Fax: (901) 495-3766
Local lymph node involvement does not predict poor outcome in pediatric renal cell carcinoma
Article first published online: 26 AUG 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 7, pages 1575–1583, 1 October 2004
How to Cite
Geller, J. I. and Dome, J. S. (2004), Local lymph node involvement does not predict poor outcome in pediatric renal cell carcinoma. Cancer, 101: 1575–1583. doi: 10.1002/cncr.20548
- Issue published online: 17 SEP 2004
- Article first published online: 26 AUG 2004
- Manuscript Accepted: 23 JUN 2004
- Manuscript Revised: 21 JUN 2004
- Manuscript Received: 19 APR 2004
- National Institutes of Health. Grant Numbers: T32-CA70089, CA-21765
- American Lebanese Syrian Associated Charities of St. Jude Children's Research Hospital.
- renal cell carcinoma;
- lymph node;
Local lymph node involvement in adults with renal cell carcinoma (RCC) is associated with poor outcome. The prognostic significance of local lymph node involvement in children with RCC has not been studied systematically.
A retrospective review of patients treated at St Jude Children's Research Hospital (Memphis, TN) and an extensive review of the medical literature were undertaken to evaluate the prognostic significance of local lymph node involvement in pediatric RCC.
Thirteen patients with the diagnosis of RCC were treated at St. Jude since the hospital's inception in 1962. Four patients presented with lymph node-positive, distant metastasis-negative (N+M0) disease, and all 4 remain disease free after resection without adjuvant therapy (follow-up duration, 2–9 years). A systematic review of the literature including 243 pediatric patients with RCC revealed stage-specific survival rates of 92.5%, 84.6%, 72.7%, and 12.7% for Stage I–IV disease, respectively. Of 58 children with N+M0 RCC for whom outcome data were available, 42 (72.4%) were alive without disease at last follow-up. Among patients whose therapy could be discerned, those who received no adjuvant therapy fared as well (15 of 16 alive) as those who received various adjuvant treatments (22 of 31 alive).
Children with lymph node–positive RCC in the absence of distant metastatic disease had a relatively favorable long-term prognosis, with survival rate nearly triple those of adult historical controls. Until highly effective therapies for RCC are identified, these children should not be exposed to adjuvant treatment. Further investigation of the biologic differences between adult and pediatric RCC is warranted. Cancer 2004. © 2004 American Cancer Society.
Renal cell carcinoma (RCC) is predominantly an adult cancer, with a median age at diagnosis of 65 years.1, 2 The stage-specific, 5-year overall survival approximates 85–94%, 84–90%, 61–63%, and 11–28% for patients with International Union Against Cancer (UICC)/American Joint Committee on Cancer (AJCC; TNM) Stage I–IV disease, respectively.3–5 One of the most important adverse prognostic factors for adult RCC is local lymph node metastasis. Patients with tumor involvement of local lymph nodes have 5-year survival estimates of approximately 20% compared with survival estimates of 80% for patients without lymph node involvement.5, 6 Lymph node–positive disease remains significantly associated with poor prognosis in multivariate analyses that adjust for other known prognostic factors including the presence of distant metastases, tumor size, tumor stage, and disease grade.5–7 Given the unfavorable prognosis conferred by lymph node positivity, recent clinical trials have evaluated the use of adjuvant immunotherapy for patients with local lymph node involvement. Such therapies have failed to prove effective in preventing disease recurrence, but efforts to improve adjuvant therapy regimens continue.8, 9
RCC in childhood is rare. Surveillance, Epidemiology, and End Results program data collected from 1975 to 2000 demonstrate that pediatric RCC comprises 5.9% of renal malignancies in children and adolescents.10 In the pediatric population, the median age at diagnosis is 9 years and at age 15 years, RCC becomes as common as Wilms tumor,11 but no prospective studies have been performed.12–16 Children with RCC tend to have a similar overall prognosis when compared with adults, with one large study recently reporting a 54.9% overall survival rate at 20 years.17 As seen in adult RCC, prognosis worsens with increasing stage, although direct comparisons of adult and pediatric data are confounded by the finding that most reviews of pediatric RCC used the modified Robson staging system rather than the TNM system.18
In the past 5 years, we have treated 3 patients with RCC and local lymph node involvement. Because the optimal management for this group of patients is unclear, we reviewed the institutional experience with pediatric RCC at St. Jude Children's Research Hospital (Memphis, TN). In addition, we conducted a systematic review of the literature to characterize outcome data for patients with pediatric RCC, with the specific aim of evaluating the prognostic significance of local lymph node involvement.
MATERIALS AND METHODS
Institutional review board approval was obtained for a retrospective review of pediatric patients with RCC diagnosed at St. Jude Children's Research Hospital since the institution's inception in 1962, with the specific intent to extract the following data: predisposing factors, presenting symptoms and signs, gender, race, sites of disease (enabling TNM and modified Robson staging), disease histology, treatment, and outcome. No patient identifiers were extracted during the chart review process in accordance with the Health Information Portability and Accountability Act as well as with good clinical research practice.
For the systematic review of the literature, Medline (PubMed) searches were undertaken to identify all cases of pediatric RCC (age < 21 years of age) desribed in the English literature during or after 1974. This cutoff date was chosen due to the presence of an adequate retrospective review encompassing published cases before that time point, as well as the desire to select a cohort assessed and treated during a more contemporary time period.12 PubMed search phrases included, but were not limited to, pediatric renal cell carcinoma, childhood renal cell carcinoma, and adolescent renal cell carcinoma. For the purpose of formulating approximate overall outcome data in pediatric RCC according to stage, a modified Robson stage was applied to each case as previously described.18 Briefly, Stage I disease represents localized disease confined by the renal capsule, Stage II suggests invasion of the renal capsule but the tumor is confined by Gerota's fascia, Stage III involves either invasion of the renal vein or inferior vena cava, or invasion of local lymph nodes, and Stage IV disease is defined by the presence of distant metastases. Cases without adequate staging information or follow-up, or publications reporting patients with underlying conditions such as tuberous sclerosis or previous malignancy, were excluded. In the rare situation when only a fraction of the cases reported by a single publication met all criteria necessary to be included in this review, only the cases with complete data were included.19 For the purpose of constructing Table 1 (which summarizes lymph node–positive, distant metastasis–negative [N+M0] cases), only publications reporting pediatric RCC cases in which information enabled accurate NM staging and assessment of outcome were included. TNM staging was applied in accordance with the 1997 UICC/AJCC standard TNM staging for RCC.3
|Authors||Year||No. of patients||No. with NED||No. AWD|
|Indolfi et al.17||2003||9||5||1|
|Huang et al.25||2001||1||1||—|
|Uchiyama et al.26||2000||2||1||1|
|Asanuma et al.28||1999||2||2||—|
|Chen et al.29||1999||1||0||—|
|Androulakakis et al.30||1999||2||1||—|
|Carcao et al.18||1998||6||6||—|
|Al-Saleem et al.33||1998||1||1||—|
|Freedman et al.35||1996||3||3||—|
|MacArthur et al.39||1994||1||0||—|
|Eckschlager and Kodet22||1994||1||1||—|
|Fenton et al.57||1993||1||1||—|
|Kabala et al.41||1992||2||1||—|
|Bruce and Grugh42||1990||1||0||—|
|Malaga et al.43||1990||1||1||—|
|Taylor et al.58||1989||1||1||—|
|Lack et al.47||1985||4||2||—|
|Chan et al.21||1983||5||3||—|
|Laurenti et al.51||1980||1||1||—|
|Hicks et al.56||1975||1||1||—|
|Cassady et al.59||1974||2||2||—|
|Castellanos et al.12||1974||2||2||—|
|Total||58||42 (72.4%)||3 (5.2%)|
Pediatric Renal Cell Carcinoma: The St. Jude Children's Research Hospital Experience
Thirteen children were diagnosed with RCC at St. Jude between 1962 and February 2004, accounting for 2.2% of all primary renal malignancies treated at the institution. Patient characteristics, disease characteristics, and patient outcomes are summarized in Table 2. The median and mean age of presentation were 12 years and 10.6 years, respectively. The male-to-female ratio approximates 1 (7:6), consistent with other reviews of pediatric RCC and in contrast to the typical male predominance seen in adult RCC. Presenting symptoms and signs in decreasing frequency were hematuria (77%), flank pain (61.5%), anemia (46%), fever (30.8%), abdominal trauma (15.4%), and abdominal fullness—described by only 1 patient (7.7%). No patient demonstrated significant hypertension, abnormal renal function, or abnormal calcium levels. Although five patients demonstrated a combination of hematuria and flank pain at presentation, no patient demonstrated the classic presenting triad of abdominal pain, hematuria, and a palpable abdominal mass. In fact, an abdominal mass was not documented by physical examination in any patient. No patient was diagnosed with or had a significant family history for von Hippel–Lindau syndrome, tuberous sclerosis, or any other condition associated with RCC.
|Patient no.||Age (yrs)||Gender||Race||Presenting symptoms and signs||Metastases||Histology||MR stage||TNM status||Therapy||Response||Outcome (FU)|
|1||13||F||C||Hematuria, pain anemia||Lung, liver, omentum||Clear cell||IV||Stage IV, T3bN0M1||N, CT + RT||PD||DOD (2 mos)|
|2||10||M||A||Abdominal trauma, pain, fever, anemia||None||Papillary||I||Stage I, T1N0M0||N||CR||NED (23 yrs)|
|3||12||M||A||Hematuria, pain, anemia||None||Clear cell||IIIa||Stage III, T3bN0M0||N||CR||NED (20 yrs)|
|4||8.5||M||A||Pain, fever||None||Clear cell||II||Stage III, T3bN0M0||N||CR||NED (16.5 yrs)|
|5||5.5||M||C||Hematuria||Lymph node||Clear cell||IIIb||Stage III, T1N1M0||N||CR||NED (9+ yrs)|
|6||10||M||A||Hematuria||Lymph node, lung||Clear cell||IV||Stage IV, T3aN1M1||N||PD||DOD (5 weeks)|
|7||17||M||A||Hematuria, pain,||Lung||Neuroendocrine||IV||Stage IV, T3bNxM1||CT + RT||PD||DOD (3 weeks)|
|fever, cachexia, anemia|
|8||7||F||A||Abdominal trauma, hematuria, fever, anemia||None initially; lung recurrence||Clear cell||II||Stage III, T3aN0M0||N, IT + CT, CT||Recurrence, PD, SD||AWD (4.7 yrs)|
|9||12||F||C||Pain, anorexia||Lymph node||Clear cell||IIIb||Stage IV, T3aN2M0||N||CR||NED (4.4 yrs)|
|10||13||F||C||Hematuria, pain||Lymph node||Papillary||IIIb||Stage III, T3aN1M0||N||CR||NED (2.6 yrs)|
|11||16||F||A||Hematuria||None||Clear cell||I||Stage II, T2N0M0||N||CR||NED (1.5 yrs)|
|12||7||M||C||Hematuria||Lymph node||Oncocytoid/histiocytic features||IIIb||Stage IV, T2N2M0||N||CR||NED (2 yrs)|
|13||17||F||A||Hematuria, pain, fatigue, anemia||None||Papillary||I||Stage I, T1N0M0||N||CR||NED (2 yrs)|
The primary tumor occurred in the right kidney in nine children and in the left kidney in four children. No patient demonstrated bilateral disease. Clear cell histology was found in eight children, papillary histology in three children, neuroendocrine RCC in one child, and oncocytoid RCC with histiocytic features in one child. Three patients had distant metastatic disease and four patients had local lymph node involvement at diagnosis. The stage distribution according to the TNM system was as follows: Stage I in two patients, Stage II in one patient, Stage III in five patients, and Stage IV in five patients. The stage distribution according to the modified Robson system18 was as follows: Stage I in three patients, Stage II in two patients, Stage III in five patients, and Stage IV in three patients. These findings demonstrate the typical downward shift in disease stage that occurs when the modified Robson system (as opposed to the TNM system) is used.
Patient 7, a 17-year-old male with neuroendocrine RCC, presented with severe cachexia (ill appearing, 18 kg weight loss, anorexia, fever, malaise, anemia), flank pain, multiple gastrointestinal and bronchial ulcerations, respiratory distress, hemoptysis, hematemesis, and hematuria (normal coagulation profile). Because he was not deemed stable enough at presentation for nephrectomy, a fine-needle aspiration biopsy was performed, enabling characterization of the tumor. Histologically, the tumor specimen demonstrated a mitotically active acidophilic undifferentiated carcinoma with rhabdoid features and neuroendocrine elements. Tumor invaded the renal vein. Immunohistochemistry showed diffuse positive staining for neural markers such as synaptophysin (+3), chromogranin (+2), and neuron-specific enolase (+3), as well as typical epithelial markers such as vimentin, desmin, and cytokeratin. Cytogenetic studies of the tumor revealed a nonspecific abnormality in chromosome 1. Serum gastrin levels were elevated at 332 pg/mL (normal range, 10–100 pg/mL) and urine VMA levels were elevated at 7.8 mg/g creatinine (normal, < 6 mg/g). Levels of serum somatomedin C (258 ng/mL) and corticotropic hormone (41 pg/mL) were normal. Urine cytology was positive for malignancy, and imaging of the right lower lung was strongly suggestive of metastatic disease. Patient 7 received 360 centi-Gray (cGy) of pulmonary radiotherapy followed by 1 course of ICE/PIXY (ifosfamide, carboplatin, and etoposide followed by granulocyte-macrophage–colony-stimulating factor and interleukin-2 [IL-2]), but died < 3 weeks after diagnosis due to multiorgan failure. As far as we are aware, this is the youngest patient to present with the rare entity known as primary renal neuroendocrine carcinoma, and demonstrates the classic aggressive behavior of this disease.
For all other patients, excluding Patient 7, treatment included a nephrectomy and various degrees of lymph node dissection. A surgical complete response was obtained in 9 of 12 (75%) patients for whom surgery was attempted. Chemotherapy combined with radiotherapy was given to one patient with Stage IV disease at presentation after nephrectomy, with no response appreciated (Patient 1). Patient 8 was treated initially with surgery only, but had disease recurrence 33 months after diagnosis with multiple small pulmonary nodules evident throughout both lungs (biopsy-proven metastasis). After disease progression after a 6-month treatment regimen employing IL-2, interferon-α, and 5-fluorouracil,20 this patient enrolled in a Phase I study of oral irinotecan (CPT-11). She currently receives CPT-11 and has had stable disease for 14 months. She has not suffered from any side effects from the CPT-11 therapy.
Overall, 9 of 13 (69.2%) patients were alive and disease free at their most recent follow-up visits, with 1 of 13 patients alive with stable disease. The mean and median follow-up in surviving patients were 8.6 and 4.6 years, respectively (range, 1.5–23 years). Of the 10 patients alive at last follow-up, 4 were followed for > 9 years and 6 were seen in our clinic in the last 2 years and continue to be monitored regularly with a mean and median follow-up of 2.9 and 2.3 years, respectively (range, 1.5–4.7 years). All three patients with papillary RCC were alive with no evidence of disease at last follow-up. Among the 8 patients with clear cell RCC, 5 were alive with no evidence of disease, 2 died of disease, and 1 remains alive with disease 4.7 years from initial diagnosis. All three patients with hematogenous metastasis at diagnosis died of disease. It is noteworthy that all four patients who presented with local lymph node involvement in the absence of distant metastases were alive without evidence of disease at their most recent follow-up visits (Patients 5, 9, 10, and 12).
Prognostic Significance of Local Lymph Node Involvement in Pediatric Renal Cell Carcinoma
The observation that all four patients with lymph node–positive disease without distant metastases were alive and disease free raised the possibility that lymph node involvement is not associated with poor outcome in pediatric RCC. To evaluate the prognostic significance of local lymph node involvement, we conducted a systematic review of the literature. We included 243 cases of pediatric RCC in the analysis of stage-specific outcomes. Stage I was the most frequent presentation, followed closely by Stage IV and Stage III disease.12–14, 17, 18, 21–56 The stage-specific overall survival rates were > 90%, > 80%, > 70%, and < 15% for Stages I–IV, respectively (Table 3). The overall survival rate for this entire retrospective cohort of patients was 63%. No association was noted between the administration of chemotherapy or radiotherapy and patient outcome. Several patients with Stage IV metastatic disease were cured by IL-2–based therapy.37, 39
|Authors||Year||No. alive/no. deada by modified Robson stage|
|Indolfi et al.17||2003||15/2||2/0||7/4||0/9|
|Shmushkevich et al.24||2002||1/0|
|Huang et al.25||2001||1/0|
|Uchiyama et al.26||2000||2/0|
|Ettinger et al.27||2000||1/0|
|Asanuma et al.28||1999||0/1||1/0||2/0|
|Chen et al.29||1999||2/1||0/1||1/0|
|Androulakakis et al.30||1999||1/0||2/0||1/1|
|Ray et al.31||1999||1/0|
|Fujii et al.32||1998||1/0|
|Carcao et al.18||1998||3/0||7/0||0/6|
|Al-Saleem et al.33||1998||1/0|
|Grant et al.34||1997||1/0|
|Freedman et al.35||1996||1/0||3/0||1/1|
|Aronson et al.36||1996||6/0||1/0||1/13|
|Bauer et al.37||1995||1/4|
|Kebudi et al.38||1995||0/1|
|MacArthur et al.39||1994||4/0||1/0||0/1||1/0|
|Eckschlager and Kodet22||1994||2/0||1/0||1/0|
|Marcus and Watt40||1994||2/0|
|Kabala et al.41||1992||1/0||1/1|
|Bruce and Gough42||1990||3/0||1/0||0/1||0/1|
|Malaga et al.43||1990||1/0|
|Mancini and Paolucci44||1989||1/0|
|Tsai et al.45||1988||1/0|
|Goto et al.13||1986||1/0|
|Lindop et al.46||1986||1/0|
|Lack et al.47||1985||6/0||2/1||2/2||0/3|
|Weiss et al.48||1985||0/1|
|Oesch et al.49||1983||1/0||0/1|
|Raney et al.19||1983||5/0||2/6|
|Chan et al.21||1983||3/0||1/1||3/4||1/4|
|Horn et al.50||1981||0/1|
|Laurenti et al.51||1980||1/0|
|Herschorn et al.52||1979||3/0||0/1||1/1||1/1|
|Abrams et al.53||1979||3/0||0/1|
|Futrell et al.54||1978||1/0|
|Fisher et al.55||1977||1/0||0/2|
|Hicks et al.56||1975||2/0|
|Castellanos et al.12||1974||2/0||2/1||0/1|
|Overall survival||153/243 (63%)|
Fifty-eight patients with N+M0 pediatric RCC, including the 4 patients reported in the current study, were identified.12, 14, 17, 18, 21–23, 25, 26, 28–30, 33, 35, 39–43, 47, 51, 56–59 In total, 77.6% of these patients were alive at the time of their case publication, and the majority of patients were alive and free of disease (72.4% overall; Table 1). Several authors suggested that radiotherapy might have had a positive impact on survival in these patients.17 To assess the impact of adjuvant therapy on survival, we analyzed the outcome for this cohort in relation to treatment intervention. Table 4 demonstrates a lack of correlation between any therapy and outcome. Among the patients clearly identified as having received some form of adjuvant therapy (radiotherapy, immunotherapy, and/or chemotherapy), 22 of 31 (71%) were reported to be alive without disease compared with patients who clearly received no adjuvant therapy, of whom 15 of 16 (93.8%) were reported to be alive without disease.
The experience at St. Jude Children's Research Hospital and an extensive review of the literature indicate that the adverse prognostic significance associated with local lymph node involvement in adult RCC is not duplicated in pediatric RCC. Castellanos et al.12 reviewed cases of pediatric RCC reported between 1934 and 1974. Of the 150 cases included in their analysis, 7 were identified as having local lymph node involvement in the absence of distant hematogenous metastasis (Robson Stage IIC). Six of those patients were reported to be alive without disease at their latest follow-up. Our updated analysis corroborates this finding. Of 58 patients with local lymph node involvement, 42 (72.4%) were alive without evidence of disease at last follow-up. This overall survival rate is nearly triple that reported for adult patients with local lymph node involvement in the absence of distant metastasis.5
The distinct clinical behavior found in children with lymph node–positive RCC may reflect biologic differences between adult and pediatric RCC. Evidence is accumulating that many pediatric patients with RCCs have histologic and cytogenetic features that are distinguishable from adult patients with RCCs. One distinct subtype of pediatric RCC is characterized by translocations involving chromosome Xp11.2,60 the locus of the TFE3 gene. The most prevalent fusion partners are the PRCC gene at chromosome 1q2161, 62 and the ASPL gene at chromosome 17q25.63, 64 Tumors with t(X;1)(p11.2;q21) translocations typically have a papillary architecture,60 whereas pseudopapillary or clear cell histology has been reported in tumors with the t(X;17)(p11.2;q25) translocation.64, 65 A systematic evaluation of Xp11 translocations in pediatric RCC is required to define the histologic spectrum of these tumors. It is noteworthy that although the t(X;17)(p11.2;q25) tumors seem to present at an advanced stage (Stage III disease with lymph node involvement or Stage IV disease), most patients were either cured or had prolonged survival with disease.64 Another distinct variant of RCC seen in the pediatric population is RCC with epithelioid morphology that stains positively for HMB45, but negatively for epithelial markers.66 The prevalence and clinical behavior of this variant, which is associated with translocation t(6;11)(p21.1;q12), have not been characterized, but we now know that the fusion product involves TFEB with the Alpha gene on chromosome 11.67 It has been reported that papillary histology is more common in pediatric RCC than clear cell histology,68, 69 although our experience and other case series indicate that clear cell histology is more prevalent.16–18, 23
Genetic susceptibility to cancer is typically associated with a young age at tumor presentation. Several heritable syndromes have been associated with RCC, including von Hippel–Lindau disease and tuberous sclerosis. One might predict a disproportionately high prevalence of such syndromes in children with pediatric RCC, but this does not appear to be the case. With the exception of tuberous sclerosis70 and a single publication reporting cases of pediatric RCC among patients with von Hippel–Lindau disease,68 pediatric RCC is not typically associated with heritable syndromes.71 It is noteworthy that nearly 20 cases of RCC have been reported to occur in children and young adults after neuroblastoma.72, 73 The absence of an obvious association with chemotherapy or radiotherapy and the observation that 31% of these RCCs were bilateral are suggestive of a genetic susceptibility in these patients.
The relatively favorable prognosis associated with regional lymph node involvement in pediatric RCC has important therapeutic implications. Whereas adjuvant therapy typically would be recommended for malignancies associated with a 70% survival rate, highly effective treatments for RCC remain elusive. The current front-line treatment for metastatic RCC is IL-2, either alone or in combination with IFN-α or other agents, but such regimens produce response rates of only 20% and durable survival in < 20% of patients.74 Two recent clinical trials demonstrated that adjuvant immunotherapy does not improve survival for adult patients with RCC with local lymph node involvement.8, 9 This lack of efficacy, combined with the relatively favorable outcomes for children with lymph node–positive RCC, suggests that adjuvant therapy is not indicated for these children. In contrast, the outcomes for children with Stage IV RCC are poor, with survival rates of < 15%. Two children with metastatic RCC were cured after treatment with high-dose IL-2, but this therapy was associated with significant toxicity.37, 39 Prospective studies to evaluate new treatment regimens for children with metastatic or recurrent RCC are warranted.
An issue worthy of brief discussion is the role of lymph node dissection in the up-front management of pediatric RCC. In adult patients with RCC, the role of lymph node dissection has been investigated extensively and remains controversial.75–78 Ebert et. al.79 suggested that lymph node dissection has a positive effect on the survival of children with RCC and that children with RCC may warrant more aggressive surgery. However, not many children with RCC in our review underwent lymph node dissection, perhaps because they were suspected of having Wilms tumor, the pediatric surgeons were unfamiliar with the potential role of lymph node dissection in RCC, or the benefit of lymph node dissection for RCC is unclear, even in the adult population. Based on our observation that the majority of children with lymph node–positive RCC survive (and the likelihood that some of these survivors did not have complete resection of their positive lymph nodes), we propose that second-look lymph node dissections are unnecessary for children who did not undergo lymph node dissections as part of their primary surgery. However, if suspicious lymph nodes are visualized on postoperative imaging studies, surgical resection is warranted.
In conclusion, children and adolescents with lymph node–positive RCC in the absence of distant metastases have a favorable prognosis, with overall survival estimates nearly triple those for adult patients with a similar presentation. Treatment decisions for these children should not be confined by approaches that are applied to adult patients. Specifically, children and adolescents with N+M0 RCC should not be exposed to adjuvant therapy until highly effective treatments are identified. It remains unclear whether the poor prognosis suggested by the presence of lymph node invasion in adults applies to young adults. Further investigation of the biologic differences between adult and pediatric RCC is warranted.
- 101975- 2000 [monograph online]. Available from URL: http://seer.cancer.gov/csr/1975_2000/sections.html [accessed April 13, 2004]., , , et al. SEER cancer statistics review,
- 37A phase II trial of human recombinant interleukin-2 administered as a 4-day continuous infusion for children with refractory neuroblastoma, non-Hodgkin's lymphoma, sarcoma, renal cell carcinoma, and malignant melanoma. A Childrens Cancer Group study. Cancer. 1995; 75: 2959–2965., , , et al.