New Approaches to Risk Stratification for Wilms Tumor

Abstract

Purpose of review:

The treatment of Wilms tumor (WT) is one of the great achievements in the field of oncology. One of the key success factors has been improved risk stratification, enabling augmentation or reduction of therapy depending on a patienťs risk of relapse. This article highlights the evolution of clinical and biological prognostic markers that have been applied in the treatment of WT.

Recent findings:

Historically, tumor stage and histology were the sole determinants of WT treatment. Recent clinical trials conducted by the Children's Oncology Group (COG) and the International Society of Pediatric Oncology (SIOP) Renal Tumor Study Group have expanded the menu of prognostic factors to include histologic and volumetric response to therapy and tumor-specific loss of heterozygosity (LOH) at chromosomes 1p and 16q. Augmentation of therapy has been able to overcome the adverse risk factors. An emerging prognostic marker is chromosome 1q gain, will be incorporated into future clinical trials.

Summary:

The application of new clinical and biological prognostic factors has created unprecedented ability to tailor therapy for WT, accompanied with improved outcomes. Current and future trials will continue to enhance precision medicine for WT.

Introduction

Wilms tumor (WT) is the second most common extracranial solid tumor and the most common malignant renal tumor in children. It accounts for 5% of all childhood malignancies and 80% of all diagnosed renal cancers in children and teenagers. Because WT is a malignant embryonal tumor, most cases are diagnosed in children under the age of 5 years. In the United States and Canada, the estimated incidence is about 9 per million children under 15 years old, affecting 1 in 10,000 children [1–2]. Similar rates are reported in Europe, Australia, and New Zealand, with lower rates in Asia, Central and South America. Areas such as Harare, Zimbabwe, in Africa, have higher incidence of 16.5 per million [1]. Most cases of WT are unilateral, with 5–10% of cases affecting both kidneys. Bilateral WT is more common in patients with underlying genetic syndromes.

More than 15 different syndromes are associated with WT, including WAGR (Wilms tumor, aniridia, genitourinary abnormalities, and a range of developmental delays), Denys-Drash (Wilms tumor, diffuse mesangial sclerosis leading to early-onset renal failure, and intersex disorders that can range from ambiguous to normal-appearing female genitalia in both XY & XX individuals), and Beckwith-Wiedemann (embryonal tumors, macrosomia, macroglossia, hemihypertrophy, visceromegaly, omphalocele, neonatal hypoglycemia, and ear creases/pits) [3]. Only 10% of WT cases are associated with an underlying constitutional mutation, and therefore the etiology of most cases is unknown [4].

Patients with WT are usually asymptomatic at initial diagnosis. In most cases, a parent will identify an abdominal mass upon bathing or dressing their child, or a pediatrician will palpate a mass upon examining the child during a routine well-child visit. However, up to 35% of patients can present with either hematuria, hypertension, fever, or flank pain [5]. Rarely, a patient can present with an acute abdomen in the setting of tumor rupture and bleeding into surrounding tissue. If the diagnosis of WT is clinically suspected, an initial ultrasound of the abdomen with doppler is indicated, followed by referral to a pediatric general surgeon and oncologist for additional work-up and treatment.

The advancement in treatment of Wilms tumor (WT) is one of the great achievements in the field of oncology [6–7]. Stepwise advances made by the National Wilms Tumor Study Group (NWTSG), which was succeeded by the Children’s Oncology Group (COG), and the International Society of Pediatric Oncology (SIOP) Renal Tumor Study Group (RTSG) have not only improved survival to 90%, but have also decreased the burden of therapy [6–7]. A key factor to success has been the refinement of clinical and biological prognostic markers that have enabled risk-directed therapy. This article reviews the evolution of clinical and biological factors that have been adopted for WT.

The original NWTSG and SIOP studies relied solely on tumor stage to define treatment. Over time, additional factors were incorporated into the risk stratification schema, allowing for a multifactorial precision medicine approach (Figure 1). It is important to recognize that prognostic factors must be interpreted in the context of the accompanying treatment regimen. This principle is relevant to WT because COG studies advocate for immediate nephrectomy for most patients whereas SIOP studies advocate for preoperative chemotherapy.

Figure 1.

Evolution of clinical, pathological and molecular prognostic factors over successive studies conducted by the National Wilms Tumor Study Group /Children's Oncology Group and the International Society of Pediatric Oncology (SIOP) Renal Tumor Study Group. The prognostic factors depicted in red are new prognostic factors that were added for a given generation of study.

COG Approach to Risk Stratification

The COG approach of upfront nephrectomy allows for immediate histologic diagnosis, molecular analysis, and accurate local staging assessment. This knowledge can identify a subset of patients with very low-risk WT who may be treated with nephrectomy alone.

Tumor Stage

Tumor stage was the earliest known prognostic factor for WT, with outcomes worsening with advancing stage [7]. Stage V designates bilateral disease, which has been associated with the least favorable outcome, though survival was markedly improved in the recent COG AREN0534 study [8]. Recent analyses have uncovered prognostic differences within the same stage category. A retrospective analysis of NWTS-5 demonstrated that the disparate factors defining stage III (positive surgical margins, positive lymph nodes, tumor rupture, peritoneal implants) impact the risk of recurrence to a varying and often additive extent [9]. Among patients with stage III disease enrolled on COG study AREN0532, 4-year event-free survival (EFS) was 82% with positive lymph nodes, compared to 94% with negative lymph nodes (p<0.01) [10]. An additive effect was observed when lymph node positivity was combined with loss of heterozygosity (LOH) at either chromosome 1p or 16q. Four-year EFS was 74% with both positive lymph nodes and LOH at either 1p or 16q, 86–87% with either positive lymph nodes or LOH at one of the loci, and 97% with negative lymph nodes and lack of LOH (p<0.001). Based on this observation, patients with positive lymph nodes and LOH at either 1p or 16q will receive additional chemotherapy drugs in the next COG study.

The ideal number of lymph nodes that should be sampled remains unknown, though studies have suggested that sampling more nodes increases the yield of detecting tumor [11]. Investigators have suggested that sampling 6–10 lymph nodes [12*] or >7 nodes [11] is optimal to detect WT involvement.

Histology

The presence of anaplasia has long been recognized as an adverse prognostic factor for WT [13–16*]. Although anaplastic histology is a marker of resistance to therapy, recent studies have shown that this resistance can be partially overcome with additional chemotherapy drugs. For stage I anaplastic WT, the addition of flank radiation and doxorubicin to vincristine/dactinomycin improved the four-year EFS from 70% on NWTS-5 to 100% on COG AREN0321 [14,16*]. Retrospective analyses of stage I anaplastic WT suggested a clear contribution of doxorubicin, with an uncertain role for flank radiation [16*,17*]. For stage II-IV diffuse anaplastic WT, inclusion of carboplatin in addition to vincristine/doxorubicin/cyclophosphamide/etoposide, the inclusion of vincristine/irinotecan for stage IV disease, and a higher dose of flank radiation for stage III disease improved four-year relapse-free survival from 58% to 73%, though the new therapy was associated with increased toxicity [15*].

Two studies demonstrated that detection of a TP53 mutation in anaplastic WT was associated with adverse outcome [18–19]. However, careful analysis including copy number assessment and immunohistochemistry, as well as confirming that anaplasia was present in the portion of tumor undergoing mutation analysis, indicated that nearly all anaplastic WT have TP53 mutations if one looks hard enough [19]. Hence, the detection of a TP53 mutation in a randomly selected tumor sample may represent a surrogate for the burden of anaplasia.

A retrospective review of stage I epithelial-predominant WT revealed outstanding prognosis, with four-year EFS and OS of 96% and 100%, respectively [20*]. The upcoming COG study will prospectively assess whether this group of patients can be treated with nephrectomy only.

Age

Increasing patient age is associated with increased risk of recurrence and worse outcome. This is partially attributed to the fact that anaplastic histology is rare in very young patients, however older patients with favorable histology WT also have worse outcomes than younger patients [21]. On NWTS-5, patients less than 2 years old with stage I disease, favorable histology, and tumor weight <550 grams were treated with nephrectomy alone. A lower than expected 4-year EFS of 84% triggered predefined stopping rules and the study was discontinued [22–23]. However, nearly all patients ultimately survived, leading the COG AREN0532 study to revisit this question. Among 116 patients enrolled, four-year EFS and OS were 89.7% and 100%, confirming that nephrectomy-only is an acceptable approach for this group [24]. Interestingly, tumor-specific LOH or loss of imprinting (LOI) at chromosome 11p15 (the Beckwith-Wiedemann Syndrome locus) was associated with an increased risk of relapse. While only 3% of patients without LOH/LOI relapsed, 20% with LOH 11p15 and 25% with LOI 11p15 relapsed [24], confirming earlier observations [25–26]. Based on these findings, patients whose tumors demonstrate LOH 11p15 will not be candidates for nephrectomy-only in the upcoming COG study. Conversely, tumor weight will no longer be a criterion for the surgery-only approach and patients up to age 4 years will be eligible for the study based on findings from the United Kingdom Children’s Cancer Study Group [24]. Expanding the age range and eliminating the tumor weight criteria will double the number of patients eligible for surgery only.

Molecular Genetic Markers: LOH of 1p and 16q and gain of chromosome 1q

NWTS-5 prospectively evaluated the prognostic significance of LOH for chromosomes 16q and 1p in favorable histology WT. LOH for either chromosome segment was found to correlate with increased risk of relapse and death, however the most significant impact was in groups with LOH for both 1p and 16q [27]. Based on these findings, AREN0532 and AREN0533 assessed escalation of therapy for patients with combined LOH of 1p and 16q (patients with LOH at either 1p or 16q received the standard treatment regimens). Patients with stage I and II favorable histology WT who had LOH of 1p and 16q received intensified therapy with Regimen DD4A (vincristine/dactinomycin/doxorubicin) instead of the historic Regimen EE4A (vincristine/dactinomycin). Patients with stage III and IV favorable histology WT received four cycles of cyclophosphamide and etoposide in addition vincristine/dactinomycin/doxorubicin (Regimen M) instead of the historic Regimen DD4A. Four-year EFS for stage I/II improved to 87% (versus 69% in NWTS-5, p=0.42) and 4-year EFS for stage III/IV improved to 90% (versus 61% in NWTS-5, p=0.001). OS estimates were also improved but not statistically different [28*]. Based on these findings, LOH at 1p and 16q will be carried forward as a risk stratification factor in the upcoming COG studies.

Combined LOH 1p and 16q has limited impact as a prognostic marker because only 5% of favorable histology WT carry this molecular change. By contrast, gain of chromosome 1q is one of the most commonly observed cytogenetic abnormalities in WT, seen in as many as 30% of patients [29]. A retrospective analysis of more than 1000 patients in NWTS-5 demonstrated that 1q gain was associated with inferior EFS and OS across all tumor stages [29]. There also was a correlation between LOH 16q/1p and gain of 1q because a translocation involving chromosomes 1p and 16q can result in both molecular lesions [30]. In the next generation of COG studies, patients with 1q gain will generally receive augmented therapy and will not be eligible for planned reductions in therapy, such as surgery-only for patients with stage I favorable histology WT.

Lung Nodule Response

On NWTSG studies, all patients with pulmonary metastasis were subjected to whole lung radiation (WLI). A subgroup of the SIOP 93–01 and 2001 populations analyzed by the German Society of Pediatric Oncology/Hematology indicated that patients with complete lung nodule response after 6 weeks of chemotherapy had superior EFS and OS compared to patients with incomplete response [31]. Based on these findings, COG trial AREN0533 used a risk-adapted approach to stage IV favorable histology WT and pulmonary metastases. Patients with incomplete response of lung nodules after six weeks of DD4A chemotherapy received WLI and escalated to chemotherapy Regimen M. With this approach, 4-year EFS was 90% (versus an expected 75% based on NWTS-5, p<0.001) and 4-year OS was 96%, strongly suggesting a benefit to Regimen M. A pitfall to Regimen M is its cumulative cyclophosphamide dose of 8.8 g/m², placing patients at risk for late effects including infertility, especially males. Future COG studies plan to limit the cyclophosphamide cumulative dose to 4.4 g/m², replacing two cyclophosphamide/etoposide cycles with vincristine/irinotecan.

Patients with complete lung nodule response after 6 weeks of DD4A chemotherapy had WLI omitted, resulting in a 4-year EFS of 79% (versus an expected 85% based on NWTS-5, p=0.052) and 4-year OS of 96% [32]. Although the OS was excellent, the number of relapses was greater than expected with omission of WLI. A post-hoc analysis showed that chromosome 1q gain provided a valuable marker to select patients with complete lung nodule response who should not be candidates for omission of WLI; among complete responders, EFS was 57% in those with 1q gain compared to 86% in those without 1q gain (p=0.001). By contrast, there was no difference in outcome according to 1q gain status in the incomplete lung nodule responders who received Regimen M, suggesting that Regimen M overcame the negative effect of 1q gain.

A summation of prognostic factors, treatments and outcomes utilized in recent COG trials is provided in Table 1.

Table 1.

Risk stratification, treatment, and outcomes for Wilms Tumor on recently completed COG Studies

Stage	Histo	Patient Age	Tumor Weight	LOH 1p/16q	Lung Nodule Response	Study	4-year EFS (%)	4-year OS (%)	Chemo	Radiation	Reference
I	FH	< 2 yrs	< 550 g	Any	N/A	AREN0532	90	100	None	•None	19
I	FH	<2 yrs	≥ 550 g	No	N/A	None	94*	98*	EE4A	• None	27
		≥2 yrs	Any	No	N/A
II	FH	Any	Any	No	N/A	None	86*	98*	EE4A	• None	27
I/II	FH	Any	Any	Yes	N/A	AREN0532	87	100	DD4A	• None	25
III	FH	Any	Any	No	N/A	AREN0532	88	97	DD4A	• Flank/Abd	10
IV	FH	Any	Any	No	CR	AREN0533	80	96	DD4A	• Flank/Abd for local stage III • WLI omitted	24
III/IV	FH	Any	Any	Yes	N/A	AREN0533	90	96	M	• Flank/Abd for local stage III • WLI for lung metastasis • RT to other metastatic sites	25
IV	FH	Any	Any	Any	IR	AREN0533	89	95			24
I	DA FA**	Any	Any	Any	NA	AREN0321	100	100	DD4A	• Flank/Abd	16
II	DA	Any	Any	Any	Any	AREN0321	87	86	UH-1	• Flank/Abd • WLI for lung metastasis • RT to other metastatic sites	15
III	DA	Any	Any	Any	Any	AREN0321	81	89	UH-1
IV	DA	Any	Any	Any	Any	AREN0321	43	49	UH-1/UH-2
V	Any	Any	Any	Any	Any	AREN0534	82	95	Variable	• Variable	8

^*Data reported are from NWTS-5 (without LOH 1p) because these groups were not studied by COG

^**Data for stage II-IV FA not reported due to small numbers

Histo: histology; FH: favorable histology; DA: diffuse anaplasia; FA: focal anaplasia; WLI: whole lung irradiation; RT: radiation therapy; N/A: not applicable; CR: complete response; IR: incomplete response; LOH: loss of heterozygosity

EE4A: vincristine/dactinomycin x 19 weeks/ DD4A: vincristine/dactinomycin/doxorubicin x 24 weeks

M: vincristine/dactinomycin/doxorubicin/cyclophosphamide/etoposide x 33 weeks

UH-1: vincristine/ doxorubicin/cyclophosphamide/carboplatin/etoposide x weeks

UH-2: UH-1 with irinotecan

SIOP Approach to Risk Stratification

The SIOP nephroblastoma studies and trials have centered around preoperative therapy since their inception in 1971. The early studies showed that neoadjuvant chemotherapy prevents intraoperative tumor rupture and induces favorable stage distribution. Subsequent trials modified post-operative treatment based on tumor stage and histology, leading to the current UMBRELLA protocol, which will identify and refine new biomarkers for the management of nephroblastoma independent of tumor type, socio-economic status, or geographic region [33].

Tumor Stage

In the SIOP system, patients are staged at diagnosis by imaging studies into local (stage I-III), metastatic (stage IV), or bilateral disease (stage V). To define metastatic disease by chest CT-scan, a clear description of lung lesions and their interpretation is done in a prospective way. Standardization and quality improvement of imaging studies are of utmost importance, therefore, analysis by reference radiology is mandatory in the UMBRELLA protocol. After preoperative chemotherapy and surgery, the local stage is redefined, which determines the intensity of postoperative treatment, which varies between no further treatment in totally necrotic stage I tumors, to more than 6 months intensive chemotherapy [33].

Histology

As in the COG system, diffuse anaplasia is a high-risk histological subtype. Additionally, preoperative chemotherapy allows for assessment of histologic response to treatment. Low-risk histology, defined as completely necrotic tumor, occurs in 5% of tumors. High-risk histology includes blastemal-type tumors, as a predominance of residual blastemal cells is a marker of chemotherapy resistance. Intermediate-risk histology represents all others [34]. Stromal and epithelia-type nephroblastoma show an excellent outcome after preoperative chemotherapy [35]. In retrospective analyses, the absolute blastemal volume after preoperative chemotherapy seems to be of prognostic relevance, showing that patients with more than 20 ml blastema have a poorer outcome than those with less blastema [36]. Importantly, an increase in therapy on the SIOP-2001 study improved event-free survival in patients with localized blastemal-type WT [37] whereas outcomes for stage IV blastemal-type or diffuse anaplastic WT remained unsatisfactory [38*].

If available, diffusion-weighted MRI is recommended as this may help characterize the tumor. The apparent diffusion coefficient (ADC) value correlates with cell density and may help to correlate with histology after preoperative chemotherapy [39–42].

Age

Data from the SIOP 93–01 and 2001 cohorts (5631 patients, median age 3.4 years) were explored to study the independent prognostic value of age on survival [43*]. Although optimal cutoffs for age could not be identified, previously described cutoffs for EFS (2 and 4 years) and OS (4 years) demonstrated the prognostic significance of age in children with WT treated with preoperative chemotherapy. These findings encourage the consideration of age in the design of future SIOP-RTSG protocols especially in combination with molecular markers, such as 1q gain.

Treatment Response

In vivo response to preoperative chemotherapy is an important stratification parameter in the SIOP trials [44]. Treatment response encompasses both volumetric and histologic response (described above). Residual tumor volume above 500 ml is an adverse prognostic factor for intermediate-risk histology other than epithelial and stromal-type tumors [33]. This is thought to be due to the absolute volume of remaining blastema [36]. Tumor shrinkage also allows for nephron sparing surgery (NSS), but this should be performed by experienced surgeons to avoid a high postoperative local stage [45]. To better assess the impact of NSS on late effects, a new predictive formula has been developed and evaluated. This formula takes into account not only the surgical and histological resection margin, but also the remaining normal kidney [46].

Response of distant metastases is also prognostically important. Patients with stage IV WT achieving a complete remission after preoperative chemotherapy do not require lung irradiation [47]. In patients with remaining lung nodules, metastastectomy and histological confirmation of metastasis is advised. Irradiation can be avoided for completely necrotic metastasis and complete response achieved with surgical resection.

Molecular markers

In contrast to COG, molecular markers are not used for adapting postoperative treatment in WT. In a retrospective analysis of different molecular markers in tumor samples from 586 patients enrolled on SIOP-2001, an association of 1q gain with poorer EFS retained significance in multivariable analysis adjusted for 1p and 16q loss, sex, stage, age, and histologic risk group. In addition, MYCN gain and TP53 loss were associated with poorer outcome [48–49]. These results and their association with the blastemal subtype are being evaluated prospectively in the ongoing UMBRELLA protocol. It is important to take the intratumoral heterogeneity into account to get reliable results [50].

Liquid biopsies show promise to play an important role to diagnose kidney tumors together with imaging studies, for surveillance and predicting earlier relapse, and for overcoming tumor heterogeneity [51–52]. Circulating tumor DNA (ctDNA) and microRNA (miRNA) have shown promise to detect WT, distinguish histologic subtypes, and correlate with chemosensitivity [53–54].

A summary of risk-groups, treatments, and patient outcomes for recent SIOP studies is depicted in Table 2.

Table 2.

Risk stratification, treatment, and outcomes for Wilms Tumor on recently completed SIOP Studies

Stage	Histo	Patient Age	Tumor volume+	Tumor histologic response	Lung Nodule response	Study	EFS (5-yrs) %	OS (5-yrs) %	Pre-operative chemo	Post-operative chemo	Radiation	Reference
I	LR	>6mo & <16yrs	Any	CN	-	SIOP 2001	96 (2 yrs)	99 (2 yrs)	AV 4w	None	None	**
I	IR	>6mo & <16yrs	<500ml++	No BT	-	SIOP 2001	92 (2 yrs)	98 (2 yrs)	AV 4w	AV 4w	None	**
I	HR	>6mo & <16yrs	Any & >500ml IR++	HR*	-	SIOP 2001	BT: 94.7	BT: 100	AV 4w	AVD 28w	None	55
II	IR	>6mo & <16yrs	<500ml++	No BT	-	SIOP 2001	AVD: 91.6 AV: 84.8	AVD: 97,6 AV: 95,5	AV 4w	AVD (R) AV	None	56
II	HR	>6mo & <16yrs	Any & >500ml IR++	HR*	-	SIOP 2001	81 (2 yrs)	78 (2 yrs)	AV 4w	4 drugs	Only DA	**
III	LR	>6mo & <16yrs	Any	CN	-	SIOP 2001	100 (2 yrs)	100 (2 yrs)	AV 4w	AV 28w	None	**
III	IR	>6mo & <16yrs	Any	No BT	-	SIOP 2001	AVD: 90,5 AV: 85,1	AVD: 93,8 AV: 96,0	AV 4w	AVD (R) AV	Flank/Abd	56
III	HR	>6mo & <16yrs	Any	HR*	-	SIOP 2001	67 (2 yrs)	70 (2 yrs)	AV 4w	4 drugs	Flank/Abd	**
IV	LR	>6mo & <16yrs	Any	CN	CR#	SIOP 2001	91,5	94,8	AVD 6 w	AVD 28w	None	**
IV	IR	>6mo & <16yrs	Any	No BT	CR#	SIOP 2001	84#	92#	AVD 6 w	AVD 28w	Flank/Abd	42
IV	HR	>6mo & <16yrs	Any	HR*	CR#	SIOP 2001	n.d.	82	AVD 6 w	4 drugs	Flank/Abd and WLI	42,38
IV	IR	>6mo & <16yrs	Any	No BT	Non-CR	SIOP 2001	78,6#	88,8#	AVD 6 w	4 drugs	Flank/Abd and WLI	42
IV	HR	>6mo & <16yrs	Any	HR*	Non-CR	SIOP 2001	33,3#	PR:39 PD: 0	AVD 6 w	4 drugs	Flank/Abd and WLI	42, 38
V	Any	Any	Any	Any	Any	SIOP 2001	75,6	87,7	AV / AVD	Variable	Variable	**

⁺Tumor volume: after preoperative chemotherapy

^**H. van Tinterin, SIOP 2001 statistical report, personal communication

⁺⁺not including stromal or epithelial type (only GPOH in SIOP 2001, all participants in Umbrella); LR: low risk; IR: intermediate risk; HR: high risk; CN: complete necrotic; BT: Blastemal Type; HR*: diffuse anaplasia & BT; (R): Randomization; DA: Diffuse anaplasia; CR^#: CR achieved by chemotherapy alone or chemotherapy plus metastectomy; WLI: whole lung irradiation; n.d. not done;

^#:SIOP 93–01

Conclusion

The application of new clinical and biological prognostic factors has created unprecedented ability to tailor therapy for WT. Recent clinical trials have demonstrated that augmented therapy can overcome the effect of certain adverse prognostic factors. The ongoing SIOP UMBRELLA protocol and future COG trials seek to further refine prognostic factors to enhance precision medicine for WT.

Key Points.

• Recent clinical trials have incorporated tumor weight or volume, patient age, response to therapy, and loss of heterozygosity (LOH) of chromosomes 1p and 16q into the risk stratification schema for Wilms tumor, supplementing the long-standing prognostic factors of tumor stage and histology

• Trials from the Children's Oncology Group (COG) have demonstrated that augmented therapy improved outcomes for Wilms tumor with anaplastic histology, incomplete lung nodule response, and LOH at 1p and 1p.

• Trials from the International Society of Pediatric Oncology (SIOP) have focused on adjusting therapy based on histologic and volumetric response to preoperative chemotherapy. Augmentation of therapy improved outcomes for patients with localized high-risk blastemal-type Wilms tumor.

• Chromosome 1q gain has shown promise as a prognostic marker in both the COG and SIOP treatment contexts.