Corrected: Publisher Correction COnSEnSUS

Statement

P O S I T I O N PA P E R

The UMBRELLA SIOP–RTSG 2016

Wilms tumour pathology and
molecular biology protocol
Gordan M. Vujanić1,12*, Manfred Gessler   2,3,12, Ariadne H. A. G. Ooms4, Paola Collini5,
Aurore Coulomb-​l’Hermine6, Ellen D’Hooghe7, Ronald R. de Krijger4, Daniela Perotti8,
Kathy Pritchard-​Jones   9, Christian Vokuhl10, Marry M. van den Heuvel-​Eibrink4 and
Norbert Graf   11, on behalf of the International Society of Paediatric Oncology–Renal
Tumour Study Group (SIOP–RTSG)
Abstract | On the basis of the results of previous national and international trials and studies, the
Renal Tumour Study Group of the International Society of Paediatric Oncology (SIOP–RTSG) has
developed a new study protocol for paediatric renal tumours: the UMBRELL A SIOP–RTSG 2016
protocol (the UMBRELL A protocol). Currently , the overall outcomes of patients with Wilms
tumour are excellent, but subgroups with poor prognosis and increased relapse rates still exist.
The identification of these subgroups is of utmost importance to improve treatment stratification,
which might lead to reduction of the direct and late effects of chemotherapy. The UMBRELL A
protocol aims to validate new prognostic factors, such as blastemal tumour volume and molecular
markers, to further improve outcome. To achieve this aim, large, international, high-quality
databases are needed, which dictate optimization and international harmonization of specimen
handling and comprehensive sampling of biological material, refine definitions and improve
logistics for expert review. To promote broad implementation of the UMBRELL A protocol, the
updated SIOP–RTSG pathology and molecular biology protocol for Wilms tumours has been
outlined, which is a consensus from the SIOP–RTSG pathology panel.

Paediatric renal tumours account for ~6% of all paediatric renal tumours. Molecular biology research
paediatric malignant tumours, of which ~90% are within the protocol is primarily focused on the vali-
Wilms  tumours (nephroblastoma) 1. Other renal dation of the prognostic value of 1q gain, which might
non-​Wilms tumours are rare entities and include meso­ lead to a more personalized treatment approach (Box 1).
blastic nephroma, clear cell sarcoma of the kidney, rhab- Consequently, short-​term and long-​term outcomes
doid tumour of the kidney, renal cell carcinoma and might be improved for all children with renal tumours
few other, even rarer tumour types1. The Renal Tumour by increasing survival, but also by reducing treatment
Study Group of the International Society of Paediatric in specific subgroups, resulting in diminished direct
Oncology (SIOP–RTSG) has developed a new study for and late adverse effects. Timely genetic analysis and
all renal tumours of childhood, the UMBRELLA SIOP– step-​wise extension to additional targets such as TP53
RTSG 2016 protocol (the UMBRELLA protocol)2,3, on (refs5–7) or several of the newly identified driver candi-
the basis of the experiences from SIOP–2001 and the UK dates for stratification and inclusion of liquid biopsies
Improving Population Outcomes for Renal Tumours of might help to reach this goal8–10.
Childhood (IMPORT) 2013 studies2–4. The aim of the The UMBRELLA protocol includes updated guide-
UMBRELLA protocol is to collect all clinical, biolo­ lines for pathologists for the handling and processing of
gical and outcomes data from children with primary tissue as well as criteria that are important for postop-
renal tumours in a comprehensive data registry, with erative histological classification, staging and treatment
*e-​mail: gvujanic@sidra.org central review of diagnostics (radiology, patho­logy stratification. These recommendations were established
https://doi.org/10.1038/ and surgery), standardized biobanking and precise by a consensus of pathology experts within the SIOP–
s41585-018-0100-3 treatment recommendations for the most common RTSG (chaired by G. M. Vujanić and I. Leuschner).

NATuRe RevIeWS | UROLOGy volume 15 | NOVEMBER 2018 | 693
C o n S e n S u S S tat e m e n t

Box 1 | Relevant aims of the SIOP–RTSG UMBRELLA 2016 protocol feedback on institutional pathological diagnosis in a
clinically relevant time frame. The UMBRELLA protocol
Pathology aims addresses both Wilms tumour and non-​Wilms tumours;
To optimize pathological diagnostics by rapid central review in order to: in this Position Paper, we focus on the recommendations
• Classify and stage tumours accurately in order to optimize treatment and avoid for Wilms tumours.
undertreatment or overtreatment.
• Monitor and give appropriate feedback on local pathological diagnosis and stage. Guidelines for specimen handling
• Accurately assess the blastema in postchemotherapy specimens. As in previous SIOP trials and studies14, the UMBRELLA
• Correlate blastema volume and biomarkers with event-​free and overall survival. protocol mandates preoperative chemotherapy on the
basis of clinical and radiological diagnoses in most patients.
Biobanking aims
Thus, pathological diagnosis is mainly based on pre-
• To store biological material from all participating centres. treated nephrectomy specimens. Pretreatment biopsy is
• To show the feasibility of storing serial blood and urine samples and tumour and not routinely recommended. Percutaneous cutting needle
germline material at diagnosis and at specific time points during treatment for biopsy (tru-​cut biopsy) can be considered in young chil-
international collaborative studies.
dren with stage IV disease and in children >10 years of
• To establish national biobanks (if not yet present) and generate a study-​wide virtual age, as the frequency of non-​Wilms tumours (rhabdoid
biobank for translational research.
tumour of the kidney and renal cell carcinoma, respec-
• To test the feasibility of returning biomarker results to treatment centres within a tively) is increased in these age groups16. Biopsies are parti­
clinically relevant time frame.
cularly not advised in children <6 months of age and in
Molecular biology aims fully cystic tumours; immediate surgery is recommended
• To assess genomic 1q gain as a prognostic marker in Wilms tumour. in these children. If performed, open wedge biopsy auto-
• To analyse biomarkers with potential prognostic relevance (simultaneous loss of 1p matically upstages tumours to stage III, irrespective of
and 16q, MYCN gain and 17p loss encompassing the TP53 locus). other findings, as it is regarded as an (artificial) breach
• To assess molecular characteristics of blastemal-​type Wilms tumours. of the capsule (rupture). However, percutaneous cutting
• To explore whether aberrations in WT1, CTNNB1, AMER1, TP53, MYCN, FBXW7, GPC3,
needle biopsies do not upstage tumour.
MLLT1, DIS3L2, DICER1, DROSHA, DGCR8, SIX1, or SIX2 considerably affect event-​
free or overall survival. Institutional pathologist’s role. The institutional patho­
• To explore circulating biomarkers (microRNA and DNA) in blood, and/or plasma, logist samples and processes the fresh tumour according
and/or urine and to support research on epigenetics, tumour cell culture, and/or to the protocol. The surgical specimen should be worked
organoids and xenografts. up without delay to minimize degradation, especially of
SIOP–RTSG, Renal Tumour Study Group of the International Society of Paediatric RNA. After visual inspection, weighing, measuring and
Oncology. photographing, the entire surface is inked. The speci­
men is bisected longitudinally, and the macroscopic
appearance — specifically the percentage of necrotic
As well as the aim of validating 1q gain as a prognostic areas — is recorded and photographed. Samples for
marker11,12, one of the main focuses of the UMBRELLA biological studies are taken from viable and, if present,
protocol is assessment of the independent prognostic macroscopically distinct areas. The whole specimen is
value of the absolute volume (rather than the percent- fixed in formalin for 24–48 hours and, then, at least one
age) of the blastemal component that survives preop- complete longitudinal slice of the whole specimen is sam-
erative chemotherapy13–15. Furthermore, in order to pled for histological evaluation to capture tumour hetero-
optimize pathological diagnostics and treatment, rapid geneity17. Most importantly, a guide block of the selected
central pathology review will be obligatory to enable samples from the tumours is made by the pathologist to
assist tumour staging17 (Supplementary Figure 1).
Author addresses
SIOP histological classification
1
Department of Pathology, Sidra Medicine, Doha, Qatar. Wilms tumour is a histologically heterogeneous embry-
2
Theodor-​Boveri-Institute/Biocenter, Developmental Biochemistry, Wuerzburg
onal tumour composed of blastemal, epithelial and
University, Wuerzburg, Germany.
3
Comprehensive Cancer Center Mainfranken, Wuerzburg University,
stromal components17. In Europe, patients diagnosed
Wuerzburg, Germany. with Wilms tumour are treated with chemotherapy for
4
Princess Maxima Centre for Pediatric Oncology, Utrecht, Netherlands. 4 or 6 weeks before surgery, depending on metastatic
5
Department of Diagnostic Pathology and Laboratory Medicine, Fondazione IRCCS status14. Postoperatively, treatment stratification relies
Istituto Nazionale dei Tumori, Milan, Italy. on overall and local stage, and also on histological clas-
6
Sorbonne Université, Department of Pathology, Hopitaux Universitaires Est Parisien, sification into low-​risk, intermediate-​risk and high-​risk
Paris, France. Wilms tumours18. As nearly 40% of all relapses occur in
7
Department of Pathology, Oslo University Hospital, Rikshospitalet, Oslo, Norway. children whose tumour was not histologically classified
8
Molecular Bases of Genetic Risk and Genetic Testing Unit, Department of Preventive and as high-​risk Wilms tumour, a need to improve treatment
Predictive Medicine, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy.
stratification is evident2,19,20.
9
Great Ormond Street Institute of Child Health, University College London, London, UK.
10
Kiel Paediatric Tumour Registry, Department of Paediatric Pathology, University
Histological assignment to the different risk groups
Hospital of Kiel, Kiel, Germany. is based on quantification of chemotherapy-​induced
11
Department of Paediatric Oncology & Haematology, Saarland University, changes, the percentages of the different viable Wilms
Homburg, Germany. tumour components (epithelial, blastemal and stromal)
12
These authors contributed equally: Gordan M. Vujanić, Manfred Gessler. and the presence or absence of anaplasia17,18 (Table 1).

694 | NOVEMBER 2018 | volume 15 www.nature.com/nrurol

 C o n S e n S u S S tat e m e n t

Table 1 | Histological criteria for Wilms tumour subtyping in SIOP pretreated patients
Tumour typea Chemotherapy-​induced change Histological features (% of viable tumour)
Blastema Epithelium Stroma
Completely necrotic 100 0 0 0
Regressive >66 0–100 0–100 0–100
Mixed <66 0–65 0–65 0–65
Mixed <66 11–65 0–89 0–89
Epithelial <66 0–10 66–100 0–33
Stromal <66 0–10 0–33 66–100
Blastemal <66 66–100 0–33 0–33
SIOP, International Society of Paediatric Oncology. aThe presence of diffuse anaplasia in any of the tumour types supersedes the
subtypes; focal anaplasia also needs to be specifically mentioned.

Histological classification after preoperative chemo- prognosis and is, therefore, classified as a high-​risk
therapy defines three major subgroups of Wilms tumour according to the SIOP–2001 working classifi-
tumours: low-​risk (completely necrotic Wilms tumour), cation18. These blastemal-​type Wilms tumours were
high-​risk (blastemal type and diffuse anaplasia) and treated with more intensive postoperative treatment than
intermediate-​risk tumours (all other types)17,18. To cor- other tumour types in the SIOP–2001 trial, resulting in
rectly subclassify the Wilms tumour, the percentages improved outcomes15. This improvement highlights
of chemotherapy-​induced changes and viable tumour the need to identify high-​risk, chemotherapy-​resistant
components are assessed and taken into account as per blastema to avoid unnecessary failure of first-​line ther-
the histological classification criteria17,18 (Table 1). The apy, especially as doxorubicin is omitted in treatment of
low-​risk group includes Wilms tumours that become stage II and stage III intermediate-​risk tumours19. A ret-
completely necrotic owing to preoperative treatment. rospective analysis of the SIOP–2001 trial data showed
Tumours in other risk groups are subclassified on the that a threshold of ~20 ml residual blastemal volume in
basis of viable tumour components. The intermediate-​ localized Wilms tumour could be used to improve strat-
risk group includes the epithelial-​type, stromal-​type, ification13,14. An accurate recognition of blastema needs
mixed-​type and regressive-​type tumours and Wilms more objective criteria than are currently available to
tumours with focal anaplasia (Supplementary Box 2). minimize interobserver variation and to enable insti-
In the intermediate-​risk group, in addition to histo- tutional pathologists to make a correct diagnosis. The
logical subclassification, the tumour volume after pre- SIOP–RTSG agreed that the blastemal volume threshold
operative chemotherapy (measured using imaging) at which the risk of relapse rises dramatically needs to
is of importance for treatment stratification. If the be established before its implementation as a prognostic
tumour volume is >500 ml in stage II/III mixed-​type, factor. Thus, one of the aims of the UMBRELLA protocol
regressive-​type, or focal anaplasia-​type tumours, these is to optimize the definition of high-​risk, blastemal-​type
tumours are considered to have an increased risk of Wilms tumour.
poor outcome and are treated aggressively3. Some 5% In the UMBRELLA protocol, institutional pathol-
of Wilms tumours are multifocal (both unilateral and ogists will not be required to calculate the volume of
bilateral tumours), and they are more difficult to assess. blastema in tumours, as it is not currently a criterion for
Currently, treatment of these tumours depends on other treatment stratification. However, the estimated percent-
parameters, such as their histological type and response ages of the nonviable and viable components (especially
to preoperative chemotherapy, and the best way for a blastema) will be recorded, and blastemal volume will
pathologist to manage them is to adopt two methods of then be calculated in the protocol database. This calcu-
analysis: one method is to assess all nodules together and lation will be based on both pathological and imaging
calculate the percentages of nonviable and viable com- measurements of the tumour.
ponents as if they represented one nodule, and the other
method is to assess them individually (also commenting Anaplasia. Anaplasia has been recognized as a high-​risk
on their size). The treatment decision should be made Wilms tumour feature for many years, and it confers a
at multidisciplinary team meetings and in consultation worse prognosis14,18,21–24. Anaplasia can occur in any
with national or international lead experts. component of Wilms tumours, and it can be focal or
diffuse21,22. Importantly, in order to diagnose anaplasia,
Blastemal type and blastemal volume. The most all three criteria need to be present: large atypical tri­
important viable component to recognize in pretreated polar and/or multipolar mitotic figures; marked nuclear
Wilms tumours is blastema, which is the most undif- enlargement, with nuclear diameters at least three times
ferentiated tumour component and is composed of those of adjacent cells; and hyperchromatic tumour cell
primitive, undifferentiated cells that are arranged in nuclei21,22. Although the criteria for anaplasia are well
no particular pattern. Blastemal-​type Wilms tumour established, in many circumstances diagnosis can be
(in which >66% of the viable tumour consists of blas- challenging, resulting in a considerable number of undi-
tema in a tumour that is >33% viable) confers a worse agnosed or overdiagnosed patients25. This misdiagnosis

NATuRe RevIeWS | UROLOGy volume 15 | NOVEMBER 2018 | 695
C o n S e n S u S S tat e m e n t

illustrates the value of rapid central pathology review for Histological diagnosis of cystic partially differenti-
correct postoperative treatment stratification. ated nephroblastoma. Cystic partially differentiated
Focal anaplasia is defined as the presence of one or nephroblastoma is a distinct variant of Wilms tumour
two foci showing the abovementioned nuclear criteria that usually occurs in children <3 years of age and is
with sharp demarcation within the primary intrarenal composed entirely of cysts with septa without tumour
tumour and without evidence of anaplasia or promi- nodules inside them34. Molecular studies have shown
nent nuclear atypia (nuclear unrest) in other areas22. that cystic nephroma and cystic partially differentiated
By consensus of the SIOP–RTSG pathology panel, the nephroblastoma are unrelated tumours35. Intermediate-​
size of the anaplastic focus does not exceed 15 mm risk or high-​risk Wilms tumours can also present with
(which was previously undefined). Wilms tumour numerous cysts, especially after preoperative chemo-
with focal anaplasia is regarded as an intermediate-​risk therapy, but they also contain solid, nodular areas34.
tumour in the UMBRELLA protocol, but the tumour The presence of numerous cysts in these tumours is of
should still be subclassified according to other compo- no prognostic significance34.
nents (for example, focal anaplasia in mixed-​type Wilms
tumour). If focal anaplasia is present in blastemal-​type Histological classification of primarily operated
Wilms tumour, it is regarded as a high-​risk tumour. Wilms tumours. In rare instances, for example in
Diffuse anaplasia is defined as nonlocalized or multi­ children <6 months of age, immediate surgery is reco­
focal anaplasia, focal anaplasia with marked nuclear mmended in the UMBRELLA protocol, as opposed
unrest in the rest of the tumour, or anaplasia outside to preoperative chemotherapy. This course of action
of the kidney (anaplastic tumour in intrarenal or extra­ leads to a different histological classification, as
renal vessels, renal sinus, extracapsular sites or meta- chemotherapy-​induced changes are not present. Of
static deposits). If anaplasia is present in a biopsy sample note, the presence of substantial amounts of viable
or other incomplete tumour sample, the diagnosis of blastema is not of prognostic significance in primarily
diffuse anaplasia is warranted22. operated Wilms tumours14. Histological classification
after immediate nephrectomy is based only on the pres-
Histological classification of bilateral Wilms tumour. ence or absence of diffuse anaplasia: if absent, the Wilms
Synchronous bilateral Wilms tumours (stage V) occur tumour is classified as intermediate risk and if present,
in ~5–8% of patients, and these children are more the tumour is high risk14,17.
likely to have an underlying genetic predisposition26.
These patients are treated with preoperative chemo- Histological staging
therapy for 6–12 weeks and nephron-​sparing surgery At diagnosis, the tumour is staged as localized (stages I–III),
(NSS) is considered by the surgical panel, taking into metastatic (stage IV), or bilateral (stage V) disease in
account tumour response to chemotherapy, to spare order to decide on duration of preoperative chemo­
as much renal function as possible. Each tumour is therapy 14. The histological staging and treatment
subclassified and staged separately according to the stratification of Wilms tumours after preoperative chemo-
histological classification and SIOP staging criteria. therapy has changed to some extent in the UMBRELLA
For multifocal tumours, the approach described above protocol on basis of the results of the previous tri-
should be used. als and studies and by consensus of the SIOP–RSTG
pathology panel3,15,19,23. More detailed definitions have
Histological classification of nephrogenic rests. been introduced to aid in correct staging by local
Nephrogenic rests are foci of embryonal cells that pathologists (Box 2).
persist after 36 weeks of gestation and are considered
precursors of Wilms tumours. Nephrogenic rests are Stage I tumours. Stage I tumours are confined within
identified in 35–40% of patients with unilateral and the (pseudo)capsule and, even if the tumours are out-
>90% of patients with bilateral Wilms tumours and are side of the kidney in the perirenal fat but surrounded
often associated with different syndromes or anoma- by the (pseudo)capsule and completely removed, they
lies such as overgrowth syndromes (hemihypertrophy are regarded as stage I. The vessels or soft tissues of the
and Beckwith–Wiedemann syndrome), Denys–Drash renal sinus are not involved. The presence of necrotic
syndrome and Wilms tumour, aniridia, genitourinary tumour or chemotherapy-​induced changes in the renal
abnormalities and mental retardation (WAGR) syn- sinus, renal veins and/or within the perirenal fat is not a
drome27–30. The two main types of nephrogenic rests reason for upstaging. Botryoid tumour growth into the
are perilobar and intralobar rests. The term nephro- renal pelvis is also not a reason for upstaging the tumour.
blastomatosis is used to describe multiple, multifocal, Furthermore, infiltration of the soft tissues between the
or diffuse nephrogenic rests27,28. Panlobar nephroblasto- kidney and the adrenal gland or of the adrenal gland itself
matosis encompasses complete replacement of the renal does not cause upstaging of the tumour if the external
lobe or cortex by nephrogenic rests and is recognized capsule of the adrenal gland is intact. However, the pres-
as a high-​risk factor for developing Wilms tumour31–33. ence of viable tumour in the lymphatic or blood vessels in
Prolonged chemotherapy is recommended in patients this area is regarded as stage II. Tumour adhesion to the
with nephroblastomatosis in the UMBRELLA protocol. liver capsule is not regarded as infiltration of an adjacent
Most patients are treated with NSS, as this therapy ena- organ; the tumour is upstaged (to stage II, if completely
bles disease-​free survival while sparing a maximum of resected, or to stage III, if incompletely resected) only
healthy parenchyma in both kidneys32,33. if clear infiltration of the liver parenchyma is present.

696 | NOVEMBER 2018 | volume 15 www.nature.com/nrurol

 C o n S e n S u S S tat e m e n t

Box 2 | Staging criteria in the SIOP UMBRELLA protocol or chemotherapy-​induced changes only at a resection
margin is not regarded as stage III. The finding of via-
Stage I ble or nonviable tumour thrombus bulging out at the
• Tumour is limited to the kidney. resection margin of the renal vein or inferior vena cava
• Tumour is present in the perirenal fat but is surrounded by a fibrous (pseudo)capsule. needs to be discussed with the surgeon. If the surgeon
The (pseudo)capsule might be infiltrated by viable tumour, which does not reach the confirms they resected the vein away from the throm-
outer surface. bus, then protruding thrombus at the vascular resection
• Tumour might show protruding (botryoid) growth into the renal pelvis or the ureter margin does not upstage the tumour to stage III. If the
but does not infiltrate their walls. thrombus is removed with force, owing to attachment to
• The vessels or the soft tissues of the renal sinus are not involved by tumour. Intrarenal the vascular wall, or by a piecemeal resection, stage III
vessel involvement might be present. must be considered at a multidisciplinary meeting.
Stage II The UMBRELLA protocol now explicitly states that the
presence of prechemotherapy tumour rupture at diag-
• Viable tumour is present in the perirenal fat and is not covered by a (pseudo)capsule,
but is completely resected (resection margins are clear). nosis on imaging studies is only considered as patho-
logical stage III if viable tumour is seen microscopically
• Viable tumour infiltrates the soft tissues of the renal sinus.
at the rupture site of the nephrectomy specimen. If not,
• Viable tumour infiltrates blood and/or lymphatic vessels of the renal sinus or of the
the tumour is staged on the basis of the other pathologi-
perirenal tissue, but it is completely resected.
cal criteria, but the final treatment stage must be decided
• Viable tumour infiltrates the wall of the renal pelvis or of the ureter.
after discussion at a multidisciplinary team and/or
• Viable tumour infiltrates the vena cava or adjacent organs (except the adrenal gland) tumour board meeting.
but is completely resected.
The presence of chemotherapy-​induced changes
Stage III (even without viable tumour) in a lymph node is
• Viable tumour is present at a resection margin. Nonviable tumour or chemotherapy-​ regarded as proof of tumour metastasis and, therefore,
induced changes present at a resection margin are not regarded as stage III. the tumour is assigned stage III. Mature tubules can be
• Abdominal lymph node involvement is present by either viable or nonviable tumour. found in lymph nodes, often associated with Tamm–
• Preoperative or intraoperative tumour rupture, if confirmed by microscopic Horsfall protein deposits, but this finding should not be
examination (viable tumour at the surface of the specimen at the area of the rupture). considered as lymph node metastasis36. Accumulation
• Viable or nonviable tumour thrombus is present at resection margins of ureter, renal of foamy macrophages within the subcapsular or inter-
vein, or vena cava inferior (always discuss resection margins with the surgeon). follicular sinus should not be regarded as a metastasis.
• Viable or nonviable tumour thrombus, which is attached to the inferior vena cava wall, Nonviable metastasis is regarded as replacement of nor-
is removed piecemeal by a surgeon. mal lymph node architecture with foamy macrophages
• Wedge or open tumour biopsy before preoperative chemotherapy or surgery. and/or chemotherapy-​induced changes.
The renal parenchyma should also be examined
• Tumour implants (viable or nonviable) are found anywhere in the abdomen.
for the presence of nephrogenic rests and for indica-
• Tumour (viable or nonviable) has penetrated through the peritoneal surface.
tions of an underlying predisposing syndrome37. For
Stage IV example, patients with Denys–Drash syndrome often
• Haematogenous metastases (for example, lung, liver, bone and brain) or lymph node demonstrate diffuse mesangial glomerulosclerosis,
metastases outside the abdominopelvic region. and Beckwith–Wiedemann syndrome is associated
with nephrolithiasis, medullary dysplasia and other
Stage V
renal abnormalities38.
• Bilateral renal tumours at diagnosis. Each side should be substaged according to the
above criteria.
Staging in nephron-​sparing surgery. Whenever NSS
is pursued, assessing the resection margins very care-
Tumour attachment to other retroperitoneal organs, fully is extremely important. Often, tumour nodules are
such as the colon, should be assessed in the same manner. resected with a small rim of renal parenchyma, especially
Fine needle aspiration or percutaneous cutting needle in patients who have multiple nodules within one kidney
biopsy (tru-​cut biopsy) does not upstage the tumour. or bilateral tumours. Procedures are encoded as NSS (A),
which is partial nephrectomy (resection of tumour with
Stage II tumours. Stage II tumours are defined by via- a rim of normal renal parenchyma), and NSS (B), which
ble tumour infiltration of the renal sinus (soft tissues is enucleation (resection of tumour without a rim of
and any vessels), and/or perirenal fat and/or adjacent normal renal parenchyma)39. Histopathological exam-
organs (except the adrenal gland) with no clearly identi- ination in NSS includes evaluation of the complete cir-
fiable (pseudo)capsule but with clear resection margins. cumference of the lesion. Small lesions are embedded
The presence of a viable tumour thrombus in the renal completely. The minimal distance of the Wilms tumour
vein or inferior vena cava is also stage II if completely to the resection margin will be measured. A safe rim has
resected in one piece. been defined as a margin of at least 1 mm, as is generally
accepted in tumour pathology. The finding of nephro-
Stage III tumours. Stage III tumours are defined accord- genic rests at the resection margin is not regarded as
ing to a number of criteria (Box  2). Several of these a positive resection margin and does not upstage the
criteria have undergone important changes in com- Wilms tumour.
parison with the SIOP–2001 trial criteria18. Now, in the Histopathological assessment clearly states one of
UMBRELLA protocol, the presence of nonviable tumour the following findings — safe rim of renal parenchyma

NATuRe RevIeWS | UROLOGy volume 15 | NOVEMBER 2018 | 697
C o n S e n S u S S tat e m e n t

Nephrectomy more countries than it was available to in SIOP–2001

will enable clinicians to modify treatment in a timely
manner, if required. Rapid central pathology review
Institutional pathologist undertakes biobanking and gross
should be undertaken before postoperative treatment
examination according to the UMBRELLA protocol and has started (which, according to the UMBRELLA
writes their report, which includes: treatment protocol, has to start 21 days after the last
• Histological subclassification dose of preoperative chemotherapy). All participat-
• Percentage of chemotherapy-induced changes
• Percentage of blastema ing centres will be required to send their samples to
• Presence or absence of diffuse anaplasia the dedicated review pathologist for central pathology
• Local stage
review. As >50 countries will be participating in the
UMBRELLA study, central pathology review will be
provided by national and/or regional pathology panels
Institutional pathologist sends the complete set of (Fig. 1; Supplementary Table 1).
histological slides without any delay for rapid central
pathology review to their designated chair of the national Additionally, all cases are peer reviewed by the
and/or regional pathology panel international SIOP–RTSG pathology panel. This
review is considered a quality control, as the individual
pathologists do not receive a report from this panel.
National or regional pathology panel chair promptly
performs rapid central pathology review and the result is Biobanking and biomarker analysis
sent back to the institutional pathologist (within 1 week)
Two of the primary aims of the UMBRELLA protocol
are biobanking and biomarker analysis (Box 1; Fig. 2).
The biobanking aim is to demonstrate the feasibil-
SIOP Pathology Panel reviews cases from national and/or
regional pathology panels retrospectively ity of storing serial blood and urine samples, as well
as tumour and germline material, at diagnosis and at
Fig. 1 | Flow diagram of rapid central pathology review. specific time points during treatment for international
After nephrectomy , the institutional pathologist biobanks collaborative studies. The molecular biology aim is to
the specimen and examines it according to the UMBRELL A validate the prognostic value of 1q gain and other copy
protocol, producing a report. The institutional pathologist number alterations (1p/16q, MYCN and TP53) that
then sends the complete set of slides for rapid central were identified in the previous SIOP–2001 trial and to
pathology review to their regional or national panel. The
demonstrate feasibility of testing in a clinically relevant
chair of the regional or national pathology panel promptly
undertakes central pathology review and sends back their
time frame5–7,11,12,40.
results to the institutional pathologist within 1 week. Achieving comprehensive collection of biomate-
The International Society of Paediatric Oncology (SIOP) rials in a uniform manner across the broad panel of
pathology panel reviews cases from the regional and participating countries will be challenging, but success
national pathology panels restrospectively. in this goal will be key for future large-​scale molecu-
lar profiling and implementation of biomarker testing.
To achieve this aim, a standard for biobanking and for
on resection margin (except nephrogenic rests), intact the submission of samples to the reference biological
(pseudo)capsule along the resection margin, or tumour laboratories has been developed. Virtually linked bio-
breach and/or rupture. banks will be created at a national level. These biobanks
will be instrumental to reach the aims of the study,
Rapid central pathology review including evaluation of novel genomic driver candidates,
As in the previous SIOP trials and studies, all renal informative microRNA (miRNA) patterns, epigenome
tumours and their assessment by the institutional analyses and panel sequencing that will be performed
pathologists will be reviewed by their national and/or in national laboratories. Individual approaches will also
regional pathology panel (Fig. 1). The results of pre- cover in vitro and in vivo models and other non-Wilms
vious Wilms tumour trials and studies have shown renal tumours.
considerable discrepancies in diagnosis and staging of
renal tumours between the institutional patho­logists Sample collection and storage. Sample collection
and central pathology review 25. In the SIOP–2001 requirements for the UMBRELLA protocol are divided
trial, rapid central pathology review was introduced in into an essential minimal set and optional extensions to
some parts of the trial, and it proved to be feasible and answer specific research questions. For baseline diagnos-
extremely valuable, avoiding undertreatment and over- tics, EDTA blood samples and tumour tissue together
treatment caused by misdiagnosis and/or mis-​staging by with healthy kidney tissue (if available) will be col-
institutional pathologists25. The SIOP–2001 study also lected from primary tumours and relapses and stored at
resulted in an increase in submissions for central patho­ –80 °C. Wherever possible, additional materials includ-
logy review from 76% (in the UKW3 trial) to 100% in ing blood samples for miRNA analysis and liquid biopsy
the UK part of the SIOP–2001 trial25. testing of circulating tumour DNA, as well as urine for
Discrepancies in diagnosis and staging between proteomic analyses, should be collected at multiple time
institutional pathologists and central pathology review points. This collection will facilitate parallel and future
are likely to remain, but introduction of rapid cen- translational research, which are detailed in the aims of
tral pathology review in the UMBRELLA protocol in the UMBRELLA protocol.

698 | NOVEMBER 2018 | volume 15 www.nature.com/nrurol

 C o n S e n S u S S tat e m e n t

Multiple tumour UK, Germany, the Netherlands, France and Italy. Sample
UMBRELLA protocol question:
sampling to capture
• 1q gain collection in this virtual biobank, including medical
heterogeneity Local collection
• Additional biomarkers records, informed consent for biobanking by parents
Sequential blood Transfer to and (if applicable) children, will be documented in the
samples (DNA and/or national web-​based SIOP UMBRELLA database.
biobanks • Ancilliary molecular analyses
RNA and liquid biopsy) • Extraction of subcohorts for The current proposal aims to further increase
Virtual study focused studies (such as inclusion of multiple samples, especially from hetero­
anaplastic Wilms tumour,
Urine samples biobank
mesoblastic nephroma and geneous Wilms tumours, and to enable additional
(liquid biopsy) rhabdoid tumour of the kidney) countries to provide the required samples and data.
Depending on each national coordinating centre,
Fig. 2 | Biobanking and biomarkers in the UMBRELL A protocol. Tissue, blood frozen blood and urine samples can be either directly
and urine samples are collected locally and included in national biobanks that are sent or stored on site until tissue samples are availa-
virtually connected. These materials are primarily used to answer UMBRELL A ble for batched collection and shipping to the national
protocol questions. Any surplus materials will be made available for ancillary and biobank. Participating countries that lack appropriate
focused studies.
infrastructure are encouraged to collect samples and
forward batches to larger, host biobanks that can pro-
Blood samples will be asked for at diagnosis, after vide the necessary storage capacity and expertise in
2 weeks of chemotherapy and before surgery. In addi- subsequent laboratory work-​up.
tion, collection of samples at follow-​up time points of
1 week after surgery and at the end of treatment will be Biomarker analysis. Children’s Oncology Group (COG)
requested. Separation of plasma and the cellular frac- protocols in the USA already include molecular mark-
tion by centrifugation within <2 hours after venepunc- ers in clinical decision making14. The results from our
ture is preferred to enable liquid biopsy analysis of previous SIOP studies are equally encouraging but not
circulating tumour DNA, as this process might provide yet definitive enough to mandate a change in clinical
an attractive option for upfront diagnostics or follow-​ approaches12,14. With the even larger community of par-
up monitoring. For miRNA studies, blood collected ticipating countries in the UMBRELLA protocol than
in PAXgene tubes, serum samples and urine samples in the SIOP–2001 study, we expect to collect sufficient
are needed. materials for studies to provide a definitive answer on
To guarantee correct histological staging, tumour whether or not we need to include certain biomarkers in
sampling is performed by the institutional patholo- the SIOP protocol as well. The two primary biomarkers
gist. Given the phenotypic and genetic heterogeneity of under study — blastemal volume and 1q gain — must
Wilms tumours, extensive sampling is needed41. Samples be considered in the context of any co-​dependencies
of ~1 cm3 each need to be selected from three or more with existing risk factors (tumour stage and histology)
spatially distinct sites within the tumour, as well as from and the treatment given to prove their additional value
macroscopically different lesions (if present) to capture for clinical decision making. On the basis of currently
potential heterogeneity. To make sure that representa- available data and agreed statistical parameters (power
tive and viable tissue has been selected in all samples, of 80% and a significance threshold of α = 0.05), up to
frozen sections or touch imprints can be prepared for 380 (for blastema volume) or 850 (for 1q gain) sam-
evaluation of cellular content, viability and proportion of ples will be needed to test an association with outcome.
tumour cells. Besides tumour tissue, collecting matched This analysis entails a prospective duration of 5 years to
normal tissue for comparative analyses from all patients collect the samples. In the case of 1q gain, testing will
is advised. be performed on frozen and FFPE material to increase
Snap freezing of tissue in liquid nitrogen must be numbers, as several of the participating countries
done as soon as possible after excision to ensure mini- will not be able to guarantee successful provision of
mal sample degradation. Duration from resection until high-​quality frozen samples.
freezing of the samples will be recorded for quality con- No final decision has yet been made on the best
trol. Importantly, all surgically removed metastases or approach to test for 1q gain and other copy number
relapses must be collected using the same protocol to alterations. Multiplex ligation-​dependent probe ampli-
facilitate identification of drivers of therapy resistance fication analysis proved to be cost-​effective and reliable
and metastatic potential. Biobanking of frozen mat­ in SIOP–2001 (ref.12), but the number of targets is lim-
erials will be complemented by surplus formalin-​fixed, ited. Single nucleotide polymorphism arrays or focused
paraffin-​embedded (FFPE) blocks that are also a val- next-​generation sequencing (NGS) panels might pro-
uable resource for research, including the molecular vide improved resolution to identify critical subregions,
analysis of microdissected nephrogenic rests, tumour including copy number neutral allele loss.
subcompartments and anaplastic foci. Several additional genomic copy number biomarkers
Clinical data and biological samples will be collected beyond 1q gain have potential prognostic relevance in
on a national basis in the paediatric oncology centres Wilms tumour (Box 1). Simultaneous allele loss of 1p
where the children are treated. To achieve inclusion of and 16q is associated with poor outcome in patients
all registered patients, each country will set up a network included in COG analyses treated with immediate
for the provision of biological samples. Such a network for nephrectomy, although the rarity of this aberration
submission and central storage of frozen tumour sam- limits its usefulness42. MYCN gain was found to be
ples from ~80% of children is already established in the associated with adverse outcomes in a retrospective

NATuRe RevIeWS | UROLOGy volume 15 | NOVEMBER 2018 | 699
C o n S e n S u S S tat e m e n t

SIOP–RTSG study published in 2015 (ref.40). Loss of newly discovered molecular aberrations and biomarkers
p53 function might be associated with poor outcome and screening for novel therapeutic options.
in anaplastic tumours, which was suggested to be asso- Last but not least, the comprehensive collection of
ciated with the volume of anaplastic regions and their paediatric renal tumours will also aid in the assembly
incomplete removal5–7. of sufficiently large cohorts to analyse rare entities such
Comprehensive NGS analysis of two Wilms tumour as mesoblastic nephroma, cystic partially differentiated
cohorts has yielded a large number of potential driver nephroblastoma, clear cell sarcoma of the kidney, rhab-
genes mutated in Wilms tumour, including AMER1, doid tumour of the kidney, or renal cell carcinoma, for
BCOR, CTNNB1, DGCR8, DICER1, DIS3L2, DROSHA, which the molecular basis is often not known. This
FBXW7, GPC3, MLLT1, MYCN, SIX1, SIX2, TP53 and analysis might also uncover genetic alterations that
WT1 (refs8–10). Evaluating whether aberrations in any can be assayed by liquid biopsies, generating future
of these genes, alone or in combination, have a con- options for upfront noninvasive discrimination of
siderable effect on clinical course and survival will be these tumours and paving the way for improved ther-
interesting. With the exception of TP53, mutations in apeutic approaches that might lead to either omission
individual known Wilms tumour genes have not gen- of chemotherapy or a change to more appropriate drug
erally been associated with adverse outcome. However, regimens.
combined SIX1 or SIX2 mutations and microRNA pro-
cessing pathway mutations did confer a worse outcome Conclusions
in at least one of two studies8,9. Thus, a planned tar- Successful implementation of the UMBRELLA proto-
geted NGS panel for Wilms tumour will facilitate these col will have the potential to shape future therapeutic
biomarker studies. approaches and to improve outcomes through several
The secondary molecular aims include a much measures. The UMBRELLA protocol provides guidelines
broader scope of studies, covering exploratory analy- for clinical practice and will hopefully stimulate further
ses of potential novel biomarkers, including circulating international collaboration with harmonization of treat-
nucleic acids detectable in blood and urine, for diagno- ment protocols and research. The UMBRELLA study
sis and prognosis. Liquid biopsies, especially, offer the will provide increased patient cohorts, and the rapid
potential to provide a global view on the frequently het- central pathology review, the standardization of speci-
erogeneous tumours, reducing the risk of overlooking men handling and the improved collection of biological
relevant tumour subclones. samples by pathologists will provide large numbers of
On the national level, multiple efforts are underway to specimens with increased homogeneity compared with
explore the role of miRNAs in blood and tumour tissue other studies collected prospectively. These improved
as biomarkers. A range of initiatives to molecularly char- procedures will be of utmost importance to validate
acterize all subtypes of Wilms tumour and non-​Wilms biomarkers such as 1q gain, to find driver gene muta-
tumour and their associated nephrogenic rests using tions, to evaluate the prognostic significance of blastemal
whole-​genome, epigenomic and proteomic approaches volume, and to unravel tumour heterogeneity with the
have been planned. This characterization extends to ultimate goal to improve treatment stratification and to
in vitro and in vivo cell models, in which tumours are delineate novel therapeutic targets.
used to establish cell cultures, organoid cultures, or xeno-
graft models that should support functional validation of Published online 11 October 2018

1. Pastore, G. et al. Malignant renal tumours incidence 8. Walz, A. L. et al. Recurrent DGCR8, DROSHA, and SIX treated according to intensified treatment in the SIOP
and survival in European children (1978–1997): homeodomain mutations in favourable histology Wilms tumour 2001 protocol, a report of the SIOP
report from the Automated Childhood Cancer Wilms tumors. Cancer Cell 27, 286–297 (2015). Renal Tumour Study Group (SIOP-​RTSG). Eur. J.
Information System project. Eur. J. Cancer 42, 9. Wegert, J. et al. Mutations in the SIX1/2 Pathway and Cancer 51, 498–506 (2015).
2103–2114 (2006). the DROSHA/DGCR8 miRNA microprocessor complex 16. van den Heuvel-​Eibrink, M. M. et al. Malignant
2. Spreafico, F. et al. Paediatric renal tumours: underlie high-​risk blastemal type Wilms tumors. rhabdoid tumours of the kidney registered on recent
perspectives from the SIOP-​RTSG. Nat. Rev. Urol. 14, Cancer Cell 27, 298–311 (2015). SIOP protocols from 1993 to 2005: a report of
3–4 (2017). 10. Gadd, S. et al. A Children’s Oncology Group and the SIOP renal tumour study group. Pediatr. Blood
3. van den Heuvel-​Eibrink, M. M. et al. Position paper: TARGET initiative exploring the genetic landscape of Cancer 56, 733–737 (2011).
rationale for the treatment of Wilms tumour in the Wilms tumor. Nat. Genet. 49, 1487–1494 (2017). 17. Vujanic, G. M. & Sandstedt, B. The pathology of
UMBRELLA SIOP-​RTSG 2016 protocol. Nat. Rev. Urol. 11. Gratias, E. J. et al. Association of chromoseme 1q gain Wilms’ tumour (nephroblastoma): the International
14, 743–752 (2017). with inferior survival in favourable-​histology Wilms Society of Paediatric Oncology approach. J. Clin.
4. Cancer Research UK. A study looking at improving tumor: a report from the Children’s Oncology Group. Pathol. 63, 102–109 (2010).
treatment for children with kidney cancer (IMPORT). J. Clin. Oncol. 34, 3189–3194 (2015). 18. Vujanic, G. M. et al. Revised International Society
Cancer Research UK https://www.cancerresearchuk. 12. Chagtai, T. et al. Gain of 1q as a prognostic biomarker of Paediatric Oncology (SIOP) working classification of
org/about-​cancer/find-​a-clinical-​trial/study-​improving- in Wilms tumors treated with preoperative renal tumors of childhood. Med. Pediatr. Oncol. 38,
treatment-​children-kidney-​cancer (2013). chemotherapy in the International Society of 79–82 (2002).
5. Maschietto, M. et al. TP53 mutational status is Paediatric Oncology (SIOP) Wilms tumour 2001 Trial: 19. Pritchard-​Jones, K. et al. Omission of doxorubicin
a potential marker for risk stratification in Wilms a SIOP renal tumours biology consortium study. from the treatment of stage II-​III, intermediate-​risk
tumour with diffuse anaplasia. PLOS ONE 9, e109924 J. Clin. Oncol. 34, 3195–3203 (2016). Wilms’ tumour (SIOP WT 2001): an open-​label,
(2014). 13. Graf, N. et al. Is absolute blastema volume after non-​inferiority, randomised controlled trial. Lancet
6. Ooms, A. H. et al. Significance of TP53 mutation in preoperative chemotherapy in nephroblastoma 386, 1156–1164 (2015).
Wilms tumors with diffuse anaplasia: a report from the relevant for prognosis. Pediatr. Blood Cancer 57, 20. Reinhard, H. et al. Outcome of relapses of
Children’s Oncology Group. Clin. Cancer Res. 22, 741–742 (2011). nephroblastoma in patients registered in the SIOP/
5582–5591 (2016). 14. Dome, J. S., Perlman, E. J. & Graf, N. Risk GPOH trials and studies. Oncol. Rep. 20, 463–467
7. Wegert, J. et al. TP53 alterations in Wilms tumour stratification for Wilms tumor: current approach and (2008).
represent progression events with strong intratumour future directions. Am. Soc. Clin. Oncol. Educ. Book 34, 21. Beckwith, J. B. & Palmer, N. F. Histopathology and
heterogeneity that are closely linked but not limited 215–223 (2014). prognosis of Wilms tumors: results from the first
to anaplasia. J. Pathol. Clin. Res. 3, 234–248 15. van den Heuvel-​Eibrink, M. M. et al. Outcome of national Wilms’ tumor study. Cancer 41, 1937–1948
(2017). localised blastemal-​type Wilms tumour patients (1978).

700 | NOVEMBER 2018 | volume 15 www.nature.com/nrurol

 C o n S e n S u S S tat e m e n t

22. Faria, P. et al. Focal versus diffuse anaplasia in Wilms 33. Furtwängler, R. et al. Pretreatment for bilateral (SIOP–RTSG) Steering Committee for their contribution, sup-
tumor - new definitions with prognostic significance: nephroblastomatosis in an independent risk factor port and critical input to the UMBRELLA protocol. The
a report from the National Wilms Tumor Study Group. for progressive disease in patients with stage V authors would like to dedicate this manuscript to Ivo
Am. J. Surg. Pathol. 20, 909–920 (1996). nephroblastoma. Klin. Padiatr. 226, 175–181 (2014). Leuschner, who had worked on the SIOP–RTSG for many
23. Pritchard-​Jones, K. et al. Treatment and outcome 34. Murphy, W. M., Grignon, D. J. & Perlman, E. J. in years, making considerable contributions to its success.
of Wilms’ tumour patients: an analysis of all cases Tumors of the Kidney, Bladder, and Urinary Structures
registered in the UKW3 trial. Ann. Oncol. 23, (AFIP Atlas of Tumor Pathology, 4th Series) 47–49 Author contributions
2457–2463 (2012). (American Registry of Pathology, 2004). All authors researched data for and wrote the manuscript,
24. Reinhard, H. et al. Results of the SIOP 93-01/GPOH 35. Cajaiba, M. M. et al. Pediatric cystic nephromas: made substantial contributions to discussions of content and
trial and study for the treatment of patients with distinctive features and frequent DICER1 mutations. edited and reviewed the manuscript before submission.
unilateral nonmetastatic Wilms tumor. Klin. Padiatr. Hum. Pathol. 48, 81–87 (2016).
216, 132–140 (2004). 36. Weeks, D. A., Beckwith, J. B. & Mierau, G. W. Bening Competing interests
25. Vujanic, G. M., Sandstedt, B., Kelsey, A. & Sebire, N. J. nodal lesions mimicking metastases from pediatric The authors declare no competing interests.
Central pathology review in multicenter trials and renal neoplasms: a report of the National Wilms’
studies: lessons from the nephroblastoma trials. Tumor Study Pathology Center. Hum. Pathol. 21, Publisher’s note
Cancer 115, 1977–1983 (2009). 1239–1244 (1990). Springer Nature remains neutral with regard to jurisdictional
26. Charlton, J., Irtan, S., Bergeron, C. & Pritchard-​Jones, K. 37. Dumoucel, S. et al. Malformations, genetic claims in published maps and institutional affiliations.
Bilateral Wilms tumour: a review of clinical and abnormalities, and Wilms tumor. Pediatr. Blood
molecular features. Expert Rev. Mol. Med. 19, e8 Cancer 61, 140–144 (2014).
Reviewer information
(2017). 38. Mussa, A. et al. Nephrological findings and genotype-
Nature Reviews Urology thanks C. Granberg and the other
27. Beckwith, J. B., Kiviat, N. B. & Bonadio, J. F. phenotype correlation in Beckwith-​Wiedemann
anonymous reviewer(s) for their contribution to the peer
Nephrogenic rests, nephroblastomatosis, and the syndrome. Pediatr. Nephrol. 27, 397–406 (2012).
review of this work.
pathogenesis of Wilms’ tumor. Pediatr. Pathol. 10, 39. Godzinski, J., Graf, N. & Audry, G. Current concepts in
1–36 (1990). surgery for Wilms tumour – the risk and function-​
Supplementary information
28. Beckwith, J. B. Nephrogenic rests and the pathogenesis adapted strategy. Eur. J. Pediatr. Surg. 24, 457–460
Supplementary information is available for this paper at
of Wilms tumor: developmental and clinical (2014).
https://doi.org/10.1038/s41585-018-0100-3.
considerations. Am. J. Med. Genet. 79, 268–273 40. Williams, R. D. et al. Multiple mechanisms of MYCN
(1998). dysregulation in Wilms tumour. Oncotarget 6,
29. Breslow, N. E., Beckwith, J. B., Perlman, E. J. & 7232–7243 (2015). Open Access This article is licensed under
Reeve, A. E. Age distributions birth weights, nephrogenic 41. Cresswell, G. D. et al. Intra-​tumor genetic a Creative Commons Attribution 4.0
rests and heterogeneity in the pathogenesis of Wilms heterogeneity in Wilms tumor: clonal evolution and International License, which permits use,
tumor. Pediatr. Blood Cancer 47, 260–267 (2006). clinical implications. EBioMedicine 9, 120–129 sharing, adaptation, distribution and reproduction in any
30. Vujanic, G. M. et al. Nephrogenic rests in Wilms tumors (2016). medium or format, as long as you give appropriate credit to
treated with preoperative chemotherapy: the UKSIOP 42. Grundy, P. E. et al. Loss of heterozygosity for the original author(s) and the source, provide a link to the
Wilms Tumor 2001 Trial experience. Pediatr. Blood chromosomes 1p and 16q is an adverse prognostic Creative Commons license, and indicate if changes were
Cancer https://doi.org/10.1002/pbc.26547 (2017). factor in favourable-​histology Wilms tumor: a report made. The images or other third party material in this article
31. Hennigar, R. A., O’Shea, P. A. & Grattan-​Smith, J. D. from the National Wilms Tumor Study Group. J. Clin. are included in the article’s Creative Commons license, unless
Clinicopathologic features of nephrogenic rests and Oncol. 23, 7312–7321 (2005). indicated otherwise in a credit line to the material. If material
nephroblastomatosis. Adv. Anat. Pathol. 8, 276–289 is not included in the article’s Creative Commons license and
(2001). Acknowledgements your intended use is not permitted by statutory regulation or
32. Perlman, E. J. et al. Hyperplastic perilobar The authors thank all International Society of Paediatric exceeds the permitted use, you will need to obtain permission
nephroblastomatosis: long-​term survival of 52 Oncology (SIOP) pathologists and biologists, as well as the directly from the copyright holder. To view a copy of this
patients. Pediatr. Blood Cancer 46, 203–221 (2006). members of the Renal Tumour Study Group of the SIOP license, visit http://creativecommons.org/licenses/by/4.0/.

NATuRe RevIeWS | UROLOGy volume 15 | NOVEMBER 2018 | 701