The rationale for nephron-sparing surgery in unilateral non-syndromic Wilms tumour

Survivors of childhood cancer remain at risk of reduced life expectancy and morbidity due to treatment-related effects [1]. In 2016, The Renal Tumour Study Group of the International Society of Paediatric Oncology developed a protocol that established new pathways in diagnosis and treatment of children with renal tumours, to continue the iterative process of collaboration and research. This was termed the UMBRELLA protocol and one of its goals was to decrease acute toxicity and long-term treatment-related sequelae [2]. The chosen surgical strategy plays an important role in preserving functional kidney and reducing morbidity through the selective utilization of nephron-sparing surgery (NSS) [3]. To preserve kidney function, NSS was established in children with bilateral Wilms tumours (WT), or unilateral WT with a syndrome predisposing to metachronous Wilms tumours. However, in the recent UMBRELLA protocol, NSS has become an acceptable treatment option in non-syndromic unilateral WT when certain criteria are met [2]. Adoption of NSS has been slow [3] due to concern around surgical complications, and positive surgical margins necessitating additional therapy [4].

The principles that have established the application of NSS in children include a physiological rationale for its use, consistent case selection, maintenance of the oncological principles of surgery, and surgery that is performed in a high-volume centre by an experienced team. These themes will subsequently be discussed in detail. The goal of this review was to determine a rationale for the adoption of NSS in a wider context and to specify the conditions, extended indications, technical considerations, and expected outcomes.

The goal of NSS is to safely preserve as much functioning and draining renal parenchyma as possible while maintaining an excellent oncological outcome. To determine the rationale for this, the consequences of total nephrectomy (TN) on kidney function, and the associated cardiovascular morbidity, must be established. The evolution of kidney function with age and the consequences of TN in children or adults can be illustrated by studying various situations. The definition of kidney injury and classification of chronic kidney disease are based on the glomerular filtration rate (GFR) and albuminuria, as standardized by the KDIGO CKD work group [5, 6]. Furthermore, when eGFR is reported, it should be accompanied by the endogenous filtration marker, assay method, and estimating equation. In this consensus statement, criteria and severity for albuminuria and proteinuria are also precisely defined.

GFR is low at birth and reaches adult levels at the end of the second year of life [7]. It is stable at around 120 ml/min/1.73 m² until the fourth decade, after which time it slowly declines by 8 ml/min/1.73 m² per decade [8]. This decline relates to reduced renal mass, glomerular sclerosis, and reduced glomerular capillary flow. There is wide anatomical variation in the total glomerular number, between as little as 200,0000 and nearly 2 million [9, 10]. The number of nephrons present at birth relates to the fetal milieu and genetic factors. Of the many aspects involved, birth weight has been the most widely studied and there is a linear relationship between nephrons at birth and birth weight (260,000 nephrons per kg) [11]. However, it is noted that there is a huge variation around the line of association. Interestingly, the importance of birth weight holds for pre-term infants with appropriate weight for age, or term infants who are small for gestational age [12].

The decline of GFR with age is often undetected because of concomitant decrease in muscular mass [13], notwithstanding that creatinine is not a pure marker of glomerular filtration (it is secreted and reabsorbed by kidney tubules) [14]. Cystatin C is produced by all nucleated cells, is freely filtered by glomeruli and not reabsorbed, and the filtration rate appears unaffected by severe illness [15]. Ultimately, when an accurate GFR assessment is needed a measured GFR should be used [16, 17].

Reduction of GFR with age is of growing importance as life expectancy in developed countries has increased significantly, and kidney failure carries a significant health burden. The diminishing function of the aging kidney is also affected by common co-morbidities (hypertension, nephrotoxic medications, diabetes, and metabolic syndrome). For these reasons, the mean eGFR in the general population over 70 years of age is less than 60 ml/min/1.73 m² [13]. A reduction of GFR (< 45 ml/min/1.73 m²) in adults over 60 years of age doubles all-cause mortality and also has a significant impact on cardiovascular mortality [18]. This notion of the aging kidney is critical when we interpret the consequences of TN in paediatric patients. It underscores that even 20–30 years of follow-up is insufficient to determine the consequences of total nephrectomy.

In 1981, Brenner postulated that hyperfiltration in remnant nephrons following renal ablation contributes to progressive kidney dysfunction after studying rats [19]. Decreased functional kidney reserve leads to increased intraglomerular pressure and hyperfiltration. These physiological alterations may exacerbate any co-existent predisposition to glomerular pathology, whether genetic or acquired. Judicious and pro-active management of risk factors in patients with acquired single kidney (ASK) may slow the progression to chronic kidney disease [14].

Congenital and acquired causes of single kidney have distinct consequences. Nephrogenesis continues up to 36 weeks’ gestation and this allows an opportunity for physiological compensation in situations where there is a congenital single functioning kidney [20]. This phenomenon could explain why cases of congenital single kidney (CSK) may have a superior functional profile, although this has been variably found in comparative studies (Table 1). However, children with both CSK and ASK do show compensatory hypertrophy regardless of aetiology [21], and it may be this adaptive (or maladaptive) response to reduced kidney reserve is more age and stage dependent rather than related to aetiology per se. Furthermore, a higher GFR within the first decade in children with a single kidney needs to be interpreted with caution, as this may represent a higher degree of hyperfiltration and herald a worst prognosis, rather than reflect better kidney function.

A review of adults following childhood TN (for oncological and non-oncological indications) revealed that after > 30 years follow-up: 40% had kidney dysfunction (eGFR < 90 ml/min/1.73 m²), 22% had hypertension, and 23% had albuminuria. After 50 years of follow-up 88% of patients had an eGFR between 46 and 86 ml/min/1.73 m² [27]. These results must be put in the context of a heterogenous group, with 30 years of follow-up and historical regimens of both adjuvant chemotherapy and radiotherapy. Nonetheless, nephrectomy performed for non-oncological indications had similarly poor outcomes (kidney dysfunction 38%, hypertension 26%, albuminuria 33%) after >30 years of follow-up. Care pathways and recommended follow-up have been developed for this population in response to the risks outlined [20, 22, 27, 28].

Although the effects of nephrectomy in childhood and adulthood are quite distinct from an experimental perspective [29], living kidney donors represent a valuable model to understand the evolution of kidney function after TN, being generally healthy and well-studied adults. However, it must be noted that living kidney donors have a higher incidence of developing metabolic syndrome than matched controls [30]. This may partially explain their long-term cardiovascular morbidity, but GFR has been identified as an independent cardiovascular risk factor in this group [31]. Another major limitation of extrapolating outcomes from this group, over and above the differences between donor nephrectomy and ASK in childhood, are the requirement for at least 10 years of follow-up to allow some meaningful application.

A study of nearly 4000 living donors with a mean follow-up of 17 years illustrated some of the expected dysfunction in ASK. Proteinuria was detected in 6%, eGFR < 60 ml/min/1.73 m² in 36%, and 2.5% had an eGFR < 30 ml/min1.73 m² or kidney failure (KF). The cumulative incidence of kidney failure was 1.3 per 1000 donors at 15 years, 7.7 per 1000 donors at 30 years, and 8.8 per 1000 donors at 40 years [32]. These figures must be interpreted with some care. The risk of kidney failure is elevated in related donors, which may be due to ill-defined genetic factors, or more comprehensive and long-term study of related donors (unrelated donation was rare before the 1980s) [32, 33].

Another North American study of over 70,000 kidney donors reported the cumulative incidence of kidney failure at 15-years post-donation. This ranged from 1.1 per 1000 donors to 3.3 per 1000 donors. This was related to eGFR six months post-donation, but importantly the pre-donation eGFR was not predictive of post-donation kidney function [34].

Treatment effects from chemotherapy and radiotherapy on kidney function are an important additional consideration in this population. A cohort of children (n = 75) treated for non-syndromic WT with TN alone (not requiring nephrotoxic chemotherapy or radiotherapy) have illuminated this point. With a mean follow-up of 20 years, 21% had an eGFR of < 90 ml/min/1.73 m²; and no patient had an eGFR of < 60 ml/min/1.73 m². The authors concluded in this specific population the risk of developing significant long-term kidney dysfunction was low [35]. However, the consequences of mild kidney impairment in later adulthood remain unclear.

A smaller (n = 37) but prospective German study had a longer mean follow-up of 25 years. Of note, in this population of children treated for WT, 60% received radiotherapy (predominantly flank/abdominal), and 14% nephrotoxic chemotherapy (cyclophosphamide and/or carboplatin). They found 27% of patients had an eGFR between 60 and 90 ml/min/1.73 m², and 3% had an GFR < 60 ml/min/1.73 m². Hypertension was present in 41% of patients. The results when compared with previous studies (in terms of kidney function and hypertension) are likely explained by the extended follow-up and impact of adjuvant treatment [36].

A Dutch study recently reported the outcomes of 31 WT survivors with 15 years of follow-up. The rates of hypertension were 26% and 4% of patients had an eGFR < 60 ml/min/1.73 m² [37]. And even more recently a French childhood survivor cancer study reported outcomes of 5498 survivors and 42,118 person-years of follow-up. They reported that children who underwent nephrectomy and required ifosfamide (> 60 g/m²) had an extremely high risk of hospitalisation for renal causes [38]. From these studies, there is a proportion of patients treated for WT following TN who will have chronic kidney disease and/or hypertension. However, the relative renal impact of surgery, radiotherapy, and nephrotoxic chemotherapy needs to be more clearly delineated [16]. Furthermore, until WT1 genetic screening is uniformly applied, it is difficult to identify if there is a sub-group of children with poor outcomes who are incorrectly classified as non-syndromic [39].

A large prospective childhood cancer survivor study reported kidney outcomes in children with unilateral non-syndromic WT stratified by treatments received (n = 2008) [40]. Compared to their siblings, WT survivors experienced a ten-fold increased in the rate of kidney failure (2.4% 35-year cumulative incidence) [40]. The addition of abdominal radiotherapy doubled the relative risk of kidney failure to 20-fold [40]. Although this was a North American study, there was no survivor group that underwent nephrectomy alone. Once long-term follow-up for this group becomes available for comparison, it will further clarify the consequences of adjuvant treatments on long-term outcomes.

In the UMBRELLA protocol, post-operative local or flank radiotherapy is indicated in children with unilateral WT with:

It must be emphasised that many aspects of this decision can only be confirmed post-operatively following histological review (e.g. histological risk group, lymph node positivity, tumour at resection margins). Dosing depends on many individual and disease factors, but the protocol stipulates that the dose to the whole kidney should not exceed 12 Gy when NSS is performed, even if there is high-risk histology. The anticipated use of radiotherapy is used as an argument against NSS when there is a high suspicion of lymph node involvement because the effects of radiotherapy are said to counterbalance any benefit of nephron capital linked to NSS (extension to adjacent organs being a contra-indication of NSS per se). However, radiotherapy-associated kidney injury is a very slow process (particularly in association with the protocolised reduction in radiotherapy dose) and is imprecisely defined (but includes hypertension [41, 42] and functional effects). Therefore, although circumstances that may mandate radiotherapy should be considered in decision-making around performing NSS, a rationale for it remains so long as the oncological principles of surgery can be maintained (further discussed under “Nephron-sparing surgery” and illustrated in Fig. 2).

In the International Society of Paediatric Oncology WT trial 2001 (published in 2014) the overall survival and event-free survival in unilateral WT were compared for NSS (91 patients) and TN (2709 patients). Both overall survival (NSS: 100% vs. TN: 94.4%, p = 0.06) and event-free survival (NSS: 94.8% vs. TN: 86.5%, p = 0.06) were similar amongst the two groups. However, in the NSS group, the tumours were smaller and more often limited to the kidney (65% vs. 48%). In this study, the incidence of positive surgical margins in the NSS group was 9% (8/91) and the incidence of positive surgical margins in the TN was 13% (355/2709) [3]. In a separate study, relapse rates were also found to be identical between NSS and TN (when stage and grade of tumour are controlled for) [70].

When considering NSS, the incidence of positive margin is critical. Flank radiotherapy will have consequences on the residual renal parenchyma, although this has been difficult to define [41]. The long-term impact of radiotherapy has also been well documented, with reduced life expectancy [1], and higher rates of second primary malignancies [71].

Because of limited numbers and follow-up of patients with unilateral non-syndromic WT that have undergone NSS, broad and long-term evaluation of kidney function is not well documented. One long-term study following children after NSS or TN for renal tumours into the fifth decade showed significantly higher eGFR measurements associated with NSS, both after the second decade (110 ml/min/1.73 m² vs. 91.2 ml/min/1.73 m²), and at last follow-up (110 ml/min/1.73 m² vs. 95 ml/min/1.73 m², p = 0.02) [72]. Although other studies with shorter follow-up are less impactful, a recent meta-analysis of all studies confirmed higher eGFR in children undergoing NSS (116 ml/min/1.73 m² vs. 98 ml/min/1.73 m²) [73]. Recommendations for surveillance of kidney function and hypertension in WT survivors have been published, although they are based on personal opinion, and extrapolation from solitary kidney and cancer survivorship literature [74].

In the long term, a proportion of patients undergoing TN for unilateral non-syndromic WT will experience a decline in kidney function, hypertension, and/or cardiovascular morbidity. These patients cannot be accurately predicted prior to surgery. For these reasons, NSS must be considered and discussed with patients in whom it is technically feasible, in an effort to reduce morbidity and mortality into adulthood. To decrease the incidence of positive surgical margins, incomplete lymph node sampling, and complications, these procedures should be performed at specialised and experienced reference centres. Based on the impacts of individual treatment pathways, survivors of childhood WT need to be followed throughout adulthood for nephroprotection and to monitor kidney function, blood pressure, and albuminuria, and prevent cardiovascular events.