Diagnosis of Osteomyelitis in Children: Utility of Fat-Suppressed Contrast-Enhanced MRI

Transcription

1 Pediatric Imaging Original Research verill et al. MRI of Osteomyelitis Pediatric Imaging Original Research FOUS ON: Lauren W. verill 1,2 ndrea Hernandez 1 Leonardo Gonzalez 1 ndres H. Peña 1 Diego Jaramillo 1 verill LW, Hernandez, Gonzalez L, Peña H, Jaramillo D Keywords: children, contrast material, gadolinium, MRI, osteomyelitis DOI: /JR Received November 9, 2007; accepted after revision October 17, Department of Radiology, hildren s Hospital of Philadelphia, Philadelphia, P Present address: Department of Medical Imaging, lfred I. dupont Hospital for hildren, 1600 Rockland Rd., Wilmington, DE ddress correspondence to L. W. verill (lwaverill@yahoo.com). JR 2009; 192: X/09/ merican Roentgen Ray Society Diagnosis of Osteomyelitis in hildren: Utility of Fat-Suppressed ontrast-enhanced MRI OJETIVE. The purpose of this study was to determine whether the use of fat-suppressed contrast-enhanced MRI, compared with unenhanced MRI alone, increases reader confidence in the diagnosis of osteomyelitis and its complications in children. MTERILS ND METHODS. MRI studies of 78 skeletally immature children and adolescents (median age, 3.6 years) with suspected nonspinal osteomyelitis were reviewed in consensus by two readers. Unenhanced images were evaluated first and then contrast-enhanced MR images. Images were scored for the presence or absence of osteomyelitis, abscess, septic arthritis, and physeal involvement on a 5-point scale ranging from definitely absent to definitely present. Forty-two additional studies were evaluated to test interobserver agreement. RESULTS. Osteomyelitis was clinically diagnosed in 40 cases (51%). There was no significant difference between the sensitivity and specificity of unenhanced MRI (p = 1.0) and those of contrast-enhanced MRI (p = 0.77) for the diagnosis of osteomyelitis. Nonetheless, there was a significant (p < 0.001) increase in confidence in the diagnosis of osteomyelitis and its complications. This increase in confidence was most pronounced for the diagnosis of abscess (46%). The addition of contrast enhancement was least useful in findings deemed definitely absent on unenhanced MR images. ONLUSION. lthough it does not increase the sensitivity or specificity of the diagnosis, use of contrast-enhanced MRI does increase reader confidence in the diagnosis of osteomyelitis and its complications in cases in which bone or soft-tissue edema is found on unenhanced images. In the clear absence of edema on unenhanced images, however, contrast enhancement is not needed. M RI has been found useful in the diagnosis of clinically suspected osteomyelitis in both adults and children [1 11]. It has become a primary advanced imaging technique for the diagnosis of osteomyelitis after initial radiographs have been obtained. MRI has the advantages of excellent tissue characterization and high resolution, depicting the presence of both osteomyelitis and its complications, which include soft-tissue and bone abscesses, physeal involvement, and septic arthritis [4, 12, 13]. This added information can alter clinical management, such as determining the need for percutaneous or surgical drainage in addition to antibiotic therapy [13]. Previous investigators have found contrastenhanced MRI useful in the diagnosis of osteomyelitis [1, 2, 4, 5]. dministration of IV gadolinium contrast material increases lesion conspicuity. In one series [5], it increased specificity in the diagnosis of osteomyelitis from 53% to 93%. nother study, however, showed that bone marrow inflammation was equal in extent and conspicuity on both T2- weighted images and contrast-enhanced T1- weighted images [14]. Most of these studies included only adults or mixed populations of adults and children. lthough the use of fatsuppressed contrast-enhanced MRI for evaluation of pediatric osteomyelitis is widespread, the advantage of using this technique for detection of infection and characterization of complications has not been evaluated, to our knowledge. The single study that did focus exclusively on children included only 10 patients who underwent contrast-enhanced MRI [1], and fat saturation technique, which has become the standard of care, was not used. To our knowledge, in no previous study has the added benefit of fat-suppressed contrastenhanced MRI over unenhanced MRI been systematically investigated for the diagnosis of osteomyelitis in a pediatric population JR:192, May 2009

2 MRI of Osteomyelitis The purpose of the study was to determine whether the use of fat-suppressed contrastenhanced MRI, compared with unenhanced MRI alone, increases reader confidence in the diagnosis of osteomyelitis in children. The diagnosis of complications related to osteomyelitis, including bone and soft-tissue abscess, physeal involvement, and septic arthritis, also was investigated. Materials and Methods The study was performed after approval was granted by the institutional review board at our hospital and in full accordance with all applicable federal and state laws and regulations, including the HIP privacy rule. Patients query to the radiology database at a tertiary children s hospital from 2004 to 2006 identified 746 potential cases with the keyword osteomyelitis on the radiology request, the urrent Procedural Terminology code for extremity MRI, and an age range of 0 18 years. Research assistants reviewed the radiology requests and included only studies ordered for suspicion of osteomyelitis. If a patient had undergone multiple MRI studies of extremities for the investigation of osteomyelitis, only the initial study was included. From this population, 111 children imaged at the children s network for clinically suspected nonspinal osteomyelitis were selected randomly to be included in the study. Thirty-three patients were excluded from the study for the following reasons: lack of clinical follow-up (18 patients), skeletal maturity (seven patients), substantial coexistent bone marrow disorder such as leukemia (four patients), previous surgery at the imaged body part (two patients), and poor image quality (two patients). The other 78 patients (32 girls, 46 boys; median age, 3.6 years; range, 1 month 18 years) were included in the final analysis. MRI MRI studies of 78 children and adolescents with clinically suspected nonspinal osteomyelitis were evaluated retrospectively. Images were reviewed for the presence of osteomyelitis and its complications, including abscess, physeal involvement, and septic arthritis. Seventy studies were acquired with closed-bore 1.5-T MRI systems, and the other eight were acquired with 3-T systems. Imaging parameters, including coil selection, plane, and field of view, were tailored for each body part studied. Images were obtained in at least two orthogonal planes. ll studies includ ed the following pulse sequences: spin-echo T1-weighted images (TR range/te range, /10 20), fat-suppressed turbo spinecho T2-weighted images (3,000 6,000/50 130), STIR images (4,000 6,000/30; inversion time, 150 ms), and fat-sup pressed spin-echo T1-weighted images ( /10 20) after IV admin istration of gadopentetate dimeglumine (Magne vist, ayer Healthare) at a dose of 0.1 mmol/kg body weight. Fat suppression was achieved with selective presaturation of lipid resonances. Image nalysis Two radiologists retrospectively reviewed the studies in consensus. One observer had 20 years experience in musculoskeletal imaging, and the other was a pediatric radiology fellow. The readers were blinded to all clinical data except patient age and clinical concern about osteomyelitis. To most closely simulate clinical practice, unenhanced MR images were evaluated first and then contrastenhanced MR images. 5-point scale ranging from definitely absent to definitely present was used to evaluate all studies for the presence or absence of osteomyelitis, abscess (intraosseous, subperiosteal, or soft tissue), septic arthritis, and physeal involvement. Interobserver reliability was analyzed with 42 additional studies chosen randomly from the pool of 746 children identified in our database search. These studies were of children other than the 78 patients included in our study population. The two readers used the 5-point scale to independently review these cases for the diagnosis of osteomyelitis and abscess. MRI criteria for the diagnosis of osteomyelitis were based on those described in the literature [1, 2, 4, 5, 12, 15]. On unenhanced images, osteomyelitis was characterized by focally decreased marrow signal intensity on T1-weighted images and focally increased marrow signal intensity on fluid-sensitive images (fat-suppressed T2- weighted and STIR sequences). fter contrast administration, osteomyelitis was described as focal abnormal bone marrow enhancement on fatsuppressed T1-weighted images. MRI criteria for complications of osteomyelitis also were based on previously published descriptions [4, 14, 16, 17]. Intraosseous, subperiosteal, and soft-tissue abscesses were defined as wellcircumscribed areas of focally decreased signal intensity on T1-weighted images with increased signal intensity equal to that of fluid on fluidsensitive sequences and rim enhancement on contrast-enhanced fat-suppressed T1-weighted images. Septic arthritis was described as joint effusion with synovial thickening that became enhanced on contrast-enhanced fat-suppressed T1-weighted images. Physeal involvement was char acterized as increased signal intensity or widening of the growth plate on fluid-sensitive images and increased or rim enhancement on contrast-enhanced images. bscess and septic arthritis were evaluated independently from the presence or absence of osteomyelitis. Therefore, cases could be scored positive for abscess or septic arthritis although there was no evidence of bone marrow abnormality. Physeal involvement, however, was considered inherently linked to the presence of osteomyelitis. linical orrelation Patient charts were reviewed for clinical data. Presence or absence of osteomyelitis was determined by the final clinical impression recorded at presentation. Data available to the clinicians varied from case to case but included results of the following: history and physical examination, laboratory tests, bone biopsy, bone marrow aspiration, blood cultures, radiologic studies, and response to therapy. No correlation was available for the secondary diagnoses of abscess, septic arthritis, and physeal involvement. Statistical nalysis Sensitivity and specificity were calculated for the diagnosis of osteomyelitis, and p was obtained with the chi-square method. The z score was used to compare the 5-point-scale data collected for unenhanced MR images and contrast-en hanced MR images, including the diagnosis of osteomyelitis and its complications. The kappa value for interobserver reliability was determined for the 42 studies reviewed independently by the two readers. kappa value greater than 0.75 was defined as excellent agreement, of as good agreement, and less than 0.40 as poor agreement. Results There was good-to-excellent interobserver reliability between the two readers for the diagnoses of osteomyelitis (κ = 0.72; 95% I, ) and abscess (κ = 0.71, 95% I ). Forty of the 78 children (51%) had a final clinical diagnosis of osteomyelitis. Positive blood culture, bone marrow aspiration, or bone biopsy results were documented in 12 of the 40 cases (30%). Scores of definitely absent to indeterminate on MR images were considered no osteomyelitis. Scores of probably present and definitely present were considered presence of osteomyelitis. ccording to this categorization, unenhanced MRI had a sensitivity of 80% and a specificity of 84% (p = 1.0). Similarly, fat-suppressed contrast-enhanced MRI had a sensitivity of 83% and a specificity of 79% (p = 0.77) (Tables 1 and 2, Fig. 1). JR:192, May

3 verill et al. TLE 1: omparison of linical Findings of Osteomyelitis with Findings at Unenhanced MRI (n = 78) linical Finding Finding at Unenhanced MRI Present bsent Present 32 8 bsent 6 32 TLE 2: omparison of linical Findings of Osteomyelitis with Findings at ontrast- Enhanced MRI (n = 78) linical Finding Finding at ontrast- Enhanced MRI Present bsent Present 33 7 bsent 8 30 oth false-positive and false-negative findings were encountered. For example, a patient with juvenile rheumatoid arthritis was deemed to have MRI findings of definitely present osteomyelitis, but when all of the clinical data were taken into account, bone marrow inflammation was attributed to the underlying disorder rather than to infection. onversely, a case scored definitely absent on the basis of MRI features of a fracture was diagnosed clinically as osteomyelitis (Fig. 2). To assess the degree of reader confidence in the imaging diagnosis of osteomyelitis, any change in scoring between unenhanced and contrast-enhanced MRI was analyzed. For example, a change in score from probably present to definitely present was considered an increase in confidence. The addition of contrast-enhanced images increased reader confidence toward the correct clinical diagnosis of osteomyelitis in 11 patients (14%), a statistically significant finding (p < 0.001). In four cases, the addition of gadolinium enhancement decreased reader confidence in the presence or absence of osteomyelitis. On further review of these images and the associated clinical data, the readers disagreed with the radiologic interpretation rendered at presentation in three of these cases. Reader confidence in the imaging diagnosis of complications of osteomyelitis was investigated independently from the diagnosis of osteomyelitis itself. The addition of contrast-enhanced images increased reader confidence in the diagnosis or exclusion of abscess in 36 cases (46%), the diagnosis or exclusion of physeal involvement in 13 cases (17%), and in the diagnosis or exclusion of septic arthritis in 13 cases (17%) (Figs. 3 and 4). ll of these changes in confidence were statistically significant (p < 0.001). When the imaging diagnoses of osteomyelitis and its complications were analyzed together, the use of contrast-enhanced MRI increased reader confidence in 51 cases (65%) (p < 0.001). onversely, in the subset of studies with a score of definitely absent on unenhanced MR images, there was no significant change in reader confidence in the diagnosis of osteomyelitis, physeal involvement, or septic arthritis with the addition of contrast-enhanced images. In this subset of definitely absent scores on unenhanced images, there was no change in reader confidence in the diagnosis of osteomyelitis after administration of gadolinium in 23 of 25 cases (92%) (p = 0.47). For the detection of physeal involvement, there was no change in reader confidence in 41 of 44 cases (93%) (p = 0.24). For the diagnosis of septic arthritis, there was no change in any of the 44 patients (100%) (p = 1.0). In the detection of abscess, however, there was a significant change in confidence with Fig month-old girl with 3-week history of knee pain due to osteomyelitis of femur., Sagittal T1-weighted MR image shows decreased bone marrow signal intensity in distal femoral metaphysis., Sagittal STIR MR image shows bone marrow edema in distal femoral metaphysis. ecause of presence of fever and pain, findings on unenhanced imaging were deemed positive for osteomyelitis., Sagittal fat-suppressed contrast-enhanced T1-weighted MR image shows bone marrow enhancement in distal femoral metaphysis. Gadolinium enhancement confirms findings on unenhanced images but does not increase sensitivity for diagnosis of osteomyelitis JR:192, May 2009

4 MRI of Osteomyelitis the addition of gadolinium enhancement in cases initially scored definitely absent. change in reader confidence with the addition of contrast-enhanced sequences occurred in eight of 37 cases (22%) (p = 0.009). In retrospect, these patients had soft-tissue edema but no defined fluid collection on STIR or fat-suppressed T2-weighted images. With the addition of contrast enhancement, a small rim-enhancing collection might have been present in the soft tissues. Fig month-old boy with fever and refusal to bear weight 5 days after fall. Radiologist at presentation diagnosed osteomyelitis because of presence of striking bone marrow edema and enhancement, and patient was treated with antibiotics. In retrospect, findings on T1-weighted images are more suggestive of fracture. In study, this case was considered false-negative., Long-axis T1-weighted MR image shows linear low signal intensity (arrow) in cuboid bone in typical location and orientation for impaction fracture., Long-axis STIR MR image shows extensive bone marrow edema in cuboid bone., Long-axis fat-suppressed contrast-enhanced T1-weighted MR image shows bone marrow enhancement. Discussion Our study showed that even though it does not improve the sensitivity and specificity of the diagnosis of osteomyelitis in children, gadolinium enhancement during MRI does substantially increase the degree of confidence in the diagnosis. Gadolinium enhancement also significantly increases reader confidence in the diagnosis of complications of osteomyelitis, which include bone and softtissue abscess, physeal involvement, and septic arthritis. This effect is most pronounced in the evaluation of abscess. In our study, the addition of contrast-enhanced MR images increased reader confidence in the diagnosis of abscess in 46% of cases. ecause our interobserver agreement was good to excellent, these findings hold true for readers with extensive musculoskeletal MRI experience and readers with limited experience. The use of fat-suppressed gadolinium-enhanced MRI in the imaging of osteomyelitis in adults has been evaluated extensively [8]. Imaging of osteomyelitis in children, however, poses unique challenges. The marrow of the metaphysis, where hematogenous osteomyelitis usually originates, is highly cellular and richly vascularized [18]. For this reason, the normal metaphysis is of higher signal intensity on water-sensitive and contrastenhanced images; therefore, early marrow changes may be more subtle. The immature skeleton also contains important cartilaginous structures, such as the physis and the cartilaginous epiphysis, both of which can be involved in infectious processes. Finally, subperiosteal abscesses often occur in children because the periosteum is loosely attached to the bone. It cannot be assumed that the results of studies performed on adults can be generalized to children. The contribution of contrast-enhanced MRI to the diagnosis of osteomyelitis in children must be evaluated systematically. Gadolinium administration is expensive and not free of risk. Paradoxically, although fat-suppressed gadolinium-enhanced imaging is a well-established technique and widely used, the advantages over fluid-sensitive unenhanced imaging have been only minimally evaluated. Gadolinium enhancement helps in the detection of osteomyelitis by depicting areas of increased vascularity in the bone marrow and adjacent soft tissues. This information is JR:192, May

5 verill et al. complementary to that on fluid-sensitive images, which depict the edema associated with an inflammatory process. In cases of normal findings, neither bone marrow inflammation nor increased vascularity is present. It is not surprising, then, that our data show contrast enhancement does not add useful diagnostic information in the detection of osteomyelitis when the findings during the initial, unenhanced portion of the study are considered definitely normal. In cases in which edema is found on fluidsensitive images, it is important to find foci of decreased vascularity because the distribution of antibiotics is likely to mirror that of the injected gadolinium. Thus if necrosis is present in an area of infection, the antibiotics likely will not reach it, and the infection may not resolve without drainage. Our data support this assertion in that gadolinium enhancement had the greatest effect on reader confidence in evaluation for the presence or absence of abscess. For example, on a small number of unenhanced images, the readers were overly eager to score definite absence Fig month-old boy with refusal to bear weight for 3 weeks due to osteomyelitis of tibia with intraosseous abscess., Sagittal T1-weighted MR image shows low bone marrow signal intensity in distal tibia., Sagittal STIR MR image shows edema in bone marrow and surrounding soft tissues. Suggestion of small intraosseous abscess in metaphysis is suggested by focal high signal intensity equal to that of water (scored abscess probably present)., Sagittal fat-suppressed contrast-enhanced T1-weighted MR image shows small abscess in distal tibial metaphysis as area of low signal intensity with rim enhancement. Gadolinium enhancement increased reader confidence in diagnosis of abscess from probably present to definitely present. of abscess despite the presence of soft-tissue inflammation on the images. Gadolinium enhancement, however, revealed small areas of rim enhancement within the soft-tissue edema, suggesting the presence of an abscess. lthough more than one half of cases of acute hematogenous osteomyelitis occur in children younger than 5 years, it is difficult to study the performance of imaging of this age group. It has been historically difficult to use a single reference standard for the diagnosis of osteomyelitis because results of blood and bone cultures are positive in fewer than one half of children with infections [19]. ulture results were positive in only 30% of cases in our study population. In the cases of the other patients, the clinical diagnosis was determined with an array of clinical data, including the MRI findings. This phenomenon creates inherent reference standard bias in any retrospective study of osteomyelitis and challenges the validity of sensitivity and specificity statistics. We found several cases in which the readers disagreed with the radiologic interpretation at presentation. In three of these cases, the addition of gadolinium enhancement changed the readers confidence and led them toward an incorrect diagnosis based on the reference standard used. In other cases, there was no change in confidence but the findings still were tallied as false-positive or false-negative. For example, we encountered one case in which the initial MRI interpretation of osteomyelitis led to a clinical diagnosis of osteomyelitis and longterm antibiotic therapy. In retrospect, however, the imaging features were more characteristic of a fracture. radiology report favoring fracture over infection might have changed the final clinical diagnosis. In a case such as this, clinical follow-up is of little use; a fracture heals in a time course similar to the resolution of osteomyelitis appropriately managed with IV antibiotics. In addition to the considerable reference standard bias that occurred because MRI interpretation informed the final clinical diagnosis, there were other limitations to this retrospective study. The inclusion of only patients evaluated at a single tertiary care children s 1236 JR:192, May 2009

6 MRI of Osteomyelitis hospital in a major metropolitan location led to referral bias. In addition, patients were excluded from the study if they had a preexisting condition that might have affected the bone marrow, such as leukemia and previous surgery, or if they were lost to follow-up. Technical limitations also were inherent in this study. Imaging parameters were determined by departmental protocols and technologist expertise and therefore varied from case to case. Likewise, the delay between gadolinium administration and acquisition of fat-suppressed T1-weighted images fluctuated. This lack of uniformity might have altered the conspicuity of bone marrow inflammation and areas of necrosis. We conclude that contrast-enhanced MR images are a useful adjunct in the diagnosis of osteomyelitis, increasing confidence in the diagnosis of bone infection and its complications. Gadolinium enhancement is most useful in determining the presence or absence of abscess in a background of bone marrow or soft-tissue inflammation. It is least useful in cases in which no evidence of inflammation is present on unenhanced MR images. pplying these data to our clinical practice, Fig. 4 2-month-old boy, born at 28 weeks gestational age with unossified epiphyses and predominantly hematopoietic marrow, with pain and swelling of left knee due to septic arthritis and cartilage infection., Sagittal T1-weighted MR image of knee shows anterior soft-tissue swelling and indistinct tissue planes. Signal intensity differs little between cartilaginous structures, marrow, and inflamed adjacent soft tissues., Sagittal STIR MR image shows edema (arrows) in both intraarticular and extraarticular soft tissues of knee. Unenhanced imaging findings were considered septic arthritis probably present but abscess probably absent., Sagittal fat-suppressed contrast-enhanced T1-weighted MR image shows area of low signal intensity (arrows) with surrounding enhancement in cartilaginous femoral epiphysis and suprapatellar bursa. With gadolinium enhancement, septic arthritis and cartilage abscess are more easily defined and were scored definitely present. we will continue to administer gadolinium to children whose unenhanced images show evidence of edema in the bone or soft tissues. If the T1-weighted and fluid-sensitive images appear normal, however, we will terminate the study without administering contrast material, confident that further MRI will not add diagnostically useful information. Given the increasing safety concerns about gadolinium contrast agents, judicious use of contrast enhancement is in the best interest of our patients. lthough the algorithm requires physician supervision during image acquisition, we already monitor each patient for plane selection and body part coverage. This additional scrutiny should not add much to physician time and has the potential benefit of a shorter examination, decreasing sedation time and increasing patient safety, satisfaction, and throughput. References 1. Dangman, Hoffer F, Rand FF, O Rourke EJ. Osteomyelitis in children: gadolinium-enhanced MR imaging. Radiology 1992; 182: Hopkins KL, Li K, ergman G. Gadolinium- DTP-enhanced magnetic resonance imaging of musculoskeletal infectious processes. Skeletal Radiol 1995; 24: Mazur JM, Ross G, ummings J, Hahn G Jr, Mcluskey WP. Usefulness of magnetic resonance imaging for the diagnosis of acute musculoskeletal infections in children. J Pediatr Orthop 1995; 15: Morrison W, Schweitzer ME, atte WG, Radack DP, Russel KM. Osteomyelitis of the foot: relative importance of primary and secondary MR imaging signs. Radiology 1998; 207: Morrison W, Schweitzer ME, ock GW, et al. Diagnosis of osteomyelitis: utility of fat-suppressed contrast-enhanced MR imaging. Radiology 1993; 189: ancroft LW. MR imaging of infectious processes of the knee. Radiol lin North m 2007; 45: Erdman W, Tamburro F, Jayson HT, Weatherall PT, Ferry K, Peshock RM. Osteomyelitis: characteristics and pitfalls of diagnosis with MR imaging. Radiology 1991; 180: Kapoor, Page S, Lavalley M, Gale DR, Felson DT. Magnetic resonance imaging for diagnosing foot osteomyelitis: a meta-analysis. rch Intern Med 2007; 167: Oudjhane K, zouz EM. Imaging of osteomyelitis JR:192, May

7 verill et al. in children. Radiol lin North m 2001; 39: Pineda, Vargas, Rodriguez V. Imaging of osteomyelitis: current concepts. Infect Dis lin North m 2006; 20: Tang JS, Gold RH, assett LW, Seeger LL. Musculoskeletal infection of the extremities: evaluation with MR imaging. Radiology 1988; 166: Jaramillo D, Treves ST, Kasser JR, Harper M, Sundel R, Laor T. Osteomyelitis and septic arthritis in children: appropriate use of imaging to guide treatment. JR 1995; 165: eltran J, Noto M, McGhee R, Freedy RM, FOR YOUR INFORMTION Mcalla MS. Infections of the musculoskeletal system: high-field-strength MR imaging. Radiology 1987; 164: Miller TT, Randolph D Jr, Staron R, Feldman F, ushin S. Fat-suppressed MRI of musculoskeletal infection: fast T2-weighted techniques versus gadolinium-enhanced T1-weighted images. Skeletal Radiol 1997; 26: Umans H, Haramati N, Flusser G. The diagnostic role of gadolinium enhanced MRI in distinguishing between acute medullary bone infarct and osteomyelitis. Magn Reson Imaging 2000; 18: onnolly S, onnolly LP, Drubach L, Zurakowski D, Jaramillo D. MRI for detection of abscess in acute osteomyelitis of the pelvis in children. JR 2007; 189: Karchevsky M, Schweitzer ME, Morrison W, Parellada J. MRI findings of septic arthritis and associated osteomyelitis in adults. JR 2004; 182: Dwek JR, Shapiro F, Laor T, arnewolt E, Jaramillo D. Normal gadolinium-enhanced MR images of the developing appendicular skeleton. Part 2. Epiphyseal and metaphyseal marrow. JR 1997; 169: Vazquez M. Osteomyelitis in children. urr Opin Pediatr 2002; 14: The comprehensive book based on the RRS 2009 annual meeting categorical course on Ultrasound: Practical Sonography for the Radiologist is now available! For more information or to purchase a copy, see JR:192, May 2009