ofthe

State Keith

S. White,

#{149} Donald

MD

R. Kirks,

MD

#{149} Kevin

E. Bove,

Art

MD

Imaging of Nephroblastomatosis: An Overview’ Nephroblastomatosis ity of nephrogenesis

incomplete

N

is an abnormal-

characterized of primitive

maturation

by

nephrogenic cells. The resultant mdividual lesions have variable histologic findings and distribution in the kidney. There is wide acceptance of nephroblastomatosis as a precursor lesion to Wilms tumor. Consequently, the role of imaging in assessing patients for neoplastic trans-

formation

of nephroblastomatosis

should be tomography be superior

emphasized. Computed (CT) has been shown to to ultrasound (US) in this

regard. lack

Nonetheless,

the low

of radiation

of US

cost

make

and

it attrac-

five for serial screening studies. Although magnetic resonance imaging shows kidney

promise in assessment for nephrogenic rests,

ence

is limited

ing

patients

The

in evaluat-

for nephroblastomatosis.

authors

sessing

in its use

of the experi-

present

patients

protocols

identified

for asas being

at

either high or low risk for the development of Wilms tumor due to nephroblastomatosis. into account

These protocols take advantages of CT

the

and US. Scanning intervals are determined from epidemiologic data on the occurrence of Wilms tumor; the most

intense

screening

occurs

during

the younger ages of life, when the likelihood of Wilms tumor development is highest. Index

terms:

genitourinary

system,

neoplasms,

diagnosis,

81.1449

Nephroblastomatosis,

81.326

Radiology

I

Infants,

#{149} Kidney #{149}

81.14,

From

1992;

the

81.1449

of Radiology

(K.S.W.,

D.R.K.), Pediatrics (D.R.K., K.E.B.), and Pathology (K.E.B.), Children’s Hospital Medical Center,

Cincinnati,

College 14,

1991;

gust dress

accepted

the

University

Center, RSNA,

and

Children’s

P0

Box C-5371, 1992

of’Cincinnati

Cincinnati. revision

14; revision received reprint requests to

Radiology, s:

and

of Medicine,

Received requested

September

K.S.W.,

Hospital

Seattle,

11. Department

and

Medical

WA 98105.

DEVELOPMENT To understand nephroblastomatosis, it is helpful to review normal developmental anatomy of the kidney. The metanephric diverticulum or ureteric bud contacts aggregates of primitive nephroblasts (nephrogenic blastema) in the paraspinal region of the developing embryo. By the process of mutual induction, blastemal tissue causes the ureteric bud to branch and form the collecting system of the kidney; the uretenc bud induces the metanephric blastema to form epithehal and stromal elements of the kidney. The blastema continuously differentiates into nephrons during development so that the peripheral portions of the kidney are the most recently formed and the juxtamedullary nephrons are the oldest. The normal mature kidney consists of an aggregate during

182:1-5

Departments

is a complex abnormality of nephrogenesis; the most severe cases have features of both developmental malformation and neoplasia. Minor degrees of nephroblastomatosis are common in children with unilateral Wilms tumor and are almost universal in those with bilateral Wilms tumor.

June Au-

Adof

EPHROBLASTOMATOSIS

of 12-14 development.

renal

lobes Each

that lobe

fuse con-

sists of a medullary pyramid and its surrounding cortex. In the mature kidney, the cortical portion of the lobe is not confined to the surface of the kidney, because the process of lobar fusion produces columns of cortical tissue that extend from the surface to the pericalyceal soft tissue deep within the kidney. Although the process of glomerular maturation continues after birth, the process of nephrogenesis is complete by 34-35 weeks gestation (1). Microscopic blastemal remnants may be seen in as many as 1% of infants at autopsy (2). Beyond infancy, persistence of metanephric blastema is ab-

IIIl_I!) IIllIllIllIll IIIIIIII!IIIIIIIIIII IIliii III

normal and is the unifying feature of a spectrum of clinically and pathologically related conditions that includes Wilms tumor. Nephroblastomatosis has been defined as the persistence of metanephnc blastema into infancy and childhood. However, a confusing array of terms has appeared in the literature describing the variations in the distribution of blastemal remnants within the kidney (diffuse, superficial diffuse, multifocal, perilobar, intralobar, panlobar, and mixed nephroblastomatosis) and different histopathologic patterns of individual lesions (nodular renal blastema, sclerosing metanephnc hamartoma, Wilms tumorlet) (3-6). Although terminology debates continue, the association of independent foci of persistent blastema with Wilms tumor has been repeatedly documented; for this reason nephroblastomatosis is widely accepted as a premalignant lesion. While blastemal residues or rests have variable histologic findings and neoplastic potential, there seems to be agreement that macroscopic features of the lesions are useful as predictors of malignancy. Consequently, the role of imaging for initial diagnosis and follow-up of patients with nephroblastomatosis should be emphasized.

PATHOLOGIC

FINDINGS

A spectrum of pathologic findings of abnormal metanephric differentiation has been described. In a report of a study of pathologic lesions present in grossly uninvolved parenchyma of resected Wilms tumor specimens and in surgical biopsy specimens from the contralateral kidney, Bove and McAdams (7) described three potentially “neoplastic” lesions (nodular renal blastema, metanephric hamartoma, and Wilms tumorlet) and two “mal-

Abbreviation:

NR

=

nephrogenic

rest.

1

formative” lesions (glomerular immaturity/sclerosis and cortical cysts). Wilms tumorlets were recognized as blastema-rich lesions that had attamed sufficient size to suggest a capacity to evolve to Wilms tumor. Microscopic blastemal rests seemed in some instances to be malignant but also seemed capable of evolving to nephrogenic adenomas or to partially matured metanephric hamartomas. Criteria to establish subclasses of blastema indicating neoplastic potential were not recognized. They proposed that nephroblastomatosis consisted of a spectrum of histologically distinct, pathogenically related lesions that range from frankly benign (malformations of nephrons or end-stage hamartomas) to frankly malignant (expanding blastema containing lesions of variable size) (8). Bove and McAdams also suggested that the distribution of lesions of nephroblastomatosis might have clinical significance. Superficial cortical lesions with multifocal distribution were common in older children who had Wilms tumor (often involving both kidneys). In contrast, diffuse pancortical lesions occurred in early infancy (hence the designation infantile form). Diffuse pancortical nephroblastomatosis is rare and apparently has the least neoplastic potential, since a typical large solitary Wilms tumor is rarely associated. A third form (superficial diffuse) was proposed as intermediate between multifocal nodular and pancortical with respect to age at diagnosis, incidence, and likelihood of association with a large, localized Wilms tumor. This early classification stressed the occurrence of focal lesions at or near the surface of the kidney corresponding to the localization of the original nephrogenic zone, with allowance for the fact that the cortical septa of Bertin-which define the flanks of each renal lobe-extend deep into the renal substance. The existence of a different type of local, often cystic lesion occurring within the renal lobe was documented in a few instances. Subsequently, Kiviat and Beckwith, as well as Machin and McCaughey (9, 10), defined the histologic spectrum of these intralobar lesions and documented a high association with Wilms tumor (particularly the syndromic forms). Consequently, Beckwith (11) has proposed a modification that classifies the lesions with regard to their distribution within the renal lobe. In this scheme, nephroblastomatosis is classified as perilobar, intralobar, panlobar, and mixed. Panlobar nephro-

2

Radiology

#{149}

1.

2.

Figures

1, 2. (1) Low-power photomicrograph shows multiple perilobar NRs of variable size. A expansile nodular neoplasm several centimeters in diameter, B = small hyperplastic NRs with partial maturation, C a regressed NR. (2) Low-power photomicrograph shows a solitary intralobar NR with variable cellular composition that includes stroma, blastema, cystic nephrons, and adipose tissue.

=

=

blastomatosis is equivalent to pancortical. All intralobar lesions lie within the parenchyma of the kidney. While most perilobar lesions are on the surface of the kidney, they may lie deep in the parenchyma in zones corresponding to septa of Bertin. Furthermore, Beckwith et al (12) suggested that all foci of persistent metanephric tissue be referred to as nephrogenic rests (NRs). The presence of multiple NRs is termed nephroblastomatosis. The proposed scheme classifies lesions with regard to their position within the lobe (penlobar, intralobar, combined, and universal) and with regard to the number of NRs (unifocal, multifocal, or diffuse). In an attempt to divide lesions into clinically relevant groupings, the individual lesions are further subclassified as dormant, matuning/sclerosing, hyperplastic, and neoplastic. This subclassification requires evaluation of both microscopic and gross pathologic findings of lesions (Figs 1, 2). Dormant and sclerosing/maturing NRs are typically microscopic. Hyperplastic NRs are macroscopic and tend to be plaquelike. Any macroscopic NR that is nodular and enlarges over time is considered neoplastic (Fig 3). This latest proposal has much to offer in terms of clarity and simplicity. Nonetheless, because the biologic activity of microscopic lesions may not be unequivocally determined by examining their microscopic appearance, further refinement is likely as methods are developed for assessing cell proliferation chromosome

rates and 11p13-associated

loss

suppressor activity (a biologic that is typical of Wilms tumor)

of tumor marker (13).

Figure

3.

expansile with

Gross

hyperplastic

characteristics lesions

face

kidney

neoplastic

specimen nodules

NRs

(arrows)

of aggressive are

contour).

not

nodular Scale

three

associated that

growth and

is in

with

(N)

lack (ie,

do not alter

the

sur-

centimeters.

Some microscopic nodules are likely to be biologically equivalent to frank tumors, whereas others may be truly dormant, nonproliferating lesions that have not undergone the mutations necessary to produce a Wilms tumor. Since original observations were made, substantial evidence has accumulated to support the concept of nephroblastomatosis as a precursor of Wilms tumor. Pathologic specimens submitted to the third National Wilms Tumor Study showed a prevalence of tissue representing NRs in sites adjacent to the Wilms tumor in at least 90% of cases of synchronous bilateral Wilms tumor (12). In cases of metachronous bilateral Wilms tumor, probable NRs were found in 94% of specimens. Of patients with intralobar NRs in association with a unilateral Wilms tumor, 16% will develop metachronous Wilms tumor in the contralateral kidney (12). Lesions proved at biopsy to represent NRs have been docuJanuary

1992

Figure

4.

cortical bar

(a) US scan

lesions lesions.

with

of a patient with nephroblastomatosis (b) CT scan of the same patient

(arrows). Most

lesions

permission,

are

from

plaquelike,

indicating

shows

probable

shows multiple hypoechoic typical low-attenuation

hyperplastic

NRs.

subperilo-

(Reprinted,

19.)

reference

Figure 5. CT scan of the left kidney in a pabent with Wilms tumor in the contralateral kidney

deep

shows

within

represents

to be lower than in those with sporadic aninidia, genitouninary abnormalities, and mental retardation (16).

IMAGING

Figure NRs

in

6.

CT scan

the

left

dominant mass

lesion

expansile was

resent a Wilms background of courtesy

shows

kidney

diffuse

with

mass

an

(arrow).

subsequently tumor

The focal

proved

that

to

occurred

hyperplastic

of Kathleen

perilobar

associated

rests.

H. Caron,

on

repa

(Image

MD, Cincin-

nati.)

mented

to subsequently

transform

Wilms tumor (14). The association of Wilms tumor with sporadic aniridia, genitourinary abnormalities (cryptorchidism and hypospadias), mental retardation,

to

and

a deletion on the short arm of chromosome 11 (lip-) has also been documented in some cases of nephroblastomatosis, but the linkage is inconsistent despite the development of Wilms tumor in up to one-third of patients (15). Other syndromes (Drash [male pseudohermaphroditism, glomerulosclerosis, and Wilms tumorj, Beckwith-Wiedemann, and hemihypertrophy) have also been shown to be associated with Wilms tumor and nephroblastomatosis, but the risk for Wilms tumor in these patients seems Volume

182

#{149} Number

I

STRATEGIES

Microscopic foci of nephroblastomatosis are not identifiable with any imaging modality. In cases of macroscopic nephroblastomatosis, findings at excretory urography include irregular renal contour and distortion of the renal collecting system. In many cases of nephroblastomatosis, results of excretory urography are normal. Ultrasonography (US) typically reveals hypoechoic nodules in the kidney, although the lesions may be isoor hyperechoic compared with normal renal parenchyma (Fig 4) (17,18). The presence of a nodular renal surface is a secondary sign suggestive of the diagnosis, although this finding can be seen normally in cases of persistent fetal lobation. Occasionally, NRs may be cystic. US is less sensitive than computed tomography (CT) in the diagnosis of nephroblastomatosis (19). CT examinations should be performed with intravenous contrast matenial enhancement and dynamic table incrementation. Care should be taken to window the images so the contrast in the renal parenchyma is maximal. Foci of macroscopic NRs typically appear as low-attenuation lesions in the kidney. Although microscopic specimens of NRs occasionally demonstrate fat, microcalcification, or both, CT documentation of fat or calcification has not yet been described, to our knowledge. As with US, nodularity of the parenchymal surface may be noted with CT. CT is currently considered the modality of choice in the evaluation of nephroblastomatosis.

The nance

an

area

of low

the parenchyma, an

intralobar

appearance (MR) imaging

attenuation

which

likely

at magnetic of a single

resocase

NR.

of nephroblastomatosis has been descnibed (20). The lesions appeared as areas of low intensity on Ti-weighted images. The best pulse sequences and the ultimate role of MR imaging in evaluating these patients are not yet known. The role of imaging in evaluating patients suspected of having nephroblastomatosis, Wilms tumor, or both includes (a) preoperative assessment in a patient suspected of having Wilms tumor to exclude macroscopic NRs in the contralateral kidney,

(b) follow-up

of patients with known microor macroscopic nephroblastomatosis to detect neoplastic change, and (c) screening of patients with syndromes associated with nephroblastomatosis and Wilms tumor. Although CT is probably the most sensitive modality in the detection of nephroblastomatosis, small macroscopic foci of NRs can be missed (21). The conventional approach, therefore, has been direct inspection of the contralateral kidney at the time of surgical resection of a Wilms tumor, even if CT findings are negative. However, it is not clear that the small lesions missed at CT are biologically significant if adequate chemotherapy and further imaging follow-up are implemented. Consequently, contralateral surgical exploration is not advocated by some surgeons (22). Imaging probably has an advantage over direct inspection in the diagnosis of intralobar or perilobar NRs deep in the renal parenchyma (Fig 5). This is particularly significant because of the high incidence of development of metachronous Wilms tumor in patients with intralobar NRs. We recommend that CT be used as the initial imaging modality in assessing the contralateral Radiology

#{149} 3

a.

b.

Figure

7.

(a)

parenchyma the previously

Longitudinal

(arrow) noted

renal

US scan

consistent with lesion. Findings

areas within a suspicious lesion a reliable indicator of benignity.

Careful comparisons with measurements should be made to identify lesions that are enlarging (Fig 7). The possibility that a previously microscopic NR might undergo transformation despite the administration of chemotherapy should be acknowledged. Patients considered to be at high risk for development of Wilms tumor include those with known nephroblastomatosis. Specifically, patients who previously had a Wilms tumor and in whom nephroblastomatosis is diagnosed with imaging or biopsy, as well as patients with a de novo diagnosis of nephroblastomatosis, are classified as high risk. The imaging protocol for screening and follow-up in 4

Radiology

#{149}

with

an NR. (b) US scan at surgical resection

kidney preoperatively. Further expenence and comparative studies are needed before the role of MR can be defined. Imaging is vitally important in the follow-up of patients with known nephroblastomatosis. Not all lesions require biopsy or removal. The results of imaging studies dictate which patients should undergo resection of suspicious lesions. The hallmark of neoplastic transformation of a benign NR is enlargement. NRs that become rounded, expansile, and masslike should be removed to rule out neoplastic change (Fig 6). Needle biopsy is inadvisable, since the variable histologic characteristics of NR may result in misleadingly innocent-looking specimens. Although it is presumed that blastema containing NRs will usually be sensitive to the chemotherapy commonly employed, dormant nonproliferating rests may not be ablated by such therapy and, insofar as is known, may be significant precursors of Wilms tumor. Known lesions should therefore be followed with serial examinations. The presence of cystic is not

c.

of a patient

these

sporadic

and

and (c) CT scan obtained confirmed neoplastic

patients

ery 2 months sis,

aniridia

then

every

chromosome 6 months transformation

llpl3later

is as follows: (a) US evfor 1 year after diagno3 months for 1 year,

then every

every 4 months for 1 year, then 6 months for 2 years, then every 12 months until 10 years old; (b) CT every 6 months for 2 years after diagnosis, then every 12 months for 3 years, then again at age 10 years. Any change noted with US should be confirmed with CT. Patients in the low-risk group indude those with Beckwith-Wiedemann syndrome, Drash syndrome, or hemihypertrophy and those who previously had a Wilms tumor in whom there was no biopsy or imaging confirmation of nephroblastomatosis. The imaging protocol in these patients is as follows: (a) US every 4 months for i year after diagnosis, then every 6 months for 1 year, then every 8 months for 2 years, then every 12 months until 10 years old; (b) CT every 6 months for 1 year after diagnosis, then every 12 months for 4 years, then again at age 10 years. Any change noted with US should be confirmed with CT, and any low-risk patient in whom NRs subsequently appear should be switched to the protocol for high-risk patients. Our proposed imaging protocols are designed to balance the advantages of CT (high sensitivity) with those of US (low cost, no radiation exposure). Scanning intervals are graduated; the most intense screening is performed during the early periods following diagnosis of Wilms tumor or between the ages of 1 and 5 years, when the patient is at the highest risk for the development of a neoplasm. Screening intervals take into account the fact that metachronous tumors have been reported to occur 2-52 months following appearance of the primary Wilms tumor (23). Screening should continue to the age of 10

shows

document of this NR

a hyperechoic

marked to Wilms

focus

in the renal

enlargement tumor.

(arrow,

b) of

years, after which the risk of Wilms tumor development becomes minimal (24). MR imaging has not been included in the imaging protocols because, to our knowledge, no comparative studies are available to document its efficacy in following up patients with nephroblastomatosis.

CONCLUSION Through cooperative efforts, such as the National Wilms Tumor Study, understanding of nephroblastomatosis and its link to Wilms tumor has greatly increased during recent years. With the emergence of a more rational classification, the vital role of diagnostic imaging in the initial assessment and follow-up of these patients is being refined. The challenge remains for radiologists to become active contributors to such multicenter trials to assess the appropriate use of newer technologies such as MR imaging in the evaluation of these patients. #{149} Acknowledgment: Hunter

for

The her

authors

thank

Juanita

secretarial

assistance.

Nephrons & Row,

and kidneys. New Heober Medical Divi-

References 1.

2.

J.

Oliver York:

Harper

sion, 1968; 64. Beckwith JB. nington mor

JL, pathology:

pelvis, ington, 3,

thology, Hennigar

and

matosis. Machin

JB,

tumors ureter,

2nd

DC: Armed 1975; RA,

Clinicopathologic 4.

Nephroblastoma.

Beckwith

Urology GA.

features 1989; Persistent

5.

of Wilms’

tumor:

review.

context Pediatr

of embryology 1-lematol Oncol

cursor

of Wilms’

Benof tu-

kidney,

fasc Institute

HB,

renal

12. Wash-

Garvin

of PaAJ.

of nephroblasto33:259-270. renal

blastema

as a frequent

clinical

Machin GA. (nephroblastomatosis)

Atlas

of the ser,

Forces

32-40. Othersen

(nephroblastomatosis) cursor

In: eds.

pre-

a pathological

and

I. Nephroblastomatosis

Persistent tumor:

in

and genetics. Am 1980; 2:165-172. renal blastema as a frequent prea pathological

January

and

1992

6.

clinical review. II. Significance of nephroblastomatosis in the genesis of Wilms’ hamor. Am J Pediatr Hematol Oncol 1980; 2:253-261. Machin GA. Persistent renal blastema (nephroblastomatosis) as a frequent precursor of Wilms’ tumor; a pathological and clinical review. III. Clinical aspects of neph-

11.

roblastomatosis.

13.

col 1980;

Am J Pediatr

Hematol

On-

8.

tomatosis complex and its relationship to Wilms’ tumor: a clinicopathologic treatise. Perspect Pediatr Pathol 1976; 3:185-223. Bove KE, McAdams AJ. Multifocal neph-

9.

10.

McAdams

AJ.

The

lesion

Beckwith Nephrogenic diatr Huff

of Wilms’

novo

tumour

1990;

14.

Pe-

cal features National

17.

of nephroblastomatosis

J.

Pro-

etiologic tumor. 8:231-234.

NE, Beckwith of Wilms’ Wilms’

Tumor

and

imaging

23.

relationship to subAm J Pediatr HemaJB.

24.

Epidemiologi-

tumor:

results

Study.

of the

JNCI 1982;

.

Magnetic

tumor. Cormier

Urology 1986; PJ, Donaldson

Nephroblastomatosis:

sis.

Radiology

Koo

AS,

1988;

Koyle

MA,

of contralateral

Radiol

KA, Donaldson

JS,

comparison

Radiology

RB, Ehrlich

resonance

F.

necessity

to Wilms’

Franken EA Jr. Yiu-Chiu V. Smith WL, Nephroblastomatosis: clinicopathologic significance

21

22.

R, Bailie MD, Bernstein

Pediatr

Nephroblastomatosis:

1988; 166:153-156. Kangarloo H, Dietrich

20.

Cenet

68:429-436.

53.

et al.

Strong

origin of de of chro-

tumor (letter). J Pediatr 1980; 96:178. Heideman RL, McCavran L, Waldstein C. Nephroblastomatosis and deletion of 1 1 P:

Breslow

19.

a!.

11 p13. Am J Hum

sequent Wilms’ tol Oncol 1986;

(neph-

intralobar multifocal nephroHistopathology 1984; 8:35-

deletion

tumor.

VM,

imaging.

of CT with US and urography.

47:155-160.

Kulkarni

the potential

16.

Parental

constitutional

gression

15.

of Wilms’ 10:1-36. A, Riccardi

and

1984; 14:392-395. Fernbach SK, Feinstein

NB, Bonadio JF. nephroblastomatosis,

CF.

band

significance

In:

JB, Kiviat rests,

Pathol 1990; V, Meadows

mosonal

Inst 1978;

61:285-289. Kiviat NB, Beckwith JB. Intralobar nephrogenic rests and related Wilms’ tumors. Presented at Pediatric Pathology Club interim meeting, Durham, NC, October 28, 1981. Machin GA, McCaughey WTE. A new roblastoma): blastomatosis.

Pathological

and the pathogenesis

nephroblas-

J Nati Cancer

neoplasia.

precursor

12.

JB.

in childhood.

LC, Saunders

Bove

roblastic

aspects of renal Broecker BH, Klein FA, eds. Pediatric tumors of the genitourinary tract. New York: Liss, 1988; 38. tumors

2:353-362.

7.

KE,

Beckwith

imaging

RM,

et

of Wilms

28:203-207. JS, Conzalez-Crussi missed

diagno-

169:737-738. Hurwitz

RS, surgical

et al.

The

explora-

tion in Wilms’ tumor with modern noninvasive imaging technique: a reassessment. Urol 1990; 144:416-417. Jones B, Hrabovsky E, Kiviat N, et al. Metachronous bilateral Wilms’ tumor: National Wilms’ Tumor Study. Am J Clin Oncol 1982; 5:545-550. Breslow N, Beckwith JB, Ciol M, et al. Age distribution of Wilms’ tumor: report from the National Wilms’ Tumor Study. Cancer Res 1988; 48:1653-1657.

et al.

characteristics.

AJR 1982; 138:950-952. 18.

Montgomery

Multifocal

Volume

182 #{149} Number

1

P. Kuhn

JP, Berger

nephroblastomatosis:

PE, et al.

clinical

Radiology

#{149} 5