ORIGINAL ARTICLE

Pediatric renal leukemia: spectrum of CT imaging findings

Melissa A. Hilmes & Jonathan R. Dillman &

Rajen J. Mody & Peter J. Strouse

Received: 8 October 2007 /Revised: 4 December 2007 /Accepted: 18 December 2007 / Published online: 1 February 2008# Springer-Verlag 2008

AbstractBackground The kidneys are a site of extramedullaryleukemic disease that can be readily detected by CT.Objective To demonstrate the spectrum of CT findings inchildren with renal leukemic involvement.Materials and methods Twelve children were identifiedretrospectively as having renal leukemic involvement bycontrast-enhanced CT of the abdomen. Contrast-enhancedCT images through the kidneys of each patient werereviewed by two pediatric radiologists. Pertinent imagingfindings and renal lengths were documented. The electronicmedical record was accessed to obtain relevant clinical andpathologic information.Results Five patients with renal leukemic involvementpresented with multiple bilateral low-attenuation masses,

while three patients demonstrated large areas of wedge-shaped and geographic low attenuation. Four otherpatients presented with unique imaging findings, includ-ing a solitary unilateral low-attenuation mass, solitarybilateral low-attenuation masses, multiple bilateral low-attenuation masses including unilateral large conglomer-ate masses, and bilateral areas of ill-defined parenchymallow attenuation. Two patients showed unilateral nephro-megaly, while eight other patients showed bilateralnephromegaly. Two patients had normal size kidneys.Two patients had elevated serum creatinine concentrationsat the time of imaging.Conclusion Renal leukemic involvement in children canpresent with a variety of CT imaging findings. Focal renalabnormalities as well as nephromegaly are frequentlyobserved. Most commonly, renal leukemic involvementdoes not appear to impair renal function.

Keywords Leukemia . Kidneys . Children . CT

Introduction

Acute lymphoblastic leukemia (ALL), acute myelogenousleukemia (AML), and juvenile myelomonocytic leukemia(JMML) are forms of leukemia that commonly affectchildren. Unlike patients with lymphoma, children withleukemia do not generally require routine CT imaging forstaging or follow-up. Children with leukemia usuallyinstead are monitored with bone marrow aspiration, lumbarpuncture with cytology, and complete blood count withsmear and differential. When children with leukemia areimaged with CT, a variety of renal abnormalities mightsuggest the possibility of extramedullary leukemic involve-ment. Leukemic patients who are at a higher risk for

Pediatr Radiol (2008) 38:424–430DOI 10.1007/s00247-007-0741-5

M. A. Hilmes : J. R. Dillman : P. J. StrouseSection of Pediatric Radiology,University of Michigan Health System,C.S. Mott Children’s Hospital,Ann Arbor, MI, USA

R. J. ModyDivision of Pediatric Hematology-Oncology and Bone MarrowTransplantation, University of Michigan Health System,C.S. Mott Children’s Hospital,Ann Arbor, MI, USA

J. R. Dillman (*)Department of Radiology, University of Michigan Health System,1500 E. Medical Center Drive,Ann Arbor, MI 48109, USAe-mail: jonadill@med.umich.edu

Present address:M. A. HilmesSection of Pediatric Radiology,Vanderbilt University Children’s Hospital,Nashville, TN, USA

extramedullary disease, including renal parenchymal in-volvement, include those with T-cell ALL as well as thosewith the M4 and M5 subtypes of AML [1–3].

CT imaging of the abdomen in children with leukemia istypically utilized in the assessment of possible disease-related complications or in the evaluation of some otherclinical problem. Consequently, renal leukemic involve-ment might be a completely unexpected incidental imagingfinding. Based on a review of the literature, renalparenchymal leukemic deposits are sometimes associatedwith impaired renal function. There are a few case reportsof new-onset renal failure that were directly attributable torenal leukemic involvement [4–6].

The purpose of this study was to describe the CTimaging findings of renal leukemic involvement in childrenwith a variety of forms of leukemia, including ALL, AML,and JMML. In addition, we sought to correlate the presenceof renal leukemic involvement with renal function.

Materials and methods

Institutional review board (IRB) approval was obtainedprior to the initiation of this retrospective investigation.Using our Department of Radiology information system(RIS), all contrast-enhanced abdominal CT imaging reportswere identified for children with leukemia during a 10-yearperiod from 1 January 1996 to 31 December 2005. Imagingreports were then reviewed by a single author (P.J.S.) forpossible leukemic renal involvement. The review of theimaging reports identified 12 children with leukemic renalinvolvement demonstrated by CT imaging.

In 11 of the 12 children, the diagnosis of leukemia as theetiology for the child’s renal parenchymal abnormality waspresumed based upon a combination of CT imagingfindings (including follow-up studies after chemotherapyand bone marrow transplantation, BMT) and clinicalinformation documented in the medical record. The renallesions observed in three children had markedly decreasedin size/extent upon follow-up CT imaging after chemother-apy and/or BMT. All 12 children had additional sites ofconcomitant extramedullary disease. Care was taken toexclude children in whom infectious disease was a possiblecause of the renal parenchymal abnormality. Specifically,two children (in addition to the 12 described above) withleukemia and bilateral renal masses with signs andsymptoms of an infectious process (i.e. positive blood/urine cultures and fever) were excluded from this study. Asingle child underwent biopsy of a renal mass that provedthe diagnosis of renal leukemia. Biopsy was indicated asthis child had Li-Fraumeni syndrome and a history of bothrecurrent fibrosarcoma and relapsed precursor B-cell ALL(for which he had received BMT). Based on the complexity

of the clinical situation, biopsy of the renal mass waspursued in order to guide appropriate therapy.

Available contrast-enhanced CT scans of the abdomen ofeach child were reviewed retrospectively by two pediatricradiologists in consensus, and imaging findings pertainingto the kidneys and adjacent perinephric/paranephric spaceswere documented. Bilateral craniocaudad renal lengthswere measured on sagittal reformatted images and recordedfor all patients. The number of standard deviations above orbelow mean renal length for patient age was alsodetermined. Images were loaded on a picture archivingand communication system (PACS) workstation andreviewed utilizing both standard soft-tissue and narrowCT level/window settings.

Our institutional electronic medical record system wasaccessed for each child in order to elicit the specific form ofleukemia, the indication for CT imaging, laboratory valuespertaining to renal function, and additional extramedullaryextrarenal sites of disease. Serum creatinine concentrationanalysis was performed by our institutional Department ofPathology. Abnormal serum creatinine levels were institu-tionally defined as follows: greater than 0.8 mg/dl for achild younger than 5 years and greater than 0.9 mg/dl for achild older than 5 years.

The initial age at diagnosis of leukemia in our patientcohort ranged from 5 months to 18 years (mean 5.8 years).The age at presentation with renal leukemic involvementranged from 7 months to 21 years (mean 7.1 years). Tenchildren were boys and two were girls. During the studytime period, a total of 423 children with leukemia werediagnosed and/or treated at our institution (300 childrenwith ALL, 120 with AML, and 3 with JMML).

All children included in this investigation were diag-nosed according to World Health Organization (WHO)criteria [7, 8]. All ALL patients had ≥25% bone morrowinvolvement with lymphoblasts, while all AML patientshad >20% bone marrow involvement with myeloblasts.Children with <25% lymphoblasts were diagnosed as havinglymphoblastic lymphoma, while children with <20% myelo-blasts were diagnosed as having myelodysplastic syndrome(MDS). Children with lymphoma and MDS were excludedfrom this investigation. JMML is associated with <20%blasts (including promonocytes) in the blood and bonemarrow (with an average blast count of less than 2%).

Eight children (66%) included in this investigation hadsome form of ALL: four (33%) had precursor T-cell ALL,two (17%) had mature T-cell ALL, and two (17%) hadrelapsed precursor B-cell ALL. Three children (25%) hadAML, including one child (8%) with relapsed M4 subtypeand two (17%) with M5 subtype. A single child (8%) hadJMML. Prior BMT had been performed in five children(42%). Five (42%) of the 12 children had died by the timeof this image review.

Pediatr Radiol (2008) 38:424–430 425

A review of electronic medical records revealed that themost common indication for contrast-enhanced CT exam-ination of the abdomen in children included in this studywas to “evaluate for extramedullary leukemic involvement”(five children with known mediastinal masses and one childwith an orbital mass). Indications for examinations per-formed in other children included: “lactic acidosis, concernfor a leukemic relapse,” “acute abdominal pain,” “back andabdominal pain,” “fever, rule out infection,” “persistentemesis,” and “pre-bone marrow transplant evaluation.”

Nine of the 12 children’s CT scans evaluated in thisstudy were performed at our institution. Three scans wereperformed at other institutions, for which the exact imagingtechnique was unknown. Scans included in this study thatwere performed at our institution were obtained usingroutine pediatric CT abdomen and pelvis examinationtechniques. Scans were performed from the lung basesthrough the ischia using either a 2.5-mm or 5-mm sectionthickness with no section overlap. Imaging was performedapproximately 65 s following the initiation of intravenouscontrast material administration. The mAs was selected perthe standard departmental protocol based on patient weight,and ranged from 50 to >170 (but never exceeding thepatient’s weight in pounds if the patient weighed greaterthan 150 pounds). A kVp of 120 was used for all children.

Both oral (either diatrizoate sodium for hospital inpa-tients or dilute iohexol for outpatients) and intravenouscontrast materials were administered to all children includ-ed in this investigation who were imaged at our institution.Two different intravenous low-osmolality nonionic iodinat-ed contrast agents were used in pediatric patients during thestudy period: iohexol 300 mg I/ml (Omnipaque 300; GEHealthcare, Princeton, NJ) and iopromide 300 mg I/ml(Ultravist 300; Bayer HealthCare, Wayne, NJ). The volumeof intravenous contrast material administered was based onthe standard departmental protocol related to patient weight.In general, we administer an intravenous contrast materialdose of 2 ml per kilogram in children weighing 15 kg orgreater (up to 75 kg). Slightly more contrast material perkilogram is used in children weighing less than 15 kg, and amaximum dose of 150 ml of contrast material is adminis-tered to patients weighing 75 kg or greater. Intravenousinjections of contrast medium greater than 20 ml areadministered utilizing a power injector unless specificallycontraindicated, while intravenous injections less than orequal to 20 ml are typically hand-injected.

Results

Renal leukemic involvement presented with a wide varietyof contrast-enhanced CT imaging findings. The mostcommon focal parenchymal abnormality was that of

multiple bilateral renal low-attenuation masses (5 of 12patients, 42%; Figs. 1, 2, 3, and 4). Three children (25%)presented with large areas of bilateral wedge-shaped andgeographic low attenuation (Fig. 5). A single child (8%)presented with multiple large conglomerate low-attenuationmasses in a single kidney with additional small discretelow-attenuation masses within the contralateral kidney.Other individual children presented with a unilateralsolitary low-attenuation mass (8%), bilateral solitary low-attenuation masses (8%; Fig. 6), and bilateral ill-definedareas of abnormal low attenuation (8%).

Abnormally increased renal length (nephromegaly, or arenal length more than two standard deviations greater thanthe expected size for patient age) was present in themajority of children with renal leukemic involvement(Table 1). Bilateral nephromegaly was observed in eightchildren (66%), while unilateral nephromegaly was ob-served in two children (16.5%). Only two children (16.5%)had normal bilateral renal length.

All 12 children with renal involvement, whether atprimary leukemia presentation or relapse, had additionalevidence of extramedullary leukemic involvement. Theseadditional areas of disease were identified by imaging andphysical examination. Areas of extramedullary leukemicinvolvement, other than the kidneys, included the medias-tinum, pericardium, peritoneum, spleen, liver, pancreas,lymph nodes, spinal canal, brain, skin, orbit, testes, andgingivae.

Two children had abnormally elevated serum creatinineconcentrations at the time of CT imaging. A single childwho was found to have an elevated creatinine wasdiscovered to have partial obstruction of the right renal

Fig. 1 T-cell ALL and a mediastinal mass in a 9-year-old boy.Contrast-enhanced CT examination was performed to evaluate foradditional sites of extramedullary disease. Axial image showsscattered small bilateral low-attenuation renal masses (arrows) andan enlarged spleen (S)

426 Pediatr Radiol (2008) 38:424–430

collecting system and globally decreased ipsilateral renalperfusion (Fig. 7). This patient presented with back andabdominal pain as well as evidence of impaired renalfunction (creatinine 1.6 mg/dl, baseline creatinine 0.8 mg/dl).Another child who was undergoing chemotherapy at thetime of imaging had a minimally elevated serum creatininelevel for age (creatinine 0.9 mg/dl, baseline creatinine0.6 mg/dl). This child had multiple bilateral renal massesthat completely regressed with chemotherapy and BMT.Follow-up creatinine upon resolution of the renal masseswas normal (Fig. 8). Eight children had normal serumcreatinine levels at the time of the CT scan thatdemonstrated renal leukemic involvement. A serum

creatinine level that coincided with the date of imagingwas not available for two children.

Discussion

Multiple reports suggest that nephromegaly is the mostcommon imaging manifestation observed in the setting ofrenal leukemic involvement. This finding has been reportedfollowing the use of sonography, excretory urography, andCT [9–13]. Ten of 12 patients in our study had eitherunilateral or bilateral renal enlargement, confirming that

Fig. 2 A 3-year-old boy with T-cell ALL and a mediastinal mass.Axial contrast-enhanced CT image demonstrates numerous bilaterallow-attenuation renal masses (arrows). Abnormally enlarged retroper-itoneal lymph nodes are also seen (arrowheads)

Fig. 3 A 9-month-old boy with AML and a biopsy-proven leukemicorbital mass. Axial contrast-enhanced CT image shows numerousbilateral low-attenuation renal masses. The masses decreased in sizefollowing chemotherapy

Fig. 4 A 5-month-old girl with emesis found to have AML. Axialcontrast-enhanced CT image shows numerous bilateral low-attenuationrenal masses as well as frank bilateral nephromegaly

Fig. 5 A 12-year-old boy with relapsed T-cell ALL. The boypresented with lactic acidosis, similar to his initial presentation withleukemia. Axial contrast-enhanced CT image shows bilateral renalgeographic areas of low attenuation (arrows). There is also frankbilateral (left greater than right) nephromegaly. A conglomerateretroperitoneal lymph node mass is also present (arrowheads)

Pediatr Radiol (2008) 38:424–430 427

nephromegaly is a common imaging finding in the settingof renal leukemic involvement in children. It should benoted, however, that focal renal parenchymal abnormalitieswere even more common than nephromegaly. Such focalabnormalities were most commonly bilateral and multifo-cal, although unilateral and solitary abnormalities werenoted in a few children. Such focal renal parenchymallesions were noted to take on a variety of imaging

appearances, including small and large round low-attenuationmasses, wedge-shaped and geographic low-attenuationmasses, and ill-defined areas of low attenuation. Noperinephric leukemic involvement was identified in ourpatients.

The presence of renal leukemic involvement itself doesnot appear to commonly cause significant renal dysfunc-tion. Only two patients had abnormally elevated serumcreatinine concentrations at the time that CT imagingrevealed renal leukemic involvement. One patient’s abnor-mally elevated creatinine likely was not the result of renalparenchymal leukemic involvement but was instead causedby retroperitoneal lymphadenopathy. The abnormal lymphnodes in this patient caused upper urinary tract obstructionas well as compression of the main renal artery and vein,near the renal hilum, resulting in ipsilateral renal malperfu-sion (Fig. 7).

Another child with minimally elevated creatinine at thetime of imaging had normal renal size and multiple bilateralrenal low-attenuation masses. The elevated creatinineobtained in this child might have been the result of recentlyinitiated chemotherapy and tumor lysis syndrome [14, 15].This condition can occur after the initiation of chemother-apy, frequently is associated with T-cell ALL, and can resultin acute renal failure secondary to urate nephropathy.Chemotherapy induces rapid tumor cell necrosis or apo-ptosis that can induce numerous metabolic abnormalities,including hyperkalemia, hyperuricemia, hyperphosphate-mia, and hypocalcemia. A follow-up creatinine level

Fig. 6 A 5-year-old boy with relapsed JMML. Axial contrast-enhanced CT image shows a solitary mass in the right kidney withmeasured attenuation greater than that of water (arrow). A largeconglomerate upper abdominal lymph node mass is also identified(arrowheads). A solitary mass was present in the left kidney (notshown)

Fig. 7 A 20-year-old man with history of Li-Fraumeni syndrome,relapsed B-cell ALL, and recurrent fibrosarcoma. The man presentedwith back and abdominal pain and elevated serum creatinine level.Axial contrast-enhanced CT image shows asymmetric renal parenchy-mal perfusion, concerning for right kidney malperfusion. Enlargedright renal hilar and retroperitoneal nodes (arrowheads) are seenadjacent to the right renal artery and vein. Mild right renal collectingsystem dilation (asterisks) as well as multiple bilateral renal masses(arrows) are also present

Table 1 Maximum renal size in 12 children with renal leukemicinvolvement

Age atdiagnosisa

Craniocaudad renal length (cm)

Right kidney Left kidney

Maximum Standarddeviationsabove orbelowb

Maximum Standarddeviationsabove orbelowb

7 months 10.0 >+4.0 8.5 >+4.09 months 7.6 +2.5 7.5 +2.53 years 9.5 +3.5 9.4 +3.54 years 8.2 +1.0 9.4 +3.05 years 9.1 +2.0 8.4 +1.06 years 9.0 +1.5 9.5 +2.56 years 10.5 +4.0 10.8 >+4.08 years 9.7 +2.5 10.0 +39 years 10.2 +2.5 10.8 +3.510 years 9.0 ±0.0 8.0 −1.012 years 12.3 +4.0 12.9 >+4.021 years 15.1 >+4.0 14.2 +4.0

a Diagnosis of renal leukemic involvement.b For patient age based on reference [20].

428 Pediatr Radiol (2008) 38:424–430

obtained from this child following chemotherapy and uponresolution of the renal masses was normal (Fig. 8).

The differential diagnosis for low-attenuation renalmasses in children includes, but is not limited to, infection,lymphoma, nephroblastomatosis, simple cysts, angiomyoli-pomas, and metastases (in addition to renal leukemicinvolvement). All of these conditions can be unilateral orbilateral as well as solitary or multifocal. Infection,frequently atypical (fungal) in leukemic patients, presentswith microabscesses that can involve the liver, spleen, andkidneys. These infectious lesions typically appear as focalsmall nonenhancing areas of parenchymal abnormality onCT images [16, 17]. The diagnosis of renal parenchymalinfection can typically be confirmed with urine gram stainand culture. Focal acute bacterial pyelonephritis can alsoproduce low-attenuation renal masses that can be unilateral

or bilateral. In these patients, the clinical presentation(fever, dysuria, flank pain, etc.) and urine gram stain andculture will help to establish the correct diagnosis.

Renal lymphomatous involvement can appear quitesimilar to renal leukemic involvement [18, 19]. Renallymphoma can present with unilateral or bilateral low-attenuation masses or as a relatively diffuse infiltrativeprocess. Lymphoma also has a propensity to involve theperinephric spaces. Imaging alone unfortunately cannotdistinguish leukemic from lymphomatous renal involve-ment. Nephroblastomatosis can also be difficult to distin-guish from renal leukemic involvement based solely on CTimaging features; however, the lack of other sites ofextramedullary disease is a clue. These patients might alsohave a known syndrome such as Beckwith-Wiedemannsyndrome.

Angiomyolipomas can usually be correctly diagnosed byCT imaging with or without intravenous contrast materialwhen they contain demonstrable macroscopic fat, and,therefore, are associated with negative Hounsfield unitmeasurements. Angiomyolipomas are benign neoplasmscomposed of vascular, smooth muscle, and fatty elements.While approximately 80% of such lesions are sporadic,20% can be found in the setting of tuberous sclerosis (andare often bilateral and multiple). The diagnosis of tuberoussclerosis is usually well-established before the renal lesionsare manifest because of central nervous system abnormal-ities that often lead to mental retardation and seizures aswell as specific skin findings. Benign renal simple cysts canalso mimic renal leukemic involvement, although thepresence of multiple bilateral renal cysts is rare in children.Unless they are too small to be characterized by CT, renalcysts usually can be differentiated from neoplasm by theirlower (water) attenuation, sharp margins, and lack ofenhancement. Syndromes such as tuberous sclerosis andvon Hippel-Lindau should be considered when multiplebilateral renal cysts are observed in pediatric patients.

Our study had a few limitations. First, this investiga-tion was a retrospective review of a relatively smallnumber of patients. Although we treat many childrenwith leukemia at our institution, these children do notroutinely undergo abdominal CT imaging. When abdom-inal CT imaging is performed in children with leukemia,renal leukemic involvement is more often than not anunexpected finding. A second limitation is that directpathologic correlation was available for only a singlepatient, and diagnosis of renal leukemic involvement wasmade by correlating clinical and CT imaging features inthe remaining patients. Although the CT imaging findingsdescribed here in children with presumed renal leukemicinvolvement are clearly abnormal, it is still possible(although thought to be unlikely) that another unrecog-nized etiology was responsible for the visualized renal

Fig. 8 A 9-year-old boy with relapsed AML and abdominal pain. aAxial contrast-enhanced CT image demonstrates numerous smallbilateral low-attenuation renal lesions. In addition, there are multipleperitoneal, mesenteric, and greater omental soft-tissue attenuationmasses, best seen along the liver and lateral conal fascia (arrowheads).b Follow-up CT image 1 month later after chemotherapy and BMTshows resolution of the masses

Pediatr Radiol (2008) 38:424–430 429

parenchymal lesions. Third, the true incidence of renalleukemic involvement in children cannot be calculatedfrom this study, as only a fraction of children withleukemia at our institution undergo CT imaging of theabdomen. A final limitation is that some children withrenal leukemic involvement could have been excludedfrom this study because of a lack of overt CT imagingfindings. It is probable that at least some children withrenal leukemic involvement have normal kidneys on CTimaging.

Conclusion

Renal leukemic involvement in children can present with avariety of CT imaging findings. Both nephromegaly andfocal renal parenchymal abnormalities might be observed.When detected on CT imaging, renal leukemic involvementis frequently an unexpected incidental finding. The pres-ence of renal leukemic deposits does not appear to befrequently associated with impaired renal function.

References

1. Chessells JM, O’Callaghan U, Hardisty RM (1986) Acutemyeloid leukemia in childhood: clinical features and prognosis.Br J Haematol 63:555–564

2. Kun LE (1997) Acute lymphoblastic leukemia. Semin RadiatOncol 7:185–194

3. Ikawa Y, Saikawa Y, Horisawa T et al (2007) Pancreatic and renalinvolvement in pediatric acute lymphoblastic leukemia/lympho-ma. J Clin Oncol 25:451–453

4. Sato A, Imaizumi M, Chikaoka S et al (2004) Acute renal failuredue to leukemic cell infiltration followed by relapse at multipleextramedullary sites in a child with acute lymphoblastic leukemia.Leuk Lymphoma 45:825–828

5. Hayek M, Srinivasan A (2003) Acute lymphoblastic leukemiapresenting with lactic acidosis and renal tubular dysfunction. JPediatr Hematol Oncol 25:488–490

6. Gilboa N, Lum GM, Urizar RE (1983) Early renal involvement inacute lymphoblastic leukemia and non-Hodgkin’s lymphoma inchildren. J Urol 129:364–367

7. Jaffe ES, Harris NL, Stein H et al (eds) (2001) Pathology andgenetics of tumours of haematopoietic and lymphoid tissues.IARC, Lyon, pp 45–115

8. Vardiman JW, Harris NL, Brunning RD (2001) The World HealthOrganization (WHO) classification of the myeloid neoplasms.Blood 100:2292–2302

9. Bailey JE, Roubidoux MA, Dunnick NR (1998) Secondary renalneoplasms. Abdom Imaging 23:266–274

10. Araki T (1982) Leukemic involvement of the kidney in children:CT features. J Comput Assist Tomogr 6:781–784

11. Ali AA, Flombaum CD, Brochstein JA et al (1994) Lactic acidosisand renal enlargement at diagnosis and relapse of acute lympho-blastic leukemia. J Pediatr 125:584–586

12. Boueva A, Bouvier R (2005) Precursor B-cell lymphoblasticleukemia as a cause of bilateral nephromegaly. Pediatr Nephrol20:679–682

13. Gore RM, Shkolnik A (1982) Abdominal manifestations ofpediatric leukemias: sonographic assessment. Radiology143:207–210

14. Del Toro G, Morris E, Cairo MS (2005) Tumor lysis syndrome:pathophysiology, definition, and alternative treatment approaches.Clin Adv Hematol Oncol 3:54–61

15. Goldman SC, Holcenberg JS, Finklestein JZ et al (2001) Arandomized comparison between rasburicase and allopurinol inchildren with lymphoma or leukemia at high risk for tumor lysis.Blood 97:2998–3003

16. Lin PC, Chang TT, Jang RC et al (2003) Hepatosplenic micro-abscesses in pediatric leukemia: a report of five cases. KaohsiungJ Med Sci 19:368–374

17. Shirkhoda A (1987) CT findings in hepatosplenic and renalcandidiasis. J Comput Assist Tomogr 11:795–798

18. Chepuri NB, Strouse PJ, Yanik GA (2003) CT of renal lymphomain children. AJR 180:429–431

19. Lowe LH, Isuani BH, Heller RM et al (2000) Pediatric renalmasses: Wilms tumor and beyond. Radiographics 20:1585–1603

20. Han BK, Babcock DS (1985) Sonographic measurements andappearance of normal kidneys in children. AJR 45:611–616

430 Pediatr Radiol (2008) 38:424–430