J Clin Pathol 1982;35:1069-1073

Wilms' tumour and renal dysplasia: an hypothesisHB MARSDEN, W LAWLER*

From the Royal Manchester Children's Hospital, Manchester M27 IHA and the *Department of Pathology,University of Manchester, Oxford Road, Manchester M13 9PT

SUMMARY The incidence of renal dysplasia in a series of Wilms' tumours is presented. Thedistribution of such lesions is discussed, together with their course of development and regres-

sion. The kidney is regarded as a particularly suitable organ for studying the relation betweendysplasia and neoplasia. A schema is suggested for this association with regard to Wilms' tumour.

Paediatric oncology offers the opportunity to studythe interrelations between normal embryonicdevelopment, dysplasia and neoplasia. The kidney isa particularly suitable organ for such study as thereis a clear relation between renal dysplasia andWilms' tumour.' The dysplastic features have beenemphasised' under the general term "the nephro-blastomatosis complex," with "nodular renalblastema" and "sclerosing metanephric hamar-toma" as main subgroups and further subdivision ofthe latter into three types. Associated changes ofglomerular immaturity and sclerosis and of focalcortical cysts' are less clearly dysplastic featuresallied to the development of a Wilms' tumour andmay even be secondary to neoplasia within the kid-ney. Because the term "nephroblastomatosis" hasbeen applied specifically to multifocal and wide-spread epithelial dysplastic lesions2 and because inour experience mesodermal in addition to epithelialhamartomata may be rarely encountered, it hasbeen avoided and "renal dysplasia" used.

Dysplastic lesions consist of metanephric tissuewhich is predominantly tubular but may also showprimitive glomerular formation. The tubules are dif-ferent from those seen in Wilms' tumours and con-sist of small rounded or polygonal hyperchromaticcells rather than the columnar cells which areencountered in nephroblastomas. There is consider-able variation in the extent of dysplastic change,varying between widespread confluent lesions andinvolvement of a single tubule. The small size of theminimal lesions precludes adequate assessmentof the true incidence of dysplasia. Fibrosis is com-monly encountered in dysplastic lesions and com-plete sclerosis with loss of tubular elements appearsto be the natural course of evolution. It is thereforenot possible to exclude previous dysplastic change

Accepted for publication 28 January 1982

when isolated areas of fibrosis are seen and thisfurther compounds the difficulty of assessing thetrue incidence of dysplasia. Occasionally, bands ofsmooth muscle rather than epithelial differentiationare encountered and are difficult to interpret; it ispossible that these may represent mesenchymalelements in renal dysplasia. (Fig. 1).The location of the dysplastic lesions relative to

the Wilms' tumour has important implications con-cerning the relation between dysplasia and neoplasia(see Discussion). Three main sites for dysplasia areseen:1 The immediate subcapsular region of the kidney,

this being the most common (Fig. 2).2 The tumour-kidney margin. Compression atro-

phy of pre-existing normally developed kidneymakes assessment of dysplasia difficult at thissite (Figs. 3 and 4).

3 Within the tumour itself. This is a particular fea-ture of more highly differentiated epithelialtumours (Fig. 5) and is easily recognised infibrous septa (Fig. 6). Occasionally, there may beextension along tissue septa to the periphery andthe tumour-kidney junction. (Fig. 3).

In a recent histopathological analysis by the presentauthors of 133 consecutive unilateral Wilms'tumours, nephroblastic dysplasia was found in 24(18.0%), 22 epithelial and two mesodermal in type.

In this series, congenital anomalies were seen asfollows: naevi and haemangiomata 7; definitehemihypertrophy 4; pyloric stenosis 3; duplex kid-ney 2; Beckwith-Wiedemann syndrome 2; and oneeach of horseshoe kidney, umbilical hernia, urethralvalves, renal cysts, Hirschprung's disease and multi-ple skeletal abnormalities. Two cases of congenitalheart disease, one Ollier's disease and oneHirschprung's disease were noted in siblings. Therewas one case of congenital heart disease and one ofheterochromia iridis in parents. In one patient, renal

1069

copyright. on 19 A

pril 2018 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.35.10.1069 on 1 O

ctober 1982. Dow

nloaded from

Marsden, Lawler

giF ..f

. .jAB

.,

7.o7:.'iW <

Fig. 1 Smooth muscle bundle in thesubcapsular region in a kidney which containeda Wilms tumour. There is loose surroundingconnective tissue but no blastema.Haematoxylin and eosin x 150.

4.4., jse*.

'9

A",

Fig. 2 Subcapsular focus ofsclerosingrenal nephroblastic dysplasia.Haematoxylin and eosin xss.

carcinomata developed in two great grandparents(one paternal, one maternal). The only associationbetween nephroblastic dysplasia and congenitalanomalies in this particular series was one of the twopatients with Beckwith-Wiedemann syndrome whoalso had renal cysts and haemangiomata.

Discussion

In an attempt to correlate the various observationson dysplasia and childhood renal neoplasia, a sug-gested hypothesis is summarised in the accompany-ing diagram (Fig. 7).

1070

Rg.+., I

v. :;.,

..W

I

1. ".1, V. .:''

-4io;`

,.,.ef..

copyright. on 19 A

pril 2018 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.35.10.1069 on 1 O

ctober 1982. Dow

nloaded from

Wilms' tumour and renal dysplasia: an hypothesis

Fig. 3 Nephroblastic<J4ff g dysplasia at the

tumour-kidney junction-,Ns>;> with extension into the

Wffi-^ tumour itself. Haematoxylin^^i f b ~and eosin x55.

Fig. 4 Sclerosing nephroblastic dysplasia withfocal calcification at the tumour-kidneyjunction. Haematoxylin and eosin x 150.

During normal kidney development, pluripoten-tial metanephrogenic blastema becomes committedto form purely renal tissue. Although histologicalappearances indicate that Wilms' tumour is nor-mally derived from committed blastema, the occa-sional finding of heterologous elements (cartilage,

fat, bone etc) suggests that the neoplasm may arisefrom uncommitted metanephrogenic blastema. Themetanephric composition of the dysplastic lesionspoints to an origin from committed blastema and theoccasional finding of smooth muscle would notrefute this possibility.

1071

copyright. on 19 A

pril 2018 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.35.10.1069 on 1 O

ctober 1982. Dow

nloaded from

1072

:^A.:£" V C

44= tt-(

AK>i &.t iS Sw w

Omo

U U -J-i-~~44 w'; -V-~~~~~~~~~

4..MIX

:'.S5

00

::A

t°tf^Xo S*Xb A .. ^, . !. >* .. eh : ..... sF o:s .: ' _ :e,,#:t..............

*it tt .,,> 4eC

. ...... .4=

l @@m.

Marsden, Lawler

Fig. 5 Nephroblastic dysplasiawithin a highly differentiatedepithelial Wilms' tumour.Haematoxylin and eosin xSS.

-4-

Fig. 6 Nephroblastic- dysplasia in fibrous

septa within a Wilms'tumour. Haematoxylinand eosin x 55.

A

I4::R g.

N

Yb:

copyright. on 19 A

pril 2018 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.35.10.1069 on 1 O

ctober 1982. Dow

nloaded from

Wilms' tumour and renal dysplasia: an hypothesis

Fig. 7 Suggested relations between normal, dysplastic and neoplastic renal development. The relative width of the arrowsindicates the probability ofthe course ofevents although the arrows representing normal development have ofnecessity hadto be reduced in size.

The relatively high incidence of tumours with stri-ated muscle as the only heterologous element sug-gests that, although not shown in our Fig. 7, neo-plasms may arise from an intermediate stage be-tween pluripotential and committed metane-phrogenic blastema.The close anatomical association between dys-

plasia and neoplasia (see above) raises the pos-sibilities either that the tumour may arise directlyfrom dysplastic foci or that the two are independentbut linked aetiologically.The existence of dysplastic areas within differen-

tiated rather than undifferentiated tumours may beexplained in three ways. Firstly, the development ofneoplasia from dysplasia may lead to the formationof a highly tubular tumour; secondly, differentiatedrather than undifferentiated tumours and dysplasiamay have the same origin; thirdly, more aggressiveblastemal neoplasms may destroy pre-existing dys-plastic elements. However, it is clear that well dif-ferentiated tumours and dysplasia may coexist with-out apparent damage to the latter.The 18% incidence of nephroblastic dysplasia in

our recently analysed series is rather lower than the33*3% of Bove and McAdams,' but bilateral caseswere excluded from our series, and these authors'emphasise the high association between bilateraltumours and metanephrogenic abnormalities.Knudson and Strong3 concluded that bilateraltumours are more likely to be familial, that familialtumours result from two mutations, one germinaland one somatic, and that sporadic tumours resultfrom two somatic mutations. However, nephroblas-tic dysplasia is a feature of both unilateral and bilat-

eral Wilms' tumours. Although some family data isavailable from the patients in our series, the minimalassociation between nephroblastic dysplasia andcongenital anomalies together with the absence ofbilateral tumours and any family history of nephro-blastoma preclude significant genetic assessment.

Wilms' tumours are neoplasms associated with theembryological development of the kidney, and aretypically located deeply in the renal substance: it iseven possible, considering their well-defined mar-gins, that they are initiated prior to organoid renaldevelopment. The finding of subcapsular dysplasticfoci with a deep-seated neoplasm indicates that theformation of such foci is not confined to the initialperiod of oncogenesis but may occur subsequently.

References

'Bove KE, McAdams AJ. The nephroblastomatosis complex andits relationship to Wilms' tumor: a clinicopathologic treatise.Perspect Pediatr Pathol 1976;3:185-223.

2 Hou LT, Holman RL. Bilateral nephroblastomatosis in a prema-ture infant. J Pathol Bacteriol 1961;82:249-55.

Knudson AG, Strong LC. Mutation and cancer: a model forWilms' tumour of the kidney. J Natl Cancer Inst1 972;48:3 13-24.

Requests for reprints to: Dr HB Marsden, ConsultantPathologist, Royal Manchester Children's Hospital, Man-chester M27 IHA, England.

1073

copyright. on 19 A

pril 2018 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.35.10.1069 on 1 O

ctober 1982. Dow

nloaded from