturation. * RESTORE = robust estimation of tensors by outlier rejection. TR/TE = repetition time msec/echo time msec. 882 radiology.rsna.org n Radiology: Volume 258: Number 3 March 2011

the bone marrow was higher than that of muscle on T2-weighted fat-saturated images. The spectrum of ossification variabilities is illustrated on representative images of several knees of patients in different age groups in Figs 3 8. The follow-up MR images obtained in patients with ossification variability who underwent more than one examination were reviewed side by side with the baseline images. The evolution of the variability was described as follows: no longer visible, decreased in size, unchanged, increased in size, or new variability present. We also noted the presence or absence of focal loss of adjacent articular cartilage and of bone marrow edema. Few patients in the included study group had follow-up images. In 12 (0.06%) (16 condyles) of the 1971 patients with ossification variability at the initial examination, a second MR examination was performed a mean of 593 days (median, 460 days; range, 46 1332 days) after the first examination, and one patient underwent a second follow-up MR examination. were performed, where appropriate, with the number of study patients ( n = 910) serving as the denominator for each result. Ninety-five percent confidence intervals (CIs) were computed. Continuous variables were compared by using the Student t test, with P =.05 considered to indicate significance. With data obtained from independent readings of 100 randomly chosen MR Figure 3 studies, k values with standard errors were used to assess agreement between the two radiologists. k Values of 0 0.20 were considered to indicate slight agreement; values of 0.21 0.40, fair agreement; values of 0.41 0.60, moderate agreement; values of 0.61 0.80, substantial agreement; values of 0.81 0.99, almost perfect agreement; and a value of 1.00, perfect agreement ( 11 ). Statistical Analyses All statistical analyses were performed by one author (L.B.O.J.) by using statistical software (SPSS 15; SPSS, Chicago, Ill). Basic descriptive statistical analyses Figure 2 Figure 3: Sagittal MR images of knee show puzzle piece variability of secondary ossifi cation center in right knee of 7-year-old girl. (a) T1-weighted image (580/14) shows ill-defi ned defect in posterior third of lateral femoral condyle; ossifi cation (arrow) completely fi lls the defect. (b) Fat-saturated T2-weighted image (6080/72) shows no bone marrow edema. Figure 4 Figure 2: Sagittal fat-saturated T2-weighted MR image (6080/72) in 10-year-old boy illustrates variability site classifi cation. Anterior (A), central (C), and posterior (P) positions are defi ned in relation to a line extending from the posterior femoral diaphyseal line distally and a line extending anteriorly from the roof at the intercondylar notch. The positions were estimated by scrolling through the sagittal images. Figure 4: Sagittal MR images of knee show incomplete puzzle piece ossifi cation variability in right knee of 8-year-old boy. (a) T1-weighted image (580/14) shows ill-defi ned defect (long arrow) in posterior third of lateral femoral condyle. In cranial aspect, an ossifi cation (short arrow) is incompletely fi lling the defect. (b) Fat-saturated T2-weighted image (6080/72) shows no bone marrow edema. Radiology: Volume 258: Number 3 March 2011 n radiology.rsna.org 883

Results Variability in ossification was present in 202 (22.2%; 95% CI: 19.5%, 24.9%) patients (278 condyles). Ossification variability was significantly more common in boys ( n = 153, 33.5%; 95% CI: 30.4%, 36.5%) than in girls ( n = 49, 10.8%; 95% CI: 8.9%, 12.9%) ( P,.0001). Figure 5 Ossification Variability versus Age The mean age at which ossification variability was present was significantly older among the boys (8 years) than among the girls (6 years) ( P,.0001). The prevalence of ossification variability according to age in the study population is summarized in Figure 9. Ossification variability was not seen in patients with 10% or less residual physeal cartilage and was most common in boys and girls with 30% or more residual physeal cartilage ( Fig 10 ). Types of Ossification Variability The prevalences of various forms of variability in femoral condyle ossification are summarized according to age in Figure 11. In the 910 patients, early ossification center ( n = 172, 18.9%; 95% CI: 16.4%, 21.4%) and spiculated configuration of the secondary ossification center ( n = 151, 16.6%; 95% CI: 14.2%, 19.0%) were the most common changes. Preossification center ( n = 50, 5.5%; 95% CI: 4.0%, 7.0%) and puzzle piece ( n = 26, 2.9%; 95% CI: 1.8%, 3.9%) were less common. Extra ossification center ( n = 9, 1.0%; 95% CI: 0.3%, 1.6%) and incomplete puzzle piece ( n = 2, 0.2%; 95% CI: 0%, 0.5%) were rarely seen. Figure 5: Sagittal MR images of knee show spiculated confi guration of secondary ossifi cation center in right knee of 9-year-old boy. (a) T1-weighted image (580/14) shows irregular confi guration (arrows) in posterior third of lateral femoral condyle. (b) Fat-saturated T2-weighted image (6080/72) shows no bone marrow edema. Ossification Variability versus Site The sites of variability in femoral condyle ossification, summarized in Table 2, were calculated as percentages of the presence of every ossification variability in each ninth of the femoral condyle. Among the 910 patients, the proportion of medial condyles ( n = 169, Figure 6 Figure 7 Figure 8 Figure 6: Sagittal T1-weighted MR image (580/14) of right knee in 10-year-old boy shows extra ossifi cation centers (arrows) in nonossifi ed cartilage in posterior third of lateral femoral condyle. Figure 7: Sagittal fat-saturated T2-weighted MR image (6080/72) of right knee in 9-year-old boy shows ill-defi ned area of high signal intensity (arrow), representing early ossifi cation center, in articular cartilage in posterior third of medial femoral condyle. Figure 8: Sagittal fat-saturated T2-weighted MR image (6080/72) of right knee in 9-year-old boy shows well-defi ned area of high signal intensity (arrow), representing preossifi cation center, in articular cartilage of posterior condyle. 884 radiology.rsna.org n Radiology: Volume 258: Number 3 March 2011

18.6%; 95% CI: 16.0%, 21.1%) with ossification variability was significantly greater than the proportion of lateral condyles ( n = 109, 12.0%; 95% CI: 9.9%, 14.1%) with this variability ( P,.0001). Seventy-six patients (8.4%; 95% CI: 6.6%, 10.1%) had ossification variability in both condyles. The majority of the cases of ossification variability ( n = 277, 99.6%; 95% CI: 99.1%, 100.0%) occurred in or extended to the posterior third of the femoral condyle. No variability was seen in the anterior third of the femoral condyle. Associated Findings Bone marrow signal intensity greater than muscle signal intensity indicating bone marrow edema was not seen in association with ossification variability. No loose bodies were seen. Joint effusion was present in 147 patients (16.2%). The prevalences of joint effusion in the study group without variability of ossification (120 [17.0%] of 708 cases) and in the group with variability of ossification (27 [13.4%] of 202 cases) were comparable. spiculated configuration of the secondary ossification center in one patient and an early ossification center in the other patient. None of these patients had focal cartilage loss. Bone marrow signal intensity greater than muscle signal intensity indicating bone marrow edema was not seen in association with the sites of variable ossification. An example of the evolution of a puzzle piece defect is shown in Figure 12. Figure 9 Discussion Radiographic irregularity of the ossification of the distal femoral epiphyses is well known. Sontag and Pyle ( 12 ) were the first to note that the irregularity of the epiphyseal outline of the distal femoral epiphysis on radiographs obtained in children was normal and were able to correlate the changes with periods of rapid growth. Interobserver Agreement The k values indicating agreement between the two radiologists are shown in Table 3. There was perfect, almost perfect, or substantial interobserver agreement in measuring or determining the presence of the MR imaging findings. Follow-up MR imaging studies. In the cases in which follow-up examinations were performed, the variability in ossification on the initial MR imaging studies consisted of a spiculated configuration of the secondary ossification center (13 condyles), an early ossification center (12 condyles), a preossification center (two condyles), or a puzzle piece configuration (one condyle). On the follow-up MR imaging studies, the majority of variable ossification configurations were no longer visible (eight cases) or had decreased in size (15 cases). Ossification variability was unchanged in two condyles. In three condyles, the early ossification center had increased in size. A new ossification variability was seen in two patients: a Figure 9: Graph illustrates patient age based distribution of variability in femoral condyle ossifi cation. Ossifi cation variability was mainly seen in young children, with decreased prevalence (vertical axis) in older children. At age 9 years in girls and age 13 years in boys, ossifi cation variability was present in only about 10% of patients. After ages 12 years in girls and 14 years in boys, ossifi cation variability was no longer seen. Figure 10 Figure 10: Graph illustrates percentages of residual physeal cartilage in association with prevalence of ossifi cation variability (vertical axis). Variability in ossifi - cation was most frequent in patients with 30% or more residual physeal cartilage and was rarely seen in patients with 20% or less residual physeal cartilage. Radiology: Volume 258: Number 3 March 2011 n radiology.rsna.org 885

Caffey et al ( 13 ) examined the frequency and form of irregular ossifications by reviewing the knee radiographs obtained in 147 children aged 3 13 years, all of whom were known to be free of clinical symptoms. Irregularity was pres ent in 66% of the boys and 44% of the girls. Only the lateral condyle was involved in 44% of these children, and Figure 11 only the medial condyle was involved in 12%. The lateral predominance of radiographic changes differs from the medial predominance of MR imaging changes that we observed. We believe that this may be explained by the high prevalence of signal intensity changes in the early ossification center embedded in the residual physeal cartilage of the Figure 11: Graphs illustrate types of ossifi cation variability among (a) boys and (b) girls in each age group. The age-based distributions of types of variability in femoral condyle ossifi cation indicate that early ossifi cation center (T2 intense lesion) of residual cartilage and spiculated confi guration of secondary ossifi cation center were most common ossifi cation variabilities in both boys and girls. All ossifi cation variabilities had wide age-based distribution. Vertical axis values are prevalences of ossifi cation variability, in percentages. medial femoral condyle, which may be visible on MR images but not on conventional radiographs. Epiphyseal cartilage is the precursor to the ossified ends of bone. Increased signal intensity on T2-weighted MR images is seen during normal enchondral ossification when cartilage cells undergo hypertrophy ( 14,15 ). The area of high signal intensity is most apparent in the posterior aspect of the distal femoral epiphysis and can be quite discrete, resembling an epiphyseal cartilaginous blister ( 1,16 ). The preossification center represents the initial structural change in the development of the secondary ossification center. The short interval between preossification development and early ossification would explain why the preossification center is infrequently imaged ( 7,9 ). In our study, we observed a lower prevalence of preossification center compared with early ossification center, reflecting the short existence of the preossification center. Early ossification center was the most common ossification variability seen on the MR imaging studies of the children s knees in this study. At times, ossification of the epiphysis from the secondary physis results in numerous small foci that eventually coalescence with underlying bone. These areas of irregular ossification can be confused with subchondral bone fragments, as seen with osteochondritis dissecans ( 16 ). The lack of adjacent bone marrow edema in the epiphysis and the more posterior location on sagittal MR images indicate that this variation in ossification represents normal development ( 10,17 ). In our study, extra ossification centers were a rare (in nine [1.0%] patients) ossification variability. We believe this may have been because extra ossification centers often fuse partly to the secondary ossification center, producing the more common spiculated configuration of the secondary ossification center. Gebarski and Hernandez ( 10 ) were the first to state that ossification defects in the posterior femoral condyles with intact overlying cartilage, accessory ossification centers, spiculated configuration, residual physeal cartilage, and lack of bone marrow edema are features of ossification variability. It is important to 886 radiology.rsna.org n Radiology: Volume 258: Number 3 March 2011

differentiate variability in femoral condyle ossification from osteochondral lesions and bone lesions such as infarcts, especially in patients receiving steroid therapy. The presence of perilesional bone marrow edema may be helpful for distinguishing abnormal processes from the normal evolution of ossification of the femoral condyles ( 12 ). Our study results confirm that the described MR features of ossification variability are common in young children, especially boys, and should not be considered abnormal. We found that the prevalence of variable femoral condylar ossification decreases with age. On the follow-up MR images, the majority of the variabilities had changed in appearance, and most of them were no longer visible or had decreased in size. Table 2 Locations of Ossification Variability in Femoral Condyle Condyle Region Our study had several limitations. Since this was a retrospective review, we were not able to determine whether trauma that had occurred a relatively long time before the MR imaging studies may have been a cause of the ossification variability. The study design was further limited because the study population was identified by means of radiology report word searches. We did not have surgical or pathologic correlation of the extent of the lesions, and only limited longitudinal follow-up data were obtained. Thus, the estimated prevalences are subject to potential referral bias. Some patients in our study population may have had symptoms that the clinicians believed did not warrant MR imaging. However, the wide range of ossification variants imaged in all Medial Condyle Lateral Condyle Anterior Third Central Third Posterior Third Anterior Third Central Third Posterior Third Internal 0 7 39 0 0 20 Central 0 11 86 0 1 93 External 0 2 28 0 0 56 Note. Data are percentages of the presence of ossifi cation variability in each ninth of the femoral condyle. The majority of the cases of ossifi cation variability occurred in the posterior third of the femoral condyle. Ossifi cation variability extended to the central third of the medial femoral condyle in about 10% of patients and only very rarely in the lateral femoral condyle. Extension to the anterior third was not observed. Seventy-six patients had ossifi cation variability in both condyles. pediatric age groups suggests that there was probably a low clinical threshold for referral, and our key word search was as exhaustive as was practical. A final limitation was the unequal distribution of ages in our study group with fewer children being younger than 6 years due to the rarity of MR imaging examinations performed in this age group. Our study results indicate that there is a high prevalence of variable femoral Table 3 Agreement between Two Radiologists Regarding MR Findings in Substudy Involving 100 Patients Parameter Mean k Value Presence of ossifi cation 0.948 6 0.036 variability Residual physeal cartilage 0.917 6 0.036 measurement Type of variability 0.947 6 0.037 Presence of bone marrow 1.000 edema Presence of effusion 1.000 Intracondylar extension of 1.000 variability Site of variability, sagittal 0.946 6 0.052 plane Site of variability, coronal 0.723 6 0.184 plane Note. Mean k values 6 standard errors of the mean are given. Figure 12 Figure 12: Evolution of puzzle piece ossifi cation variability in 10-year-old boy. Sagittal T1-weighted MR images (580/14) of knee obtained (a) initially and (b, c) at follow-up 2 months ( b ) and 12 months ( c ) after initial MR imaging examination show decreasing size of puzzle piece ossifi cation variability in posterior aspect of femoral condyle. Radiology: Volume 258: Number 3 March 2011 n radiology.rsna.org 887

condylar ossification among boys aged 2 12 years and girls aged 2 10 years, and this should not be considered abnormal. We believe that this variability can be explained by advancing ossification in the physeal cartilage as a child matures. It is important to recognize this morphologic variability and the signal intensity changes that may occur with development so that these can be distinguished from epiphyseal disease. Disclosures of Potential Conflicts of Interest: L.B.O.J. No potential conflicts of interest to disclose. J.L.J. No potential conflicts of interest to disclose. M.D. No potential conflicts of interest to disclose. K.L.V. No potential conflicts of interest to disclose. References 1. Varich LJ, Laor T, Jaramillo D. Normal maturation of the distal femoral epiphyseal cartilage: age-related changes at MR imaging. Radiology 2000 ; 214 ( 3 ): 705 709. 2. Jaramillo D, Shapiro F. Growth cartilage: normal appearance, variants and abnormalities. Magn Reson Imaging Clin N Am 1998 ; 6 ( 3 ): 455 471. 3. Jaramillo D, Hoffer FA. Cartilaginous epiphysis and growth plate: normal and abnormal MR imaging findings. AJR Am J Roentgenol 1992 ; 158 ( 5 ): 1105 1110. 4. Schweitzer ME, Falk A, Berthoty D, Mitchell M, Resnick D. Knee effusion: normal distribution of fluid. AJR Am J Roentgenol 1992 ; 159 ( 2 ): 361 363. 5. Harding WG 3rd. Diagnosis of ostechondritis dissecans of the femoral condyles: the value of the lateral x-ray view. Clin Orthop Relat Res 1977 ;( 123 ): 25 26. 6. Hughston JC, Hergenroeder PT, Courtenay BG. Osteochondritis dissecans of the femoral condyles. J Bone Joint Surg Am 1984 ; 66 ( 9 ): 1340 1348. 7. Shapiro F. Developmental bone biology. In: Shapiro F, ed. Pediatric orthopedic deformities. San Diego, Calif : Academia Press, 2001 ; 21 25. 8. 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-A ( 1 ): 85 100. 9. Chapman VM, Nimkin K, Jaramillo D. The pre-ossification center: normal CT and MRI findings in the trochlea. Skeletal Radiol 2004 ; 33 ( 12 ): 725 727. 10. Gebarski K, Hernandez RJ. Stage-I osteochondritis dissecans versus normal variants of ossification in the knee in children. Pediatr Radiol 2005 ; 35 ( 9 ): 880 886. 11. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977 ; 33 ( 1 ): 159 174. 12. Sontag LW, Pyle SI. Variations in the calcification pattern in epiphyses. Am J Roentgenol 1941 ; 45 : 50 54. 13. Caffey J, Madell SH, Royer C, Morales P. Ossification of the distal femoral epiphysis. J Bone Joint Surg Am 1958 ; 40-A ( 3 ): 647 654, passim. 14. Magnusson MM, Jaramillo D, Zaleske DJ. Magnetic resonance imaging of the epiphysis: an experimental study. Iowa Orthop J 1993 ; 13 : 79 84. 15. Laor T, Jaramillo D. MR imaging insights into skeletal maturation: what is normal? Radiology 2009 ; 250 ( 1 ): 28 38. 16. Nawata K, Teshima R, Morio Y, Hagino H. Anomalies of ossification in the posterolateral femoral condyle: assessment by MRI. Pediatr Radiol 1999 ; 29 ( 10 ): 781 784. 17. Ogino S, Huang T, Watanabe A, Iranpour- Boroujeni T, Yoshioka H. Magnetic resonance imaging of articular cartilage abnormalities of the far posterior femoral condyle of the knee. Acta Radiol 2010 ; 51 ( 1 ): 52 57. 888 radiology.rsna.org n Radiology: Volume 258: Number 3 March 2011