Cancer in Children: Clinical Management, 5th Edition

Chapter 21. Nephroblastoma

Beatriz de Camargo

Kathryn Pritchard-Jones

Nephroblastoma (Wilms tumour) is the most common renal tumour in children comprising 90 per cent of kidney cancer in this age group. The survival of Wilms tumour patients has improved to 90 per cent in localized disease and >70 per cent for metastatic disease. Challenges remain in identification of histologic and molecular risk factors for stratification of treatment intensity to allow safe reduction in therapy and avoidance of late sequelae for the majority while leading to increased biologic insights and new therapies for the minority of high-risk tumours.

Epidemiology and genetics

Wilms tumour was originally proposed as an index tumour of childhood which could serve to gauge quality of case ascertainment on the grounds that it was thought to have almost constant incidence throughout the world. However, it is now clear that there is international variation. The annual incidence of Wilms tumour has been found to average 7 per million children aged >15 years. A lower incidence has been noted in Asian countries such as Japan, India, and Singapore, as well as in Asian children living in more developed countries. In contrast, a higher incidence is found in Scandinavia, Nigeria, and Brazil. The highest documented rates are among the non-White population of Great Delaware Valley, USA (13.7 per million), the Black population of Los Angeles (11.8) and the SEER Black population (11.1). The incidence rates within Europe range from 6.5 per million in Spain and Slovakia to >10 per million in Northern European countries.1 The highest rates are experienced by children aged <5 years, the incidence declines to 5–6 per million in children aged 5–9 years and approximately 1 per million in those aged 10–14 years. The frequency appears to be equal in males and females. There is some evidence of a difference in age distribution between sexes. Data from the National Wilms Tumour Study Group (NWTS) and the Brazilian Wilms Tumor Study Group suggest that median age at diagnosis is slightly older for girls than boys.2,3

Wilms tumour arises in three different clinical situations. First, the majority occur in children with no unusual physical features or family history and are considered ‘sporadic’. It can occur in association with several congenital anomalies, especially aniridia, hemihypertrophy, and genitourinary anomalies (cryptorchidism, hypospadias, horseshoe kidney), and with specific syndromes such as the Beckwith–Wiedeman, Denys–Drash and WAGR (Wilms tumour, aniridia, genitourinary abnormalities, mental retardation) syndromes.4 The third setting in which Wilms tumour arises is within certain families. These findings have helped to identify children at increased risk of Wilms tumour and allowed formulation of recommendations for screening programmes (see later).

In 1972, Knudson and Strong5 applied the two-step mutational model that had so successfully described the development of retinoblastoma to Wilms tumour. The Wilms tumour suppressor gene WT1 (at chromosome 11p13) was the second tumour suppressor gene to be cloned. This followed the observation that WAGR syndrome is associated with interstitial deletions of chromosome 11p13 and that tumour tissue at this same locus often displays loss of heterozygosity at this same chromosomal region. The WT1 gene provides a fascinating example of the close relationship between normal embryonic kidney development and predisposition to cancer. WT1 plays an essential role in genitourinary development, with different phenotypes associated with different mutations. Complete loss of one WT1 allele, as in the WAGR syndrome, gives rise to a less severe genitourinary malformation than is seen in Denys-Drash syndrome, where intragenic mutations are associated with early-onset nephritic syndrome and frequent pseudohermaphroditism. Both syndromes are associated with a substantial risk of Wilms tumour of 50 per cent. However, unlike the retinoblastoma model, the WT1 gene is somatically mutated in only 20 per cent of Wilms tumours and linkage to the WT1 locus has been excluded in most familial cases.

Several genes are now known to be involved in predisposition to Wilms tumour and/or in the somatic events occurring in sporadic tumours. The Beckwith–Wiedemann overgrowth syndrome is caused by mutation or deregulation of several genes that regulate somatic growth and are clustered in a locus more distal to the WT1 gene on chromosome 11 at 11p15.5. These genes (IGF2p57KIP2H19KVLQT1) are all subjected to imprinting during normal embryogenesis, which means that the expression from each allele depends on its parental origin. Beckwith-Wiedemann syndrome is associated with either mutation or deregulated imprinting of one or more of these genes, some of which are also affected in sporadic Wilms tumours.

A family history of Wilms tumour occurs in 1–2 per cent of all cases. Typical pedigrees are small, with only two or three affected relatives and without predisposition to other tumour types. Genetic linkage studies in two unusually large pedigrees have localized one gene for familial Wilms tumour, FWT1, to chromosome 17q. Another locus has been suggested at 19q and a recent evaluation of all available pedigrees in the UK has confirmed that there is genetic heterogeneity for FWT genes with at least four families clearly unlinked to any currently identified Wilms tumour locus.6 Until these genes are identified, it is difficult to predict what, if any, involvement they would have in sporadic Wilms tumour, at both the constitutional and somatic levels. However, the penetrance of FWT1 at least appears to be low, of the order of 15–30 per cent. Therefore it is possible that a substantial proportion of apparently sporadic cases of Wilms tumour carry a constitutional mutation in a low-penetrance familial Wilms tumour gene. Very few familial cases are due to WT1 mutation. The incidence of constitutional WT1 mutation in apparently sporadic cases is a little higher. In a study of 201 cases of Wilms tumours selected from the NWTS cases, eight constitutional WT1 mutations were found and these were virtually confined to boys with cryptorchidism as well as Wilms tumour (7/28 patients).7 A separate study from Germany showed that somatic WT1 mutation was common in Wilms tumours showing stromal predominant histology (13/26) and, remarkably, nearly all of these (10/13) had a constitutional WT1 mutation.8

Finally, allele loss studies in sporadic Wilms tumours suggest a role of other tumour-suppressed genes at 16q, 1p, and 22q that may be involved in progression to more aggressive tumours. Mutation of the p53 gene is associated with the anaplastic subtype.9

The cause of genetic events described is unknown, but several case–control studies of Wilms tumour have identified environmental exposures that may contribute to its development. These include prenatal irradiation, penthrane anaesthesia during birth, household pesticides, parental occupation to pesticides prior to birth, and maternal use of dipyrone.10,11


Clinical presentation

The most common initial manifestation of Wilms tumour is the presence of an asymptomatic abdominal mass. Other associated signs and symptoms include malaise, abdominal pain, gross or microscopic haematuria, fever, anorexia, and hypertension. Abdominal pain may be the result of local distension, spontaneous intralesional haemorrhage, or tumour rupture. The incidence of hypertension is variable because of lack of adequate documentation, but has been reported in 30–63 per cent of children. The aetiology is an increase in circulating renin or a renin-like substance either secreted by the tumour itself or secondary to impaired renal circulation from compression by the tumour mass. Paraneoplastic manifestations associated with Wilms tumour include acquired von Willebrand disease, tumour-induced glomerulone-phritis, and erythropoietin and hyaluronidase secretion.

Imaging studies

Imaging studies should be limited to identifying intraor extrarenal tumour, the presence of a normal functioning contralateral kidney, tumour thrombus in renal vein, inferior vena cava, and heart, and the presence of pulmonary or liver metastases. It is rare for Wilms tumour to metastasize to brain or bone, but imaging of these sites should be performed when other renal tumours (malignant rhabdoid tumour or clear cell sarcoma of kidney, respectively) are identified on histology. Ultrasound, CT scan, and MRI all have their particular advantages. Ultrasound has the specific advantage of assessment of blood vessels for flow and tumour thrombus, and can achieve most of the purposes defined above. CT is superior at evaluating intra-and extrarenal involvement and in demonstrating small lesions within the kidney and liver and the presence of anomalies such as horseshoe kidneys. MRI may possess the advantages of both, but the lesser availability, increased cost, and requirement for sedation probably outweigh the advantages. With current treatment regimens, it appears that chest radiography is sufficient to detect clinically significant pulmonary metastases (see the section on staging for further discussion).

Differential diagnosis

The differential diagnosis on imaging includes other abdominal neoplasms such as neuroblastoma and hepatoblastoma. For a primary intrarenal tumour, the differential includes malignant rhabdoid tumour (MRTK), clear cell sarcoma, hypernephroma (renal cell carcinoma), congenital mesoblastic nephroma (CMN), and benign processes involving the kidney. The likelihood that a primary intrarenal tumour will be a non-Wilms tumour increases with age <6 months (where CMN, MRTK, and renal dysplasia predominate) or >7 years (where other malignancies start to increase).


Immediate biopsy is not always recommended when a tumour has the typical imaging and clinical features of Wilms tumour. The UK Children's Cancer Study Group (UKCCSG) has recently analysed the role of biopsy and shown that 12 per cent of renal tumours with typical features on imaging studies proved to be non-Wilms on pre-chemotherapy biopsy; however, only 1.5 per cent were benign tumours. The morbidity associated with biopsy is small.12


Tumour stage is one of the most important criteria for determining therapy in Wilms tumour, and its accurate determination requires meticulous attention to detail. Staging is based primarily on the anatomical extent of disease, determined after histologic examination of the nephrectomy specimen, as well as surgical factors such as tumour spillage or rupture before or during surgery and completeness of excision.

The staging systems used by two major cooperative groups are somewhat different and reflect that the NWTS system is based on staging of a previously untreated tumour, while the International Society of Paediatric Oncology (SIOP) classification is based on nephrectomy following preoperative chemotherapy (Table 21.1). Staging of a tumour with extension along the renal vein into the inferior vena cava can be problematic. Provided that the tumour can be removed completely en bloc from within the inferior vena cava, it remains a stage II. However, if the tumour has to be removed piecemeal, with the likelihood of microscopic residue, or remains adherent to the vessel wall, it should be considered stage III. Cases where only fibrosis is found in the resected vessel wall can usually be considered stage II, but such cases need careful consideration, especially in relation to their prior therapy. For the SIOP staging system after preoperative chemotherapy, the presence of necrotic tumour or chemotherapy-induced change in the renal sinus and/or within the perirenal fat should not be regarded as a reason for upstaging a tumour provided that it is completely excised and does not reach the resection margins (i.e. from stage I to II). However, if evidence of necrotic tumour is found in a lymph node or at the resection margins, the tumour is assigned as stage III because of the possibility that some viable tumour is left behind in an adjacent lymph node or beyond the resection margins.

Assignment of stage IV due to lung metastases has always been based on a chest radiography assessment in all previous international trials. Therefore any move to incorporate detection of ‘micrometastases’ seen only on CT scan requires careful consideration. Thorax CT will detect pulmonary nodules of size <10 mm that would not generally be apparent on chest radiography (note that a larger nodule apparent on CT but not visualized on chest radiography solely due its position should be considered a metastasis). These nodules may represent either micrometastases or benign lesions. A retrospective review of 141 children treated in the UKW2 trial who had a normal chest radiograph and had a thorax CT performed (but were treated according to the results of their chest radiograph) revealed that 22 per cent had one or more small nodules visible on CT scan only. The median size of these nodules was 5 mm (range 2–8 mm) and all disappeared following chemotherapy. There was no significant difference in pulmonary relapse rate between children with normal thorax CT and those with nodules, with the exception of children treated for stage I tumours with vincristine monotherapy.13 A similar study from the NWTS also suggests that, provided that presumed ‘micrometastases’ are exposed to at least two-drug chemotherapy, this is probably adequate therapy. However, neither study incorporates sufficient numbers to reach a definitive conclusion on this important point and prospective studies are needed.

Table 21.1. Clinical pathologic staging



Tumour limited to the kidney and completely exciseda


Tumour extending beyond the kidney but completely excised

Invasion of renal sinus and/or extra renal vessels and/or perirenal fat


Invasion beyond capsule
Any abdominal lymph nodesb
Tumour rupturec
Peritoneal tumour implants
Incomplete excision


Haematogenous metastases (lung, liver, bone, brain, lymph nodes outside abdominal pelvic region)


Bilateral renal tumours at diagnosis

aIn the most recent NWTS 5 trial, the definition of stage I was further refined to exclude those tumours showing involvement of blood vessels in the renal sinus.
b Previous SIOP studies had a stage II N + category that is now included in stage III as there are no clear data to suggest that hilar LN metastases confer a better prognosis than para-aortic nodes and it is often difficult to make this anatomical distinction on the pathology specimen.
cThe NWTS allows minor tumour spillage confined to the flank to remain within the stage II category, whereas in SIOP this is stage III.

Assignment of the correct tumour stage requires close cooperation between the surgeon and pathologist to ensure that the nephrectomy specimen is delivered intact for proper assessment of any capsular infiltration and completeness of resection margins as well as proper orientation of any additionally resected tissues.


Nephroblastoma covers a large spectrum of special variants that differ in their morphologic features and in their prognosis and natural history. Typically, Wilms tumour presents as a solitary spherical unicentric mass, sharply demarcated from the adjacent renal parenchyma. Multicentric lesions and bilateral tumours are not infrequent. Calcification is rare, but may be detected radiographically in 10 per cent of cases, particularly after haemorrhage in the tumour. Most tumours exhibit a so-called ‘triphasic’ appearance, including blastemal, stromal, and epithelial cell types. However, it is important to recognize that many specimens express only two of these cell types and that monomorphous specimens are not rare. It is important to note that the histologic appearance of a tumour depends on its prior therapy, with different classification schemes used for immediate nephrectomy than for tumours exposed to preoperative chemotherapy (Table 21.2).

Although ‘blastemal predominance’ is recognized in chemotherapy naïve tumours, this is not the same entity as the ‘blastemal type’ defined in the new SIOP classification, where survival of large amounts of blastema following pre-operative chemotherapy is believed to confer an adverse outcome with current therapies.

The NWTS classifies tumours into ‘favourable’ and ‘unfavourable’ subtypes. Although the latter used to encompass MRTK and CCSK, these have been recognized as distinct biologic entities for many years now. However, any classification scheme of renal tumours of childhood usually retains MRTK and CCSK in a ‘high-risk’ category, while placing other entities such as congenital mesoblastic nephroma or cystic partially differentiated nephroblastoma in a ‘lowrisk’ category.

A review of the histologic patterns seen in the NWTS-4 showed that tumours composed predominantly of epithelial cell types were usually low stage (81.3 per cent stage I) and presented at a younger age (median 17 months), while predominantly blastemal tumours were much more aggressive (76.3 per cent stage II–IV) and presented at an older age (median 57 months).14 Similarly, tumours displaying large amounts of stromal elements, particularly those showing differentiation towards skeletal muscle, are known to have a good outcome despite showing little shrinkage in response to chemotherapy. A review of cases treated in a German study (GPOH) according to SIOP protocols dealt with a group of unilateral localized Wilms tumours which were treated with a standardized preoperative chemotherapy and a small group of immediately operated Wilms tumours. An important difference between the two groups was in distribution of histologic subtypes. The percentage of tumours showing blastemal predominance decreased significantly from 39.4 to 9.3 per cent.15 This latter group, with blastema resistant to chemotherapy, was more likely to relapse irrespective of stage. The SIOP Working Classification of Renal Tumours of Childhood has recently been revised to incorporate the different clinical behaviours of the various histologic subtypes defined (Table 21.3).16 Completely necrotic nephroblastoma shows an excellent prognosis and will be included in the low-risk category.17 However, this subtype tends to be overdiagnosed by local pathologists compared with the review pathologist.

Table 21.2. The Revised SIOP Working Classification of Renal Tumours of Childhood (2001)16

A. Pretreated cases:


Low-risk tumours
   Mesoblastic nephroma
   Cystic partially differentiated nephroblastoma
   Completely necrotic nephroblastoma


Intermediate-risk tumours
Nephroblastoma: epithelial type
   Nephroblastoma: stromal type
   Nephroblastoma: mixed type
   Nephroblastoma: regressive type
   Nephroblastoma: focal anaplasia


High-risk tumours
   Nephroblastoma: blastemal type
   Nephroblastoma: diffuse type
   Clear-cell sarcoma of the kidney
   Rhabdoid tumour of the kidney


Other tumours or lesions

B. Primary nephrectomy cases:


Low-risk tumours
   Mesoblastic nephroma
   Cystic partially nephroblastoma


Intermediate-risk tumours
   Non-anaplastic nephroblastoma and its variant
   Nephroblastoma: focal anaplasia


High-risk tumours
   Nephroblastoma: diffuse anaplasia
   Clear-cell sarcoma of the kidney
   Rhabdoid tumour of the kidney

Anaplastic Wilms tumour is a morphologically defined entity containing cells with very large hyperchromatic nuclei associated with multipolar mitotic figures. These changes can be focal or diffuse and occur in about 5 per cent of tumours. Anaplasia is deemed ‘unfavourable’ or ‘high-risk’ by both classification schemes. Recently, the definition of focal anaplasia has been limited to tumours showing a single focus of anaplastic features representing <10 per cent of the tumour, and these tumours are now considered to behave in the same way as the ‘intermediate-risk’ or ‘favourable histology’ groups.18 Preoperative chemotherapy does not obliterate or produce anaplasia, but it makes it more obvious because non-anaplastic areas are probably destroyed by chemotherapy while anaplastic foci remain unchanged. Anaplasia is thought to represent a more resistant rather than a more aggressive cell type. The age distribution of anaplastic tumours is distinctive, with a median age of 61 months. Anaplasia is almost never seen in children aged <2 years.

Table 21.3. Histologic risk groups in Wilms tumour




Completely necrotic
Cystic partially differentiated

Favourable histology

All non anaplastic histologies
Focal anaplasia


Epithelial type
Stromal type
Mixed type
Regressive type
Focal anaplasia


   Blastemal type
   Diffuse anaplasia

Unfavourable histology

Diffuse anaplasia

Nephrogenic rests have been considered as precursor lesions of Wilms tumour for several decades. They may become hyperplastic or neoplastic and form multicentric or bilateral renal lesions. Nephrogenic rests have been reported in 25–40 per cent of kidneys harboring Wilms tumour and in almost 100 per cent of bilateral Wilms tumour. Two types of rest have been recognized: perilobar rests (PLNRs) and intralobar rests (ILNRs). PLNRs are more common in patients with hemi-hypertrophy and Beckwith–Wiedeman syndrome, whereas ILNRs are associated with WAGR and Denys–Drash syndromes and with an earlier age of onset of Wilms tumour. Based on their morphology and clinical associations, ILNRs are thought to result from a defect early in nephrogenesis and PLNRs from a slightly later insult to the embryonic kidney.19

Clear cell sarcoma and rhabdoid tumour are now considered to be biologic entities distinct from Wilms tumour. Clear cell sarcoma of the kidney (CCSK), the bone metastasizing renal tumour of childhood, is a primitive mesenchymal neoplasm which comprises 4 per cent of renal tumours in children. Bone and brain metastases develop in 23 per cent and 15 per cent, respectively, and may be the first site of recurrence. The need to perform a skeletal survey, or a radionuclide bone scan, and a brain CTat time of diagnosis has been emphasized. The addition of adriamycin to vincristine and actinomycin D has improved survival from 25 to 71.9 per cent (NWTS-3). Rhabdoid tumour is considered to be one of the most malignant tumours of childhood, comprising 2 per cent of childhood renal cancer, and is characterized by a uniform cellular infiltrate with abundant eosinophilic cytoplasm, initially interpreted as rhabdomyoblastic or sarcomatous but which may be of neural crest origin. Most are diagnosed in the first year of life and are characterized by early widespread metastases and a poor response to therapy, with a survival of 25 per cent according to NWTS data. The finding of a common molecular defect, deletion of the SNF–INI1 gene on 22q, in rhabdoid tumours from both renal and extrarenal sites, including the associated brain tumours, shows that these constitute a biologic family of tumours regardless of their anatomical site.


Two major clinical trial groups have investigated treatment in Wilms tumour: the International Society of Paediatric Oncology (SIOP) formed in 1970 in Europe, and the National Wilms Tumour Study Group (NWTS) launched in 1969 in North America. Other cooperative groups such as UKCCSG in the UK, GCBTTW in Brazil, and GPOH in Germany have contributed to refining treatment. Overall survival is now approaching 90 per cent for children with Wilms tumour in many countries around the world. Hence the main objective of most trials is to treat patients according to well-defined risk groups in an attempt to achieve the highest cure rates, while decreasing the frequency and intensity of acute and late toxicity and minimizing the cost of therapy. There remain subgroups of Wilms tumour with considerable room for improvement in survival rates.

Combined modality strategies using surgery, radiotherapy, and chemotherapy are the key to success in treating Wilms tumour. Similar outcomes by stage are now reported by all the clinical trial groups (Table 21.4) Therefore the ongoing challenge is to identify better prognostic markers for stratification of therapy, intensifying treatment for patients with poor prognosis, while reducing interventions for those at standard risk. Extent of disease (stage) and tumour histology defined as favourable and unfavourable (NWTS) or low, intermediate, and high risk (SIOP), remain the two most significant prognostic variables for tailoring treatment. Treatment consists of chemotherapy with one to three or more drugs together with surgical excision of the affected kidney. Radiotherapy is used only if there is residual or spilled tumour in the abdomen or metastases.

Preoperative strategies

The benefit of preoperative strategies to facilitate complete surgical removal is well known. The NWTS recommends this approach for a few select groups of patients including those with bilateral renal tumours, tumour extension into the inferior vena cava above the hepatic veins, and tumours found to be inoperable at surgical exploration. The major concern is the potential for loss of important staging information. The approach of SIOP studies is to ‘downstage’ the disease with preoperative chemotherapy and to avoid radiotherapy in most of the patients. In the SIOP preoperative chemotherapy approach, tumours are downstaged so that only 15 per cent of patients have stage III disease and 50 per cent have stage I. Both approaches have similar cure rates and the advantages and disadvantages of each are constantly discussed. The UKCCSG has recently completed a randomized trial comparing the two approaches in terms of their impact on stage distribution and overall burden of treatment. One advantage of the preoperative chemotherapy approach is its potential to allow the in vivo response to treatment to be used as an individual prognostic parameter for risk-adapted postoperative treatment.

Table 21.4. Relapse-free survival with favourable histology


Survival (%)

Stage I

Stage II

Stage III

Stage IV

UK2 (2 years)





NWTS (5 years)





SIOP-93-01 (2 years)





*Note that stage is defined at time of nephrectomy and is therefore not directly comparable. In SIOP-93-01 tumour stage after preoperative chemotherapy. In UK and NWTS trials no pre-operative chemotherapy was used.


Nephroblastoma was the first paediatric malignant solid tumour found to be responsive to systemic chemotherapy using actinomycin D. Others agents such as vincristine, adriamycin, cyclophosphamide, ifosfamide, etoposide, and carboplatin have all been identified as effective in Wilms tumour. Cooperative groups have performed randomized trials to evaluate the necessity of singleor multi-agent chemotherapy in the treatment of Wilms tumour. Lessons learned from cooperative study groups (NWTS, SIOP, UKCCSG, BWTSG) include the following.20,21,22,23,24

1.   Preoperative chemotherapy is equivalent to radiotherapy in reducing surgical morbidity and tumour stage (SIOP 1 and SIOP 5 trials).

2.   Four weeks of preoperative chemotherapy with two drugs is as effective as 8 weeks in producing a more favourable stage distribution for localized tumours (SIOP 9 trial).

3.   Limited courses of two drugs (dactinomycin, vincristine) or even a single agent (vincristine) are sufficient for patients with stage I tumours after immediate surgery (NWTS, UKCCSG). Patients with stage I tumours after preoperative chemotherapy with two drugs require only a short postoperative course of the same treatment (SIOP 6 and SIOP 93–01). The blastemal type after preoperative chemotherapy is now considered high risk in the current SIOP WT 2001 trial, where stage I high-risk tumours receive postoperative treatment with three drugs (dactinomycin, vincristine, Adriamycin).

4.   A single-dose schedule of dactinomycin is as effective as the standard 5-day divided-dose regimen and is less myelosuppressive as well as reducing the number of hospital attendances for children and their parents (NWTS, BWTSG).20,21,24 The NWTS reported an unusual hepatotoxicity and reduced the single dose from 60 to 45 µg/kg.25

5.   For stage II tumours, two drugs are sufficient with excellent cure rates. The addition of adriamycin and radiotherapy did not increase survival in patients treated with immediate surgery (NWTS3). Radiotherapy is not necessary for stage II tumours after preoperative chemotherapy.20

6.   The need for adriamycin in treatment of stage III tumours is controversial. Previous studies including small numbers suggested that adriamycin was of benefit for disease-free but not overall survival (NWTS3). In view of increasing concerns about long-term cardio toxicity in Wilms tumour survivors, the current SIOP WT 2001 trial is readdressing this question but excluding the newly defined ‘high-risk’ histology tumours from the randomization between two (vincristine, dactinomycin) and three (vincristine, dactinomycin + adriamycin) drugs postoperatively.

7.   Radiotherapy is presumed to be necessary for stage III tumours, although no randomized trials addressing this question have yet been attempted. There is anecdotal evidence that radiotherapy may be avoidable in some younger children.

8.   The addition of adriamycin to vincristine and dactinomycin improved survival in stage IV patients (NWTS 2). The question of lung irradiation is discussed in the section on radiotherapy.

Patients with diffuse anaplastic tumours still have a poorer prognosis and there is a need to identify new drugs or schedules with activity. Tumours with focal anaplasia are now treated with less therapy than previously. The prognosis of patients with clear cell sarcoma has improved significantly with the addition of adriamycin. Whether cyclophosphamide should be added is still unanswered. The prognosis for rhabdoid tumours remains dismal.


Surgical staging and tumour resection remain a major component of therapy. A large transabdominal transperitoneal incision is recommended for adequate exposure. Complete excision without spillage and adequate exploration of the abdominal cavity should be done. Lymph node sampling should be performed even if the lymph nodes appear normal, but radical lymph node dissection is unnecessary. This sampling is important because 11 per cent of lymph nodes believed to be negative for disease by the surgeon will harbour metastatic disease.26 The obligatory exploration of the contralateral kidney is controversial given the accuracy of modern imaging techniques. Several recent studies have found a very high rate of accuracy of modern imaging in detecting small contralateral lesions.27 The renal vein and vena cava should be carefully palpated to exclude tumour thrombus. If present, an attempt should be made to remove the thrombus en bloc with the tumour. A thrombus extending to the infrahepatic vena cava should be removed through a vena cavotomy.


Wilms tumour is a particularly radiosensitive cancer. The addition of postnephrectomy radiotherapy increased the survival of Wilms tumour patients to 50 per cent. Clinical trials (NWTS and SIOP) have studied the interrelationship of chemotherapy and radiotherapy, and have been able to reduce the indications for radiotherapy and the doses used without apparent deleterious effect on the survival rates. The NWTS now recommends 10 Gy for flank irradiation compared with 14.4 Gy used in SIOP studies. Whole-abdominal irradiation is usually still recommended for diffuse tumour rupture, although its benefit over flank irradiation has been questioned. It is also possible that abdominal radiotherapy can be safely avoided in younger children even when they have stage III tumours.28

Current treatment recommendations for the use of radiation as part of the multimodality regimen are as follows.

1.   Stage III: positive lymph nodes and/or diffuse peritoneal spillage.

2.   Stage II–IV: diffuse anaplastic tumours.

3.   Stage IV: sites of metastatic disease (usually lungs).

It should be noted that SIOP studies limit lung irradiation to the 25 per cent of stage IV patients who do not achieve a complete remission of lung metastases following preoperative chemotherapy and metastatectomy; NWTS retains low-dose lung irradiation for all children with pulmonary metastases visible on chest radiography at diagnosis.

Prognostic factors

Many prognostic factors have been identified, although they may change as treatment becomes more effective. Older age at diagnosis has been identified as an adverse factor in many studies. Conversely, age <2 years appears to identify an exceptionally good prognosis, especially for those with smaller tumours. Extent of disease at diagnosis (stage) and tumour histology remain the two most significant prognostic variables. Although stage IV disease still clearly identifies a group with poor outcome, the prognostic significance of stage in localized tumours has reduced because of the risk-adapted therapy they receive. The most important tumour histologic feature is diffuse anaplasia (see above). Preoperative chemotherapy provides an opportunity to test in vivo the chemosensitivity of individual tumours. The new SIOP classification is based on the observation that survival or differentiation of distinct cell types is associated with different clinical outcomes. Tumours in which a substantial proportion of blastema survives preoperative chemotherapy have a poorer outcome than other subtypes.15 Conversely, tumours in which epithelial or stromal components predominate after preoperative chemotherapy appear to have a very favourable outcome. They are not yet selected for treatment reduction, but this question is being addressed in the SIOP WT 2001 trial. Complete necrosis following preoperative chemotherapy has an excellent outcome.17

Molecular prognostic factors are being identified and will be incorporated into treatment strategies in the near future. Loss of heterozygosity (LOH) of markers at chromosome 16q and 1p was associated with a significantly poorer relapse-free survival and overall survival in a retrospective study of 232 children registered on NWTS-3–4.29 Other retrospective studies have confirmed the adverse prognostic prediction for tumours with LOH on 16q.30 The NWTS-5 trial was set up to test prospectively the hypothesis that LOH at specific chromosomal loci can identify tumours with a worse outcome. The only other established adverse molecular factor is mutation of the p53 gene which occurs in the majority of anaplastic tumours and may explain their worse prognosis. Involvement of the WT1 gene, or indeed other genes at 11p15, does not appear to be associated with adverse outcome. The association of many other molecular alterations with clinical outcome is currently under investigation, including gain or overexpression of genes on 1q, telomerase activity, etc.

Another question being considered by the renal tumours committee of the Children's Oncology Group (which has taken over the role of the NWTS) is whether the absence of adverse clinical and molecular prognostic factors can be used to identify a group of stage I tumours curable by nephrectomy only.

Late effects

All treatment modalities are associated with toxic effects, some of which occur early, while others occur months or years later. Long-term deleterious effects of radiation in Wilms tumours patients have been reported in the musculoskeletal system, the gastrointestinal tract, the urinary tract, the endocrine system, and the lungs. In long-term follow-up studies of survivors of Wilms tumour, scoliosis and musculoskeletal abnormalities have been found more frequently in irradiated patients, particularly those treated with orthovoltage radiotherapy. With modern megavoltage techniques and the significant dose reductions now used, late orthopaedic effects are expected to be less frequent. Reduction in stature following megavoltage radiation therapy of the spine is dose and age dependent, and is usually clinically insignificant. Adverse pregnancy outcomes have been reported in Wilms tumour survivors treated with abdominal radiation, with increased rates of perinatal mortality, low birth weight, and congenital malformation.31 Exposure to doxorubicin and thoracic and left-flank irradiation are risk factors for cardiac toxicity.32

Recurrence of Wilms tumour

Overall relapse rates for all patients with Wilms tumour registered on NWTS and SIOP studies is 17–24 per cent. The majority of relapses occur within 2 years of diagnosis, although with the continued reductions in therapy occurring in most studies, it is possible that this pattern may change in the future. Despite the excellent prognosis at first diagnosis of Wilms tumour, long-term survival is 30 per cent for most patients with recurrent disease. Prognosis following relapse depends on initial stage, site of relapse, interval from initial diagnosis, and previous therapy. The most favourable prognostic group comprises children treated for low-stage tumours with no more than two drugs and without radiotherapy, who experience isolated pulmonary relapse >6–12 months after diagnosis. All other patients have a poor outcome and a high risk of treatment failure. This has led to the investigation of the role of ifosfamide, etoposide, and platinum compounds as single agents or in combination. These studies have demonstrated response rates >40 per cent. However, the results are transitory and the outcome has continued to be poor. The use of ICE (ifosfamide, carboplatin, etoposide) has shown an overall response rate (partial plus complete) of 82 per cent and a 3-year event-free survival and overall survival of 63.6 ± 14.5 per cent.33 High-dose chemotherapy followed by autologous stem cell rescue has been used for the treatment of patients with high-risk recurrent Wilms tumour, and preliminary results are encouraging. Pein et al34 recently reported on 29 patients with high-risk recurrent Wilms tumour who received treatment with high-dose chemotherapy followed by autologous stem cell rescue. Despite significant treatment related toxicity, the disease-free survival and overall survival at 3 years were 50 ± 17 per cent and 60 ± 18 per cent, respectively, an improvement compared with the historical controls.34 An international randomized trial is being designed to assess the value of consolidation with high dose chemotherapy after intensive re-induction.

Bilateral tumours

Between 5 and 10 per cent of Wilms tumour patients have bilateral renal lesions at diagnosis, usually of low stage individually. Multidisciplinary advances in surgery, chemotherapy, and radiotherapy have markedly improved the expected outcome for these children. The median age at diagnosis is lower than that of patients with unilateral disease and the incidence of congenital anomalies is higher. However, the heterogeneity of conditions associated with bilateral Wilms tumour, as well as the bimodal peak on the age-at-onset distribution curve, suggests more than one causative factor. The management of bilateral Wilms tumour has changed with time, especially with regard to the surgical procedure. Bilateral renal salvage procedures are technically possible and effective in controlling tumours without adversely affecting renal function or survival. Radical excision of the tumour should never be performed at the initial operation. Preoperative chemotherapy is always recommended to reduce the tumour burden and to permit bilateral renal preservation procedures, if feasible. Functional assessment of both kidneys should always be undertaken prior to any planned conservative procedure. Delayed secondor even third-look surgery to attempt tumour resection with maximum preservation of functioning renal tissue may be of benefit. The primary objective of bilateral conservative surgery is to decrease the incidence of renal failure. However, there is a fine balance between preserving nephrons and achieving adequate excision to avoid relapse, as the treatment of recurrent disease has far more deleterious consequences for renal function. Bilateral Wilms tumour patients should be treated individually. Partial nephrectomy, tumour enucleation, radiotherapy, and chemotherapy should be adapted according to response, highest stage, and highest-risk histology. Overall survival rates exceed 80 per cent at 4 years and complete nephrectomy is avoided in 60–70 per cent of the children.35,36

Bilateral Wilms tumour is frequently associated with nephrogenic rests, which are presumed to have the potential for malignant conversion. The role of chemotherapy in influencing this progression is unclear. However, it is currently suggested that prolonged treatment with vincristine and dactinomycin for up to 1 year may be of benefit.


Certain congenital malformation syndromes carry a high risk of Wilms tumour, of the order of 30–50 per cent. With the discovery of the underlying genetic defects, in some cases subgroups can be defined for screening. For example, in sporadic aniridia, high-resolution karyotyping using probes for the contiguous PAX6calmodulin and WT1 genes on 11p13 can distinguish those children whose aniridia is due to mutation confined to PAX6 and therefore do not require screening from those with a more extensive deletion involving the WT1 gene. Similarly, children with early-onset nephrotic syndrome involving diffuse mesangial sclerosis, even without ambiguous genitalia, are likely to harbour a constitutional WT1 mutation and hence carry an elevated risk of Wilms tumour. In Beckwith–Wiedemann syndrome, it appears that the children with uniparental disomy for 11p15.5 or imprinting abnormalities of the IGF2 and H19 genes are at higher tumour risk than those with other types of causative genetic lesion, such as imprinting abnormalities of LIT1 or point mutations in the p57 gene. A higher tumour risk may also be related to clinical features including the presence of hemi-hypertrophy or nephromegaly during the first year of life. Where the risk of Wilms tumour is of the order of 10–30 per cent, it is generally held that some sort of screening programme is justified. It has not been possible to perform randomized studies in this setting. However, retrospective analyses of the tumour stage at diagnosis in relation to the mode of discovery suggest that if regular ultrasound screening is to be used, it should be performed at intervals of no more than 3–4 months. It is possible to teach parents to perform regular abdominal palpation, particularly at bathtime, but there is no evidence as to whether this is equivalent to ultrasound screening. The risk of Wilms tumour in children with hemi-hypertrophy without other stigmata of Beckwith–Wiedemann syndrome or isolated overgrowth of a single limb is not easily defined, but it does appear to be at a much lower level that raises questions about the justification of an interventional imaging screening programme. Children whose Wilms tumour predisposition results from mutation in the WT1 gene have a much earlier age of onset of Wilms tumour than those with Beckwith–Wiedemann syndrome (median age at diagnosis 16 months versus 39 months).37 It is now also emerging that children whose tumorous kidney contains either multifocal tumours or nephrogenic rests are at increased risk of late development of metachronous Wilms tumour. As a safe policy, it is now recommended that all children with these features are screened for a prolonged period of time, at intervals of 3 months until the age of 7 years. These recommendations may become more lenient as the evidence base becomes firmer.

Long-term follow-up

Children who survive relapse free for >3 years after diagnosis of their Wilms tumour are unlikely to suffer a recurrence. Treatment received by the majority means they are at very low risk of developing second cancers related to their treatment. Therefore the main aim of longterm follow up is to monitor renal function. The current UKCCSG recommendations for any child having a nephrectomy are that blood pressure should be checked annually and serum creatinine measured every 5 years. An early morning urine sample should be tested annually for protein-to-creatinine ratio. The risk of renal dysfunction continues into adult life, and it is important that such information is imparted to their adult physicians or general practitioners. Analysis of the molecular genetics of Wilms tumour has provided and continues to provide a fascinating insight into the relationship between developmental abnormalities and embryonal cancers. Identification of genes involved in such processes and their impact on tumour biology may ultimately allow the safe selection of subgroups of children with Wilms tumour requiring only minimal therapy and to modify front-line therapy for those groups with a poorer prognosis. The avoidance of anthracyclines and radiotherapy for an increasing majority would be a major step forward in the successful treatment of this tumour type. The new SIOP WT 2001 trial addresses this issue.


1. Stiller CA, Parkin DM (1990). International variations in the incidence of childhood renal tumours. Br J Cancer 62, 1026–30.

2. Breslow N, Olshan A, Beckwith JB, Green DM (1993). Epidemiology of Wilms tumor. Med Pediatr Oncol 21, 172–81.

3. Franco EL, de Camargo B, Saba L, Marques LA (1991). Epidemiological and clinical correlations with genetic characteristics of Wilms' tumor: results of the Brazilian Wilms' tumor Study Group. Int J Cancer 48, 641–6.

4. Clericuzio CL (1993). Clinical phenotypes and Wilms tumor. Med Pediatr Oncol 21, 182–7.

5. Knudson A, Strong L (1972). Mutation and cancer, a model for Wilms tumour of the kidney. J Natl Cancer Inst 48, 313–24.

6. Rapley EA, Barfoot R, Bonaiti-Pellie C, et al. (2000) Evidence for susceptibility genes to familial Wilms tumour in addition to WT1FWT1 and FWT2Br J Cancer 83, 177–83.

7. Diller L, Ghahremani M, Morgan J, et al. (1998). Constitutional WT1 mutations in Wilms' tumor patients. J Clin Oncol 16, 3634–40.

8. Shumacher V, Schneider S, Figg A, et al. (1997). Correlation of germ-line mutations and two-hit inactivation WT1 gene with Wilms' tumor of stromal-predominant histology. Proc Natl Acad Sci USA94, 3972–7.

9. Bardeesy N, Falkoff D, Petruzzi M-J, et al. (1994). Anaplastic Wilms' tumour, a subtype displaying poor prognosis, harbours p53 gene mutations. Nat Genet 7, 91–7.

10. Olshan AF, Breslow NE, Falletta JM, et al. (1993). Risk factors for Wilms tumor. Cancer 72, 938–44.

11. Sharpe CR, Franco EL, de Camargo B, et al. (1995). Parental exposures to pesticides and risk of Wilms tumor in Brazil. Am J Epidemiol 141, 210–17.

12. Vujanic GM, Kelsey A, Mitchell C, Shannon RS, Gornall P (2003). The role of biopsy in the diagnosis of renal tumors of childhood, results of the UKCCSG Wilms tumor study 3. Med Pediatr Oncol 40, 18–22.

13. Owens CM, Veys PA, Pritchard J, Levitt G, Imeson J, Dicks-Mireaux C (2002). Role of chest computed tomography at diagnosis in the management of Wilms' tumor: a study by the United Kingdom Children's Cancer Study Group. J Clin Oncol 20, 2768–73.

14. Beckwith JB, Zuppan CE, Browning NG, Moksness J, Breslow NE (1996). Histological analysis of aggressiveness and responsiveness in Wilms' tumor. Med Pediatr Oncol 27, 422–8.

15. Weirich A, Leuschner I, Harms D, et al. (2001) Clinical impact of histologic subtypes in localized non-anaplastic nephroblastoma treated according to the trial and study SIOP-9/GPOH. Ann Oncol12, 311–19.

16. Vujanic GM, Sandstedt B, Harms D, et al. (2002). Revised International Society of Paediatric Oncology (SIOP) Working Classification of Renal Tumors of Childhood. Med Pediatr Oncol 38, 79–82.

17. Boccon-Gibod L, Rey A, Sandstedt B, et al. (2000). Complete necrosis induced by preoperative chemotherapy in Wilms tumor as an indicator of low risk, report of the International Society of Paediatric Oncology (SIOP) Nephroblastoma Trial and Study 9. Med Pediatr Oncol 34, 183–90.

18. Faria P, Beckwith JB, Mishra K, et al. (1996). Focal versus diffuse anaplasia in Wilms tumor—new definitions with prognostic significance. A report from the National Wilms Tumor Study Group. Am J Surg Pathol 20, 909–20.

19. Beckwith JB (1998). Nephrogenic rests and the pathogenesis of Wilms tumor: developmental and clinical considerations. Am J Med Genet 79, 268–73.

20. Green DM, Breslow N, Beckwith JB, et al. (1998).Comparison between single-dose and divided-dose administration of dactinomycin and doxorubicin for patients with Wilms' tumor: a report from the National Wilms' Tumor Study Group. J Clin Oncol 16, 237–45.

21. Green DM, Breslow NE, Beckwith JB, et al. (1998). Effect of duration of treatment on treatment outcome and cost of treatment for Wilms' tumor: a report from the National Wilms' Tumor Study Group. J Clin Oncol 16, 3744–51.

22. Tournade MF, Com-Nougue C, Voûte PA, et al. (1993). Results of the Sixth International Society of Pediatric Oncology Wilms Tumor Trial and Study: a risk-adapted therapy approach in Wilms tumor. J Clin Oncol 11, 1014–23.

23. Mitchell C, Jones PM, Kelsey A, et al. (2000). The treatment of Wilms' tumour: results of the United Kingdom Children's Cancer Study Group (UKCCSG) second Wilms' tumour study. Br J Cancer 83, 602–8.

24. De Camargo B, Franco EL for the Brazilian Wilms Tumor Study Group (1994). A randomized clinical trial of single-dose versus fractionated-dose fractionated dactinomycin in the treatment of Wilms tumor. Results after extended follow up. Cancer 73, 3081–6.

25. Green DM, Norkool P, Breslow NE, Finklestein JZ, D'Angio GJ (1990). Severe toxicity after treatment with vincristine and dactinomycin using single-dose or divided-dose schedules: a report from the National Wilms Tumor Study. J Clin Oncol 8, 1525–30.

26. Othersen HB Jr, DeLorimer A, Hrabovsky E, Kelalis P, Breslow N, D'Angio GJ (1990). Surgical evaluation of lymph node metastases in Wilms' tumor. J Pediatr Surg 25, 330–1.

27. Kessler O, Franco I, Jayabose S, Redá E, Levitt S, Brock W (1996). Is contralateral exploration of the kidney necessary in patients with Wilms tumor. J Urol 156, 693–5.

28. Pachnis A, Pritchard J, Gaze M, Levitt G, Michalski A (1998). Radiotherapy omitted in the treatment of selected children under 3 years of age with stage III favorable histology Wilms tumor. Med Pediatr Oncol 31, 150–2.

29. Grundy PE, Telzerow PE, Breslow N, Mokness J, Huff V, Paterson MC (1994). Loss of heterozygosity for chromosome 16q and 1p in Wilms tumor predicts an adverse outcome. Cancer Res 94, 2331–3.

30. Grundy RG, Pritchard J, Scambler P, Cowell JK (1998). Loss of heterozygosity on chromosome 16 in sporadic Wilms' tumour. Br J Cancer 78, 1181–7.

31. Green DM, Peabody EM, Nan B, Peterson S, Kalapurakal JA, Breslow N (2002). Pregnancy outcome after treatment for Wilms tumor: a report from the National Wilms Tumor Study Group. J Clin Oncol 20, 2506–13.

32. Green DM, Grigoriev Y, Nan B, et al. (2001). Congestive heart failure after treatment for Wilms tumor: a report from the National Wilms Tumor Study Group. J Clin Oncol 19, 1926–34.

33. Abu-Ghosh AM, Krailo MD, Goldman SC, et al. (2002). Ifosfamide, carboplatin and etoposide in children with poor-risk relapsed Wilms tumor: a Childrens Cancer Group Report. Ann Oncol 13, 460–9.

34. Pein F, Michon J, Valteau-Couanet D, et al. (1998). High-dose melphalan, etoposide, and carboplatin followed by autologous stem-cell rescue in pediatric high-risk recurrent Wilms' tumor: a French Society of Pediatric Oncology study. J Clin Oncol 16, 3295–301.

35. Cooper CS, Jaffe WI, Huff DS, et al. (2000). The role of renal salvage procedures for bilateral Wilms tumor: a 15-year-review. J Urol 163, 265–8.

36. Horwitz JR, Ritchey ML, Mokness J, et al. (1996). Renal salvage procedures in patients with synchronous bilateral Wilms tumor: a report from the National Wilms Tumor Study Group. J Pediatr Surg31, 1020–5.

37. Beckwith JB (1998) Nephrogenic rests and the pathogenesis of Wilms' tumour: development and clinical considerations. Am J Med Genet 79, 268–73.

If you find an error or have any questions, please email us at Thank you!