Overall Survival and Renal Function of Patients with Synchronous Bilateral Wilms Tumor Undergoing Surgery at a Single Institution

Abstract

Objectives

Wilms tumor (WT) is the most common renal cancer in children. Approximately 5% of children with WT present with disease in both kidneys. The treatment challenge is to achieve a high cure rate while maintaining long-term renal function. We retrospectively reviewed our institutional experience with nephron-sparing surgery (NSS) in patients with synchronous bilateral Wilms tumor (BWT) operated on between 2001-2014.

Methods

Imaging studies, surgical approach, adjuvant therapy and pathology reports were reviewed. Outcomes evaluated included surgical complications, tumor recurrence, patient survival and renal function, as assessed by estimated glomerular filtration rate (eGFR).

Results

Forty-two patients with BWT were identified. Thirty-nine (92.9%) patients underwent bilateral NSS; only three patients (7.1%) underwent unilateral nephrectomy with contralateral NSS. Post-operative complications included prolonged urine leak (10), infection (6), intussusception (2) and transient renal insufficiency (1). Three patients required early (within four months) repeat NSS for residual tumor. Long-term, seven (16.7%) patients had local tumor recurrence (managed with repeat NSS in 6 and completion nephrectomy in 1) and three had an episode of intestinal obstruction requiring surgical intervention. Overall survival was 85.7% (mean follow-up, 4.1 years). Of the 6 patients who died, 5 had diffuse anaplastic histology. All patients had an eGFR>60mL/min/1.73m² at last follow-up; no patient developed end-stage renal disease.

Conclusions

In patients with synchronous, bilateral Wilms tumor, bilateral nephron-sparing surgery is safe and almost always feasible, there by preserving maximal renal parenchyma. With this approach, survival was excellent, as was maintenance of renal function.

Introduction

Wilms tumor (WT) is the most common renal tumor of childhood, affecting approximately 650 children in the United States annually.^1,2 Four to 8% of these children present with bilateral disease.^3,4 Surgery is a critical component of the multidisciplinary approach to the management of patients with Wilms tumor, with the standard surgical approach for unilateral WT being radical nephrectomy. Extrapolating this approach to children with bilateral disease would leave affected patients anephric and dependent on renal replacement therapy. The morbidity and adverse effects on the quality of life, and mortality associated with dialysis and renal transplantation in infants and young children, the age groups most commonly affected by bilateral WT, are significant.^5,6 Alternatives to bilateral nephrectomy are essential for the management of these patients.

Bishop and Hope first reported, in 1966,a small series of 6 patients with BWT managed with nephrectomy and contralateral partial nephrectomy⁷, a technique of nephron-sparing surgery first used for BWT by Rickham, as reported in 1957.⁸ Ten years later, Bishop published outcomes of 30 patients with BWT treated on the first national cooperative group study for children with Wilms tumor (National Wilms Tumor Study [NWTS]) and suggested that pre-nephrectomy chemotherapy could effectively shrink tumors, making them more amenable to partial nephrectomy, without compromising oncologic outcomes.⁹ Blute subsequently reported similar findings for 145 children with BWT treated on NWTS-2 and -3. The latter study, in particular, recommended preservative rather than ablative surgery following initial neoadjuvant chemotherapy.¹⁰ Three-year survival for these patients was 82% with long term survival of 70% reported later by Montgomery.³ Nevertheless, two-thirds of the patients underwent initial surgical resection on at least one kidney and about half of the patients had at least one nephrectomy performed as part of the surgical management.

Several smaller, single institution experiences were reported around this time, also supporting the benefits of treating children with BWT with chemotherapy prior to definitive surgical intervention.^11-17 However, most of the patients in these series underwent total nephrectomy on the more involved side with nephron-sparing surgery reserved for the less involved, contralateral side. In 1996, Horwitz reported the NWTS-4 experience with 98 patients.¹⁸ Four-year survival was 82% and the rate of local recurrence was only 8.2%. But despite finding that the presence of positive margins following NSS did not result in a higher incidence of local recurrence, over half of the patients underwent nephrectomy on at least one side. An unacceptably high incidence of renal failure was associated with this approach; in part as a consequence of aggressive surgery, 12% of patients with BWT treated on NWTS studies 1-4 developed renal failure.¹⁹ Although the relative contributions of nephrotoxic adjuvant therapy and intrinsic renal disease are not certain, the most common reason for renal failure was surgical removal of nephrons, in some cases bilateral nephrectomies.²⁰ Despite this, a later, comprehensive review of 188 patients treated on NWTS-4 by Hamilton revealed that over 40% of patients still had upfront surgery on at least one kidney and less than 40% of the patients had bilateral NSS.²¹ Several smaller, contemporary studies have revealed the performance of bilateral NSS in 22-78% of cases of BWT.^4,22-28

We previously reported our early experience in managing patients with BWT with bilateral NSS, showing preserved short-term renal function and excellent overall survival.^29,30 Here we describe a larger experience in managing 42 patients with an aggressive surgical approach to nephron-sparing surgery in an effort to improve long term renal function in children treated for BWT, while maintaining excellent oncologic outcomes.

Patients and Methods

This study was approved by the St. Jude Children's Research Hospital (SJCRH) institutional review board. We identified 42 consecutive patients presenting with a diagnosis of synchronous BWT who received surgical management of their disease at SJCRH between 2001 and 2014. Patients with metachronous lesions were excluded. Patient demographics (age, gender, race, and height), tumor characteristics (histology, stage after NSS³¹, margin status), adjuvant therapies, operative findings, and postoperative clinical course (including complications) were recorded. All patients, except two, received pre-operative chemotherapy consisting of vincristine, dactinomycin, and doxorubicin. The remaining patients received only vincristine and dactinomycin or vincristine and doxorubicin prior to surgery. The estimated GFR (eGFR) was calculated using the modified Schwartz equation for children.³² Patients were considered hypertensive if they had repeated systolic or diastolic blood pressure readings exceeding the 95^th percentile, or pre-hypertensive if the blood pressure was between the 90^th to 95^th percentiles recorded at more than one outpatient visit, or if anti-hypertensive medication was prescribed for blood pressure control.

We have previously described our surgical approach for NSS in patients with bilateral Wilms tumor.^30,33 Briefly, we perform bilateral nephron-sparing procedures during a single operation. The kidney with the larger tumor burden is approached first and completely mobilized. The renal vessels are identified; manual compression is occasionally utilized during the procedure. The renal capsule is scored around each tumormass, with each lesion being removed with a surrounding margin of normal parenchyma when possible, with blunt dissection and electrocautery. The collecting system is carefully inspected for entry, with disruptions being closed with fine, absorbable suture. Flank Penrose drains are frequently placed to ensure adequate drainage in patients at high risk for urine leak.

Results

Patient Characteristics

Patient and disease characteristics for the 42 patients are summarized in Table 1. Thirty-five (83.3%) were diagnosed prior to age 4, with the median age at diagnosis being 2.2 (range: 0.3-10.8) years. Two patients were also diagnosed with Beckwith-Wiedemann Syndrome (BWS), one had Wilms tumor-Aniridia-Genitourinary anomalies-Mental retardation (WAGR) Syndrome, three had cryptorchidism, and one had isolated hemihypertrophy. Seven (16.7%) patients initially presented with metastatic disease, limited to the lungs in each case.

Table 1. Patient characteristics.

Characteristics	No. of patients (%)
Sex
Female	23 (54.8)
Male	19 (45.2)
Race
Caucasian	31 (73.8)
African American	4 (9.5)
Other	7 (16.7)
Age at diagnosis (years)
< 1	12 (28.5)
1-4	23 (54.8)
> 4	7 (16.7)
Metastases at presentation
No	35 (83.3)
Yes	7 (16.7)
Associated conditions
Cryptorchidism	3 (7.1)
Beckwith-Wiedemann Syndrome	2 (4.8)
WAGR Syndrome	1 (2.4)
Isolated hemihypertrophy	1 (2.4)
Pathologic Findings	No. (%)
Local tumor stage by kidney
Upfront biopsy	15 (17.8)
Stage I	45 (53.6)
Stage II	10 (11.9)
Stage III	14 (16.7)
Tumor histology at initial surgery
Bilateral favorable histology	32 (76.2)
Unilateral focal anaplasia, contralateral favorable histology	5 (11.9)
Unilateral diffuse anaplasia, contralateral favorable histology	3 (7.1)
Bilateral diffuse anaplasia	2 (4.8)

Abbreviations: WAGR – Wilms tumor-Aniridia-Genitourinary anomalies-Mental retardation

Surgery

All 42 patients underwent their definitive operation(s) at SJCRH. Sixteen (38.1%) patients received all of their therapy at SJCRH, while 26 (61.9%) patients were referred to SJCRH for surgical management only, with neoadjuvant and post-operative therapy being given at the referring hospital. Thirty-nine (92.9%) patients underwent bilateral NSS; 3 (7.1%) patients underwent unilateral nephrectomy with contralateral NSS. Of these three patients, one patient had WAGR syndrome and showed tumor progression on one side while receiving neoadjuvant chemotherapy. The second patient had BWS and a large central tumor with extension into the ureter. The third patient had a WT-1 mutation and underwent unilateral nephrectomy due to parent preference.

Nine (21.4%) patients underwent at least one repeat NSS a median of 17.5 months (range: 1.6-37.4) after the initial surgery. Three of these patients had repeat NSS within 4 months of their initial procedure and were felt to have had residual tumor rather than recurrent disease. Overall, seven patients (16.7%) had recurrent disease 4.7 – 37.4 months after the initial surgery. Six of these patients underwent repeat NSS for local tumor recurrence. The seventh patient with recurrent disease in a kidney underwent a unilateral completion nephrectomy 4.7 months after the index procedure due to rapidly progressive disease. Four of the nine patients who required repeat NSS underwent a third surgical procedure – one for a retroperitoneal recurrence, one for liver recurrence, and two for a second repeat NSS for recurrent local disease (one in the same kidney that had undergone the prior repeat NSS, the other in the contralateral kidney).

Pathology

Fifteen of 84 kidneys, in 9 patients, underwent initial biopsy with 7 of these patients having had biopsies performed prior to referral to SJCRH. Of the 69 kidneys not subjected to pre-chemotherapy biopsy, 45 (53.6%), 10 (11.9%), and 14 (16.7%) kidneys were local stage I, II, and III, respectively, by post-operative pathological review. Tumor histology at initial surgery is listed in Table 1. Three patients who had blastemal-predominant favorable histology (FH) WT at initial operation had tumor with diffuse anaplasia at the time of repeat NSS (2 patients) or completion nephrectomy (1 patient). Twenty-eight (66.7%) patients had lymph node sampling performed at the initial surgery, with only three (7.1%) patients being found to have positive lymph nodes. Thirteen (31.0%) patients had at least one surgical specimen with a positive margin at initial operation; three more patients had a specimen with a positive margin at a subsequent surgery. Three of these 13 patients had diffuse anaplastic histology in at least one resection specimen.

Adjuvant Therapy

Forty (95.2%) patients received a combination of vincristine, dactinomycin, and doxorubicin, while 15 (34.9%) patients also received some additional combination of histology-tailored post-operative chemotherapy consisting of cyclophosphamide, ifosfamide, etoposide, or carboplatin. Eighteen (42.9%) patients received radiation therapy (RT) post-operatively to the abdomen for positive margins (n=14), anaplasia (n=5), or positive lymph nodes (n=3). Twelve (66.7%) of these patients received unilateral flank irradiation while six (33.3%) received bilateral flank irradiation. One patient with bilateral positive margins did not receive radiation because of an underlying diagnosis of Fanconi anemia. Two patients received whole-abdomen irradiation – one due to progressive, diffuse anaplasia after completion nephrectomy and the other due to disease recurrence. In addition to the 18 patients who received abdominal irradiation due to their post-operative pathologic findings, six other patients received abdominal irradiation for recurrent or progressive disease. Lastly, two (4.8%) patients also received palliative abdominal radiotherapy with 25 Gy due to rapidly progressive tumor growth – one of which was extending into the spinal column.

Of the seven patients who presented with metastatic pulmonary disease, five were treated with whole lung irradiation (total 12 Gy). The other two patients were treated with either chemotherapy alone or metastatectomy in addition to chemotherapy.

Complications

There were 24 acute post-operative complications in 16 (38.1%) patients. Complications included prolonged urine leak (n=10, of which 7 required stent placement or manipulation), infection (n=6), intussusception (n=2) and transient renal insufficiency (n=1). Three of the patients with prolonged urine leak developed a urinoma requiring percutaneous drainage. Six hospital-acquired infections in five patients were observed. Four patients had urinary tract infections (UTI) and two patients developed Clostridium difficile enterocolitis; one patient had both a UTI and a C. difficile infection. Two patients exhibited symptoms of a small bowel obstruction 5-6 days after surgery and underwent exploratory laparotomy for ileo-ileal intussusception. In neither case was bowel viability found to be compromised. The patient with transient renal insufficiency required continuous veno-venous hemodialysis (CVVHD) for less than 30 hours following surgery. One patient required re-intubation on post-operative day four due to fluid overload. On long-term follow-up, three (7.1%) patients had an episode of intestinal obstruction. Two patients underwent exploratory laparotomy for lysis of adhesions; one patient required a small bowel resection.

Oncologic Outcomes

The 3-year event-free survival in our series was 64.0% while the overall survival for the entire cohort of patients was 85.7%, with a mean follow-up of 4.1 years. The six deaths occurred a median of 11.4 (range: 2.1-14.7) months from the last surgical resection (Table 2). Two of these patients initially presented with diffuse anaplasia in at least one resected tumor specimen, while an additional three developed diffuse anaplasia at recurrence. The patient who had only favorable histology WT documented in all resection specimens had positive lymph nodes at the initial operation. Seven patients had metastatic disease at presentation (all pulmonary); five (71.4%) of these patients survived. One patient with a germ line BRCA2 gene mutation died of a second malignancy (medulloblastoma) 16 months after the diagnosis of WT, without evidence of residual/recurrent WT. One other patient also developed a medulloblastoma as a second malignancy but is still alive 1.7 years after completing therapy for his brain tumor.

Table 2. Deaths.

Patient	Age at diagnosis (years)	Metastasis at presentation	Surgical Procedure(s)	Lymph Node Status	Histology	Local Tumor Stage at Index Operation	Time from Last Surgery to Death (months)	Associated Anomalies
1	4.1	Pulmonary	B NSS	Positive	B diffuse anaplasia	B previous biopsies	2.1	None
2	0.6	None	B NSS Re-do right NSS Resection of liver recurrence	Not sampled	Favorable Unilateral diffuse anaplasia Diffuse anaplasia	B stage III due to positive margins	3.2	None
3	0.9	None	B NSS	Negative	Unilateral diffuse anaplasia	B stage III due to positive margins	13.5	BRCA2 gene mutation, medulloblastoma (cause of death)
4	0.4	None	B NSS Re-do right NSS Re-do left NSS	Not sampled	Favorable Favorable Unilateral diffuse anaplasia	B stage I	9.4	Beckwith-Wiedemann
5	2.4	Pulmonary	B NSS	Positive	Favorable	B stage III due to positive margin and lymph node involvement	13.4	None
6	1.2	None	B NSS Completion left nephrectomy	Negative	Favorable Unilateral diffuse anaplasia	B stage I	14.7	None

Abbreviations: B – bilateral, NSS – nephron-sparing surgery.

Renal Function Outcomes

Long term renal function was assessed in the 36 surviving patients with a median age at follow-up of 6.7 years (range: 1.3-14.6 years) with median follow-up at the time of renal function assessment of 3.7years (range: 0.03-13.4 years) (Table 3). Eleven (30.6%) patients were on anti-hypertensive medication at last follow-up. An additional two (5.6%) patients were noted to have persistent systolic or diastolic blood pressure readings which exceeded the 95^th percentile for their age group. Five (13.9%) other patients not taking anti-hypertensive medications were noted to have persistent systolic or diastolic blood pressure readings between the 90^th and 95^th percentile for their age group, and are thus considered pre-hypertensive. Five (13.9%) patients were found to have persistent ≥1+ proteinuria by urinalysis; four of these five patients were taking anti-hypertensive medications.

Table 3. Follow-up Characteristics.

Characteristics	No. of patients (%)
Status
Alive	36 (85.7)
Deceased	6 (14.3)
Oncologic outcome
Residual tumor requiring repeat NSS	3 (8.3)
Recurrent tumor requiring repeat NSS	6 (14.3)
Recurrent tumor requiring completion nephrectomy	1 (2.4)
Characteristics for living patients*	No. of patients (%)
Time from diagnosis to last visit (years)
<1	7 (19.4)
1-5	16 (44.5)
5-10	10 (27.8)
>10	3 (8.3)
Taking medication for hypertension
Yes¥	11 (30.6)
No	25 (69.4)
Persistent ≥1+ proteinuria
Yes	5 (13.9)
No	31 (86.1)
Estimated GFR by Schwartz formula (mL/min/1.73m2)
< 60	0
60 – 89	13 (36.1)
> 90	23 (63.9)

^*N=36 living patients;

^¥Two additional patients with elevated systolic blood pressures are not yet on anti-hypertensives. Abbreviation: NSS – nephron-sparing surgery; GFR – glomerular filtration rate.

All patients were found to have an estimated GFR >60 mL/min/1.73m² as determined by the modified Schwartz formula.³² Thirteen (36.1%) patients, including one of three who had initial unilateral nephrectomy, had an eGFR between 60-90 mL/min/1.73m², with 3 of these patients also having persistent ≥1+ proteinuria. Thus, overall, 8.3% of living patients in this series have stage 2 chronic kidney disease (CKD), as defined by eGFR of <90 mL/min/1.73m² and persistent ≥1+ proteinuria.³⁴

Discussion

The focus of many current clinical trials for treating children with tumors of good prognosis has shifted to decreasing the intensity of therapy to avoid long-term complications, while maintaining excellent oncologic outcomes. This is particularly true for favorable histology Wilms tumor for which the overall survival is now >90%.¹ Although the outcome for bilateral favorable histology Wilms tumor is not as high, the overall survival is still >80%, and so the consequences of therapy are still relevant, with the development of renal dysfunction being among the greatest concerns. Thus, the role for nephron-sparing surgery for children with BWT has become increasingly important, yet bilateral NSS is still not widely performed. We published an early experience in which ten patients with synchronous BWT underwent bilateral NSS and concluded that bilateral NSS was safe and feasible in nearly all cases and maintained the good oncologic outcome. In this study, we have evaluated the functional and oncologic outcomes in a much larger series of patients, all of whom underwent definitive surgery at our institution since 2001 when we began our consistent, aggressive approach to nephron-sparing surgery in this patient population.

In this larger series of 42 patients, we were able to perform bilateral NSS in 39 (92.9%) patients. This percentage of successfully completed bilateral nephron-sparing procedures includes a large percentage of patients (61.9%) who were referred just for surgery from other institutions because of complex anatomy. Thus, these results confirm our earlier reported hypothesis that nearly all BWT lesions can be removed with a nephron-sparing approach despite occasionally foreboding pre-operative imaging.³⁰ Even central lesions can be successfully and safely removed using the method of longitudinal partial nephrectomy described by Fuchs.³⁵ The consistent use of three-drug chemotherapy prior to surgery in our series may have contributed to the greater ability to perform nephron-sparing surgery.

We have modified some of the technical aspects of our approach to nephron-sparing surgery during the past fifteen years. Several patients early in our experience had residual disease detected on the first imaging study performed after surgery, generally 3 months after the nephron-sparing procedure. These patients typically had very large lesions prior to surgery that obscured visualization on the pre-operative imaging studies of additional smaller lesions that were missed at the time of surgery. Because of this we now routinely obtain intra-operative ultrasound of the residual renal parenchyma following tumor resection and have detected, in two cases, residual lesions that were found to contain viable WT on histologic examination after resection. We also had a fairly high incidence of post-operative urine leak. These leaks were often managed by cystoscopic ureteral stent placement or manipulation; no leak required open re-exploration. However, because of these leaks early in our experience when stents were often placed at initial surgery, we now rarely place internal stents, simply closing defects with a running monocryl suture and, more recently, placing TachoSil® over the cut surface of the kidney or where the kidney has been folded over on itself. We have also continued our aggressive approach to renal preservation even after local recurrence³⁶, performing repeat NSS in six of seven patients who recurred in a kidney. Survival for these patients, despite multiple repeat nephron-sparing procedures, was still 77.8%.

Many of the lesions removed were enucleated in order to preserve renal parenchyma. Although not an acceptable oncologic approach for other histologic tumor types, and although some have expressed concern about its use in WT²⁴, we have found that the surgical margins are usually negative, despite performing a marginal resection. This appears to be due to the fact that Wilms tumors tend to grow within a surrounding fibrous capsule that does not contain malignant cells.³⁷ We have also found that the presence of a positive margin does not predict a poorer outcome for patients, consistent with the earlier results in NWTS-4, reported by Horwitz.¹⁷ In our series, 16 (38.1%) patients had at least one specimen with a margin involved with tumor, yet the incidence of local recurrence and overall survival was 18.8% and 75%, respectively, in these patients, similar to the group as a whole; one of the patients died of a secondary malignancy without evidence of WT. However, we have generally given flank irradiation to those patients with a positive resection margin.

The factor that remains the most important predictor of survival is tumor histology. Eight (19.0%) patients in our series had diffuse anaplasia in one or more resected specimens. Their overall survival was only 37.5%, with all but one patient in our series who died of disease having anaplastic histology. Interestingly, three of these patients had blastemal predominant, favorable histology initially but ultimately had anaplastic histology at relapse. These patients relapsed despite being of local stage I in two patients after their initial nephron-sparing procedures. The third patient had local stage III due to bilateral positive margins. The best management of children with BWT and tumor(s) of diffuse anaplasia is unclear given that their outcome is so poor. One patient with metastatic anaplastic BWT in our series underwent bilateral NSS despite infiltrative local disease in an effort to maintain some degree of quality of life before she succumbed. Others had negative margins in all specimens containing diffuse or focal anaplasia and so completion nephrectomy was not undertaken as we would have done, had positive margins been encountered. Although spilling of anaplastic tumor cells as a consequence of having positive margins when attempting NSS should be avoided for this chemotherapy-resistant histology, histology is rarely known prior to performing surgery because there are no pathognomonic imaging signs predictive of anaplasia and tumor biopsies have been shown to have a poor ability to detect anaplasia.³⁸ The impact of histology on outcome is so profound (89% [FH] vs 45% [UH] in NWTS-4, 96.6% vs 61.5% in our study) that the incidence of patients with anaplasia must be considered when evaluating results from smaller series. For example, Cooper reported an overall survival of only 56.5% in patients with BWT but 43.5% of patients had unfavorable histology WT in that series.³⁹

The greater number of patients and longer follow-up for many patients in this series has allowed us to look more critically at renal function outcomes. In our cohort, no patient has developed end-stage renal disease requiring ongoing renal replacement therapy or renal transplantation; only one patient required transient renal replacement therapy of less than 30 hours duration following bilateral NSS. Go has suggested a correlation of declining renal function and cardiovascular events leading to an increase in mortality.⁴⁰ Therefore, we have evaluated our patients looking for more subtle signs of renal dysfunction and found that only 3 of 36 (8.3%) surviving patients had stage 2 CKD, as defined as an eGFR of less than 90 mL/min/1.73m² and persistent proteinuria.³⁴ Thirteen (36.1%) patients have hypertension (all but 3 patients require only a single oral agent to remain normotensive), and 5 (13.9%) have persistent proteinuria. This incidence of hypertension is consistent with previous reports in smaller cohorts.⁴¹ Moreover, both chemotherapy and radiation have known effects on kidney function, as well as urinary tract infections (as occurred in 9.5% of our patients) making it difficult to isolate the effects of nephron loss from those induced independently or synergistically by nonsurgical therapy. Also critical is the impact of specific glomerulopathies that can occur in Denys–Drash (diffuse mesangial sclerosis) and WAGR (focal segmental glomerulosclerosis) syndromes, for which the incidence of renal failure after the diagnosis of BWT is 50% and 90%, respectively.^18,42,43 When comparing rates of renal insufficiency in different series of patients with BWT, one needs to consider the number of patients with these syndromes as they can significantly impact the incidence of renal failure in a patient cohort. In addition, longer follow-up of the patients in our series will be required to more fully assess the incidence and degree of renal dysfunction in this group of patients, as this process is likely progressive, and to identify risk factors for progressive renal dysfunction. Close monitoring of renal function in these patients to allow for early intervention and prevention of these sequelae will be important in the care of these patients.

Thus, bilateral nephron-sparing surgery in children with BWT is almost always feasible and can be done safely and with good oncologic outcomes. Sparing of renal parenchyma is likely to help preserve renal function in these children who are at significant risk for chronic renal insufficiency although further careful long-term follow-up is still required to fully assess the potential progression of renal dysfunction.