The association of postnatal urinary tract dilation risk score with clinical outcomes.

BACKGROUND: There are limited data on the predictive value of the consensus urinary tract dilation (UTD) score with respect to subsequent clinical diagnoses. We sought to define the relationship between postnatal UTD risk score and clinical outcomes during childhood. METHODS: Complete ultrasound image sets from a random selection of infants aged 0-90 days undergoing initial ultrasound at a single institution for prenatal hydronephrosis between 2012 and 2014 were assigned a UTD score by 1 pediatric urologist and 1 pediatric radiologist. Urinary tract dilation risk score was analyzed for association with a composite outcome comprising urinary tract infection, vesicoureteral reflux (VUR), ureteropelvic junction obstruction, non-refluxing megaureter (NRM), ureterocele, bladder outlet obstruction (BOO), and chronic kidney disease. Surgical intervention and resolution of UTD were evaluated separately. Descriptive and survival analyses were performed. RESULTS: Urinary tract dilation scores for 494 subjects were P0 in 23.5%, P1 in 26.5%, P2 in 23.5%, and P3 in 26.5%. Seventy-four percent were male. Median age at initial imaging was 28 days; median follow-up was 19.8 months. The composite outcome occurred in 138 of 494 patients (27.9%) and varied significantly (p < 0.001) by UTD score: 11.2% for P0, 10.7% for P1, 29.3% for P2, and 58.8% for P3. On survival analysis (Summary Figure), higher UTD grade was significantly associated with the composite outcome (hazard ratio for P3 vs. P0 was 7.4 [95% CI: 3.44-15.92, p < 0.001]). Urinary tract infection and VUR diagnosis varied by UTD score (p = 0.03 and p < 0.001, respectively). Ureteropelvic junction obstruction was diagnosed (based on MAG3 results) in 6.3% of patients, 84% of whom were P3. Non-refluxing megaureter was diagnosed in 7.7%. Ureterocele and BOO were uncommon (1.4%, and 0.6%, respectively). Surgical intervention was also associated with UTD risk, with 46% of P3 undergoing surgery vs. 1% of P0, 1% of P1, and 6% of P2 (p < 0.001). Resolution of UTD occurred in 41% (median 10.1 months) and varied significantly by UTD risk (p < 0.001). DISCUSSION: Urinary tract dilation risk score is associated with clinical events, although ascertainment bias may influence some of the differences in outcomes, particularly for VUR, because VCUG utilization varied by the UTD group. The lack of any significant difference in outcomes between patients with UTD P0 versus P1 suggests that the P1 category could be eliminated as it does not meaningfully distinguish between outcome risk. CONCLUSIONS: Higher UTD risk scores are strongly associated with genitourinary diagnoses during the first two years of life.

Frequency and cause of disagreements in imaging diagnosis in children with ventriculomegaly diagnosed prenatally.

OBJECTIVES: To assess the frequency and cause of variability in diagnosis on cranial sonography and magnetic resonance imaging (MRI) in children referred following prenatal diagnosis of ventriculomegaly. METHODS: Between 19 September 2003 and 16 March 2007, 119 infants with ultrasound and/or MRI studies performed within 13 months (median, 6 days) after birth, following prenatal referral for ventriculomegaly, were studied prospectively. There were 97 infants with ultrasound results and 53 with MRI, including 31 with both. Three sonologists and three pediatric neuroradiologists interpreted the postnatal ultrasound and MRI findings, blinded to prenatal diagnosis, and a final consensus diagnosis or group of diagnoses was obtained. Ventricular sizes as well as types of and reasons for any disagreement in diagnosis were recorded. Disagreements on a per patient basis were categorized as being major when they crossed diagnostic categories and had the potential to change patient counseling. Postnatal and prenatal diagnoses were compared. RESULTS: There was prospective agreement on 42/97 (43%) ultrasound and on 9/53 (17%) MRI readings. Prospective consensus was more likely when the number of central nervous system (CNS) anomalies was lower (P < 0.001 and P = 0.002 for ultrasound and MRI, respectively). In 24/55 (44%) ultrasound and 11/44 (25%) MRI examinations with disagreement in diagnosis, there was disagreement concerning the presence of ventriculomegaly. In 22/97 (23%) ultrasound studies and 22/53 (42%) MRI studies the disagreements were potentially important. Reasons for discrepancies in the reporting of major findings included errors of observation as well as modality differences in depiction of abnormalities. In comparing prenatal with postnatal diagnoses, there were 11/97 (11%) ultrasound and 27/53 (51%) MRI examinations with newly detected major findings, the most common being migrational abnormalities, callosal dysgenesis/destruction and interval development of hemorrhage. CONCLUSION: Variability in postnatal CNS diagnosis is common after a prenatal diagnosis of ventriculomegaly. This is due in part to a lack of standardization in the definition of postnatal ventriculomegaly.

Neurodevelopmental outcome of fetuses referred for ventriculomegaly.

OBJECTIVE: To characterize the delivery and postnatal neurodevelopmental outcomes of fetuses referred for ventriculomegaly (VM). METHODS: Under an internal review board-approved protocol, pregnant women were referred for magnetic resonance imaging (MRI) after sonographic diagnosis of VM and classified into one of four diagnostic groups: Group 1, normal central nervous system (CNS); Group 2, isolated mild VM (10-12 mm); Group 3, isolated VM > 12 mm; and Group 4, other CNS findings. Pregnancy outcome was obtained. Follow-up visits were offered with assessment of neurodevelopmental, adaptive and neurological functioning at 6 months and 1 year and/or 2 years of age. Atrial diameter and VM group differences in developmental outcomes were evaluated using repeated measures logistic regression and Fishers exact test, respectively. RESULTS: Of 314 fetuses, 253 (81%) were liveborn and survived the neonatal period. Fetuses in Groups 4 and 3 were less likely to progress to live delivery and to survive the neonatal period (60% and 84%, respectively) than were those in Groups 2 or 1 (93% and 100%, respectively, P < 0.001). Of the 143 fetuses followed postnatally, between 41% and 61% had a Bayley Scales of Infant Development (BSID-II) psychomotor developmental index score in the delayed range (< 85) at the follow-up visits, whereas the BSID-II mental developmental index and Vineland Adaptive Behavior composite scores were generally in line with normative expectations. Among those that were liveborn, neither VM group nor prenatal atrial diameter was related to postnatal developmental outcome. CONCLUSIONS: Diagnostic category and degree of fetal VM based on ultrasound and MRI measurements are associated with the incidence of live births and thus abnormal outcome. Among those undergoing formal postnatal testing, VM grade is not associated with postnatal developmental outcome, but motor functioning is more delayed than is cognitive or adaptive functioning.

Cleft of the secondary palate without cleft lip diagnosed with three-dimensional ultrasound and magnetic resonance imaging in a fetus with Fryns' syndrome.

We present a case of Fryns' syndrome diagnosed prenatally using three-dimensional (3D) ultrasonography and magnetic resonance imaging (MRI). A cleft of the soft palate was diagnosed using 3D thick-slice ultrasonography. Other sonographic findings included a right diaphragmatic hernia, enlarged echogenic kidneys and severe polyhydramnios. The detection of the cleft palate was instrumental in suggesting the diagnosis of Fryns' syndrome in a fetus which also had a diaphragmatic hernia. These findings were also demonstrated with prenatal MRI. The technique of imaging the soft palate en face using a thick-slice technique is presented.

99m Technetium dimercapto-succinic acid renal scintigraphy abnormalities in infants with sterile high grade vesicoureteral reflux.

PURPOSE: Although vesicoureteral reflux associated with bacteriuria may cause renal scarring, sterile reflux is thought not to cause renal injury. We determined the incidence and associated characteristics of renal abnormalities using 99mtechnetium(Tc) dimercapto-succinic acid (DMSA) renal scintigraphy in infants with high grade vesicoureteral reflux but no history of urinary tract infection. MATERIALS AND METHODS: We retrospectively reviewed the results of 99mTc-DMSA renal scintigraphy and renal ultrasonography performed during the first 6 months of life in infants with vesicoureteral reflux detected during the postnatal evaluation of prenatal hydronephrosis or sibling reflux screening. Those with a history of urinary tract infection, or evidence of ureteropelvic junction or bladder outlet obstruction were excluded from study. RESULTS: Of the 28 male and 6 female infants who met study criteria vesicoureteral reflux was bilateral in 25 and unilateral in 9. Reflux grade was IV or V, II or III and I in 38, 18 and 3 of the 59 refluxing renal units, respectively. 99mTc-DMSA renal scintigraphy revealed parenchymal abnormalities in 24 refluxing renal units (41%) in 22 patients (65%), of whom 19 (86%) were male and 15 (68%) had bilateral reflux. We noted differential uptake less than 40% with and without cortical defects in 10 and 7 refluxing units, respectively, and cortical defects only in 7. Of the 24 refluxing units with abnormalities 21 were associated with grade IV or V and 3 with grade II or III reflux. Ultrasound showed evidence of renal injury in only 7 of the 17 patients (41%) in whom 99mTc-DMSA scintigraphy was abnormal. CONCLUSIONS: In our study the majority of infants with high grade reflux had decreased differential function and/or cortical defects. Parenchymal defects detected by 99mTc-DMSA renal scintigraphy were often not identified by renal ultrasound. Therefore, 99mTc-DMSA renal scintigraphy is especially useful for initially evaluating infants with high grade, sterile vesicoureteral reflux.

Depiction of renal perfusion defects with contrast-enhanced harmonic sonography in a porcine model.

OBJECTIVE: Our purpose was to test the feasibility and optimization of contrast-enhanced gray-scale harmonic sonography for the evaluation of focal renal perfusion defects in an animal model. SUBJECTS AND METHODS: We performed arteriography and contrast-enhanced harmonic sonography on six anesthetized piglets after embolization of each kidney with an autologous clot through the main renal artery. Harmonic images were obtained with continuous (30 Hz) and ECG-triggered acquisition. The two harmonic sonography strategies were compared with respect to the number and conspicuity of devascularized areas, and correlation was made with arteriographic findings. RESULTS: Contrast-enhanced harmonic sonography showed focal areas of absent or diminished vascularity that corresponded closely with perfusion defects seen on angiography. Enhancement ratios to perfused cortex were significantly higher than to devascularized cortex in both continuous (mean +/- SD, 469 +/- 5% versus 102 +/- 8%, p < .0005 by t test) and triggered (673 +/- 7% versus 198 +/- 7%, p < .0001) modes. Triggered acquisition increased the conspicuity of perfusion defects over that obtained with continuous imaging (p < .002 by t test). CONCLUSION: Contrast-enhanced harmonic sonography is an effective method of depicting focal renal perfusion defects. Triggered acquisition further improves lesion conspicuity.

Renal blood flow in pigs: changes depicted with contrast-enhanced harmonic US imaging during acute urinary obstruction.

PURPOSE: To evaluate contrast agent-enhanced harmonic ultrasonographic (US) imaging and Doppler hemodynamics during acute urinary obstruction. MATERIALS AND METHODS: In 12 piglets, the distal ureter was obstructed for 60 minutes, followed by intravenous injection of furosemide. In six piglets, ureteral pressure was further elevated to mean arterial pressure, and in six other piglets ureteral obstruction was released. Contrast-enhanced harmonic imaging was performed, and interlobar resistive index (RI) and renal blood flow were determined at baseline and during each experimental condition. A bolus injection curve was constructed by plotting mean pixel intensity versus time, and the area under this normalized curve was compared with renal blood flow values. RESULTS: Ureteral obstruction and high ureteral pressure reduced cortical renal blood flow to 88% and 66%, respectively, of baseline values. Administration of contrast agent resulted in marked homogeneous enhancement of the renal cortex. The area under the curve diminished during ureteral obstruction and correlated well with mean cortical blood flow. RI correlated well with renal perfusion pressure but poorly with changes in renal blood flow. CONCLUSION: Contrast-enhanced harmonic US imaging depicts changes in renal blood flow during acute obstruction. Interlobar RI is a good predictor of renal perfusion pressure but not of changes in renal blood flow.

Nonradiopaque penetrating foreign body: "a sticky situation".

Foreign bodies within soft tissues are common in children. They may cause a chronic inflammatory reaction that can result in abnormal findings on radiographs, including lytic or blastic osseous changes. These radiographic findings can mimic both benign and malignant processes. In cases where the history is uncertain and the foreign body is not recognized, magnetic resonance (MR) imaging can make a specific diagnosis and direct appropriate therapy.

Experimental fetal vesicoureteral reflux induces renal tubular and glomerular damage, and is associated with persistent bladder instability.

PURPOSE: We assessed renal function and urodynamic status in animals with experimental congenital vesicoureteral reflux. MATERIALS AND METHODS: Vesicoureteral reflux was surgically induced in male sheep fetuses at 95 days of gestation. After birth the animals were maintained on antibiotic prophylaxis. At ages 1 week and 6 months reflux was assessed by fluoroscopic voiding cystography. Cystometrography was performed with the animals awake. Serum creatinine, inulin clearance and the excretion of urinary N-acetyl-beta-D-glucosaminidase were measured at ages 1 week, 1 month and 6 months by surveillance urine cultures. Urinary concentrating capacity was assessed by desmopressin testing at ages 1 and 6 months. RESULTS: Nine animals (18 renal units) were born after the induction of reflux. There was no reflux in 2 renal units, while reflux was mild in 2, moderate in 5 and severe in 9. In the 6 animals available for followup at age 6 months only severe reflux persisted. Reflux resolution was associated with normalization of bladder urodynamics. Surveillance urine cultures were negative until age 6 months, when infection developed in 3 of the 6 lambs. In all animals serum creatinine was normal during followup. Glomerular filtration rate in the lambs with reflux was no different from normal at age 1 week but it was significantly less than normal independent of infection at age 6 months (2.7 versus 3.9 ml./kg. per minute, p = 0.002). As an indicator of renal tubular injury the ratio of N-acetyl-beta-D-glucosaminidase-to-creatinine remained significantly higher in animals with reflux than in normal animals from ages 1 week to 6 months (51.0 versus 10.2 IU/mg., p = 0.03). Maximal concentrating ability after desmopressin testing was already less than normal by age 1 month with a maximal increase of 98 versus 435 mOsm./l. in lambs with reflux versus normal lambs (p <0.0001). It was further impaired by age 6 months. Urodynamic evaluation of the animals with reflux revealed decreased bladder compliance at age 1 week with normal voiding pressure. In addition, in those with reflux there was a more pronounced immature voiding pattern with multiple phasic contractions due to sphincteric activity as well as a post-void bladder contraction. CONCLUSIONS: Our model of fetal vesicoureteral reflux induces alterations in renal function that are consistent with clinical observations and marked by altered tubular function but a relatively mild decrease in glomerular filtration. Bladder dynamics are altered, consistent with observations in human neonates with high grade reflux and bladder instability. Whether this represents cause or effect remains unclear. Our model permits focused study of the interaction of these factors in neonatal reflux and may allow the application of more specific therapies, particularly those directed toward mechanisms of renal and bladder dysfunction.

Renal findings on radiological followup of patients with Beckwith-Wiedemann syndrome.

PURPOSE: The Beckwith-Wiedemann syndrome is most commonly characterized by macroglossia and abdominal wall defect(s), and it carries a predisposition to embryonal tumors, including Wilms tumor. We report our experience with the character and incidence of renal disease in patients with the Beckwith-Wiedemann syndrome, and discuss the role of radiological followup. MATERIALS AND METHODS: We reviewed the medical records of all patients diagnosed with the Beckwith-Wiedemann syndrome who were treated at our institution between March 1979 and February 1998. Radiological followup consisted of renal ultrasound at approximately 3 to 6-month intervals with the addition of computerized tomography or magnetic resonance imaging (MRI) in patients with an indeterminate lesion(s) or nephrogenic rest(s). RESULTS: A total of 29 patients were identified. Of these cases renal ultrasound showed normal kidneys bilaterally in 19 (70%), simple cysts in 5 (19%), indeterminate lesion(s) in 2 (7%) and nephrocalcinosis in 1 (4%). Nephrogenic rests were followed with MRI in 1 patient, and 1 in whom a 2 cm. mass was revealed by followup MRI underwent partial nephrectomy and chemotherapy for stage I Wilms tumor. CONCLUSIONS: The 3.7% incidence of Wilms tumor in our patients with the Beckwith-Wiedemann syndrome is similar to that in previously published reports. Aggressive follow-up by a sensitive radiological technique is warranted in cases of the Beckwith-Wiedemann syndrome, and associated hemihypertrophy and/or nephromegaly with or without evidence of a Wilms tumor precursor. The detection of suspected malignant disease at an early stage may permit curative nephron sparing surgery.

Techniques, coils, pulse sequences, and contrast enhancement in pediatric musculoskeletal MR imaging.

The first half of this article emphasizes the basic techniques involved in performing MR imaging of the pediatric musculoskeletal system. These include patient preparation, sedation issues, immobilization, coil selection, and pulse sequences. The second half of this article provides a detailed discussion of the use of Gadolinium and its enhancement characteristics in the normal developing musculoskeletal system, and its pathologic states are given.

Excitotoxin-induced cerebral hyperemia in newborn piglets: regional cerebral blood flow mapping with contrast-enhanced power Doppler US.

PURPOSE: To create regional cerebral blood flow maps with contrast material-enhanced power Doppler ultrasound (US). MATERIALS AND METHODS: In six anesthetized newborn piglets, cerebral hyperemia was induced after intrastriatal injection of 5 mumol of N-methyl-D-aspartic acid. Coronal power Doppler US was performed with a microbubble-based contrast agent, and cerebral blood flow was determined before and at 15, 30, and 60 minutes after injection. Images were digitized and analyzed for changes in mean pixel intensity. A bolus injection curve was constructed by plotting mean pixel intensity versus time, and the area under this normalized curve was compared with cerebral blood flow. RESULTS: Hemispheric cerebral blood flow increased from 40 mL/min/100 g +/- 1 (mean +/- standard error) to 90 +/- 12, 86 +/- 8, and 85 +/- 12 at 15, 30, and 60 minutes after injection, respectively (P < .0001 [analysis of variance]). Hemispheric mean pixel intensity at peak contrast also increased from 69 units +/- 5 to 120 +/- 4, 112 +/- 6, and 98 +/- 13 at 15, 30, and 60 minutes after injection, respectively (P < .003). The area under the normalized curve correlated well with changes in hemispheric and striatal cerebral blood flow (r = .73, P = .0001; r = .62, P = .0001, respectively). CONCLUSION: In the newborn brain, regional blood flow maps can be created accurately with contrast-enhanced power Doppler US.

Ossifying renal tumor of infancy presenting as a palpable abdominal mass.

We report a case of ossifying renal tumor of infancy, which presented as a palpable abdominal mass in an otherwise asymptomatic 10-month-old girl. The tumor was partially calcified and occupied the renal pelvis, causing severe hydronephrosis. The differential diagnosis for a patient this age included Wilms tumor, extra-adrenal neuroblastoma, infection, calculus, calcified hematoma and ossifying renal tumor of infancy. The child underwent heminephrectomy and is currently doing well.

Neonatal pig brain: lack of heating during Doppler US.

PURPOSE: To evaluate potential local brain-heating effects of Doppler ultrasound (US) at high power output settings. MATERIALS AND METHODS: After craniotomy, a temperature probe was inserted into deep white matter or basal ganglia in 12 anesthetized newborn piglets. The range gate and focal zone were placed directly over the temperature sensor. Brain and core (rectal) temperatures were measured before and during 5 minutes of continuous transdural duplex or color Doppler US, with transmitter power output settings of less than 500 mW in situ spatial peak temporal average or less than 800 mW in situ spatial peak temporal average. Temperature also was measured during administration of a US contrast agent with duplex US at less than 500 mW in situ spatial peak temporal average and after death at the same output setting. RESULTS: Maximal differences between the baseline and the post-US temperatures were, on average, less than 0.3 degree C (P > .5) and were highly correlated with changes in core temperature (r = .9, P < .001). Addition of the contrast agent did not result in focal brain heating during pulsed Doppler US (P > .6). There was no association between an increase in estimated power output and brain temperature change (P = .9). In addition, brain ischemia (postmortem exposure) was not associated with focal brain heating during pulsed Doppler US. CONCLUSION: Transfontanellar Doppler US alone or with administration of a contrast agent does not produce statistically significant focal brain heating at high transmitter power levels.

Imaging techniques to assess the pelvis in young females.

This article is designed to help the nonradiologist to form a logical imaging plan when presented with a diagnostic dilemma involving the pelvis of a young female. Common presenting problems are grouped under the categories of fetus, neonate and infant, child, and adolescent. Ultrasound is frequently the initial, and sometimes the only, examination indicated. The appropriateness of other modalities, such as plain radiography, genitography, computed tomography, and magnetic resonance imaging, is described. Details of examination performance and interpretation are beyond the scope of this discussion, but can be gleaned from the references cited.

Renal cortical ischemia in rabbits revealed by contrast-enhanced power Doppler sonography.

OBJECTIVE: Our goals were to describe and quantify redistribution of renal cortical blood flow during systemic hypotension in rabbits using contrast-enhanced power Doppler sonography and to explore the feasibility of creating real-time regional blood flow maps of the kidney. MATERIALS AND METHODS: Mild, moderate, and severe systemic hypotension (70%, 50%, and 40% of baseline measurements, respectively) were induced by controlled exsanguination in nine anesthetized rabbits. Sagittal power Doppler sonograms of the kidney were obtained during and after i.v. injection of 0.2 ml per kilogram of body weight of an experimental microbubble-based contrast agent. Doppler sonograms were analyzed for anatomic distribution of contrast agent and changes in mean pixel intensity over time. Cortical time-intensity curves were drawn and the area under the curve was compared with regional renal blood flow as measured by radiolabeled microspheres using linear regression in eight of the nine rabbits during graded hypotension. RESULTS: Graded hypotension resulted in increasing thickness of the nonperfused peripheral cortex (p < .001 by analysis of variance) with preservation of deeper cortical flow on enhanced Doppler sonograms. Other changes included decreased peak enhancement ratios and area under the curve values (p < .0001) and increased time to peak enhancement for cortex (p < .01). We found a strong correlation between cortical area under the curve and regional renal blood flow (r = .86, p < .0001). These changes were not shown without contrast enhancement. CONCLUSION: Peripheral cortical perfusion is disproportionately affected during systemic hypotension. Contrast-enhanced power Doppler sonography shows potential for depicting changes in regional renal blood flow in real time and without radiation.

Radiologic-Pathologic Conference of Children's Hospital Boston: bilateral asymptomatic renal enlargement.

An otherwise healthy 11-month-old boy presented with bilateral abdominal masses. Imaging findings, differential diagnosis, histological findings, and pertinent discussion are presented.

Normal gadolinium-enhanced MR images of the developing appendicular skeleton: Part I. Cartilaginous epiphysis and physis.

OBJECTIVE: We have used gadolinium-enhanced MR imaging to define the expected normal appearance of the developing cartilaginous epiphyses and physes in neonates, infants, and children and to define the changes with maturity in epiphyseal vascular pattern. MATERIALS AND METHODS: We analyzed gadolinium-enhanced MR images of 80 normal epiphyses in 48 neonates, infants, and children who were 1 month to 15.5 years old. We studied the differences in enhancement ratios for the epiphyses and physes and the epiphyseal vascular pattern at various development stages. We correlated the MR imaging findings with histologic and injection studies of immature epiphyses. RESULTS: Gadolinium enhancement allowed differentiation between physeal and epiphyseal cartilage and revealed epiphyseal vascular canals. Enhancement proved to be greater in the physeal than in the epiphyseal cartilage (p < .001). In the unossified epiphysis, the vascular canals were mainly parallel. After the development of the secondary ossification center, these canals came to have a radial pattern (p < .0001). Comparison with cadaveric specimens confirmed how, with age, the arrangement of these canals changed. Also, physeal enhancement decreased with physeal closure. CONCLUSION: Gadolinium-enhanced MR imaging reveals differential enhancement of the physis, epiphyseal vascular canals, and epiphyseal cartilage. The pattern of epiphyseal vessels and degree of enhancement of the physis change with maturity.

Normal gadolinium-enhanced MR images of the developing appendicular skeleton: Part 2. Epiphyseal and metaphyseal marrow.

OBJECTIVE: We have studied how gadolinium enhancement of T1-weighted MR images affects the expected normal differences in signal intensity between metaphyseal hematopoietic and epiphyseal fatty marrow. We have also analyzed how enhancement affects the expected normal changes in the MR images of the marrow due to fatty conversion. MATERIALS AND METHODS: We analyzed gadolinium-enhanced MR images of normal distal femurs in 18 immature rabbits that were 5-11 weeks old and of normal proximal femurs in 18 infants, children, and young adults who were 2 months to 21 years old. In all subjects, we studied the change with age in signal intensity and enhancement ratio of the epiphyseal and metaphyseal marrow. In the rabbits, marrow composition and transformation were histologically verified. RESULTS: On unenhanced T1-weighted MR images of the rabbits and of the infants, children, and young adults, epiphyseal signal intensity always exceeded metaphyseal signal intensity; however, the enhancement ratio was always greater in the metaphysis. The signal intensity in metaphyseal and epiphyseal marrow on unenhanced MR images increased with age. However, enhancement ratios decreased with age in both areas. In the rabbits, histologic studies showed more fatty marrow in the epiphysis than in the corresponding metaphysis and an age-related increase in marrow fat at both sites. CONCLUSION: In the marrow of the extremities, gadolinium enhancement is greater in the (hematopoietic) metaphysis than in the (fatty) epiphysis. In both areas, enhancement decreases as the marrow becomes more fatty. On T1-weighted images, administration of a gadolinium-containing contrast agent reduces the normal contrast between hematopoietic and fatty marrow and obscures the changes in marrow signal intensity due to fatty conversion.