Departments of 1HematologyOncology, 6Surgery and 4Pathology, Giannina Gaslini Childrens Hospital, Genova; 2Epidemiology and Clinical Trials Unit and 3Laboratory of Population Genetics, National Institute for Cancer Research, Genova; 5Departments of Pediatrics and Pediatric Surgery, Universities of Padova, Bologna, Napoli, Torino, Brescia, and Catania, Italy
Received 3 July 2001; revised and accepted 15 January 2002.
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
We previously reported that stage 3 neuroblastoma comprises (i) a low-risk group including all infants (age 011 months) as well as older children with non-abdominal primaries, and (ii) a high-risk group made up of children >1 year of age with abdominal primaries. Aggressive chemotherapy was effective only in the latter group.
Patients and treatment
On this basis, in 1990 we designed a new protocol by which all low-risk patients received standard-dose chemotherapy, while the high-risk ones received very aggressive chemotherapy.
Results
Between November 1990 and December 1997 a total of 95 eligible and evaluable children were enrolled: 47 were low-risk (35 infants and 12 >1 year of age at diagnosis and having non-abdominal primaries), and 48 were high-risk (being >1 year of age and having abdominal primaries). Of the 47 low-risk patients, five relapsed and four subsequently died. The 5-year overall survival (OS) was 91%. Of the 48 patients in the high-risk group, 22 relapsed or progressed, 18 of whom died from their disease and two from toxicity, and one was lost to follow-up. The 5-year OS was 60%. Univariate analysis showed that age, site of primary, risk-group, urine vanillylmandelic excretion, plasma levels of lactate dehydrogenase, ferritin and neurone-specific enolase, and MYCN status correlated with outcome. However, multivariate analysis showed that only MYCN status retained prognostic value.
Conclusions
In low-risk stage 3 neuroblastoma, standard-dose chemotherapy is associated with an excellent chance of being cured. Aggressive chemotherapy is effective for high-risk patients, but results are still unsatisfactory. MYCN gene amplification is a prognostic indicator for most, but not all, treatment failures.
Key words: prognostic factors, treatment, unresectable neuroblastoma
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In the range from the almost benign to the almost incurable tumors there is a small population of patients who present unresectable tumor masses but who have no metastatic spread. The outcome for these patients is less predictable, as shown by the rather different results of the few recent publications reporting long-term survival as ranging from as low as 60% [5] to as high as 88% [6].
On behalf of the Italian Cooperative Group for Neuroblastoma (ICGNB), we previously reported that this population includes two subsets of patients with a distinctively different survival expectancy [7]: (i) a low-risk group that includes all infants and older children with non-abdominal primary tumors; and (ii) a high-risk group comprising children 1 year of age or older with primaries located in the abdomen. Rather interestingly, the use of an intensified chemotherapeutic regimen was effective in improving survival results in the high-risk patients but not in the low-risk ones. On the basis of these data, the treatment protocol designed in 1990 by the ICGNB for unresectable neuroblastoma consisted of standard-dose chemotherapy for patients with low-risk features, and very aggressive chemotherapy for patients with high-risk features.
The main aims of this study were (i) to both confirm that standard-dose chemotherapy is suitable for low-risk patients, and to possibly identify clinical and biological characteristics associated with eventual failure, and (ii) to evaluate whether the modifications made to the previous high-dose chemotherapy regimen actually led to further improvement of results.
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Diagnosis was usually made on histological grounds. A few patients presenting features that strongly supported the hypothesis of neuroblastoma, but who were in poor general conditions making surgery risky, were included as well. Diagnostic work-up included evaluation of the primary tumor by computed tomography and/or magnetic resonance imaging, scintigraphic study with 123I-metaiodobenzylguanidine, one to four bone marrow aspirates and at least one bone marrow biopsy (the latter was limited to children 1 year of age or older), and the assay of vanillylmandelic (VMA) and homovanillic (HVA) acids in the urine, and of ferritin, neurone-specific enolase (NSE) and lactate dehydrogenase (LDH) in the serum.
Owing to the lack of revision of the histological material on a national level and the difficulties in classifying neuroblastoma on small and often scarcely representative biopsy specimens [9], the histological classification of tumor specimens was not taken into account when analyzing prognostic factors. Biological studies were performed on almost all tumor specimens sent to the national reference laboratory located in Genova [10].
Patient stratification
Patients with unresectable neuroblastoma were divided into two groups according to the clinical factors we had previously found to correlate with outcome. The low-risk group included all infants (011 months) and older patients with primary tumors located in the thorax, neck or pelvis. The high-risk group was made up of patients 1 year of age or older, with abdominal primary tumors.
Neo-adjuvant chemotherapy (Table 1)
Low-risk patients were treated with 3-monthly chemotherapy cycles made up of vincristine 1.5 mg/m2 on day 1, cyclophosphamide 150 mg/m2/day orally on days 17 and doxorubicin 35 mg/m2 on day 7. Patients achieving at least minor tumor response were entitled to receive three additional cycles, while patients with stable or progressive disease were switched to the high-risk treatment.1
|
All patients with a residual or still inoperable tumor after chemotherapy were treated individually with further chemotherapy or radiotherapy, or 131I-metaiodobenzylguanidine (MIBG) therapy according to the characteristics of the residual tumors and to the decision of the referring physicians. Patients achieving complete response after chemotherapy received no further therapy.
Surgery of primary tumor
Complete resection was defined as a radical resection or a resection leaving a residual tumor of <2 ml volume. Partial resection was a resection >50%, but less than complete. Biopsy was defined as a resection ranging from a fragment suitable for histological diagnosis to 50% of the primary mass.
Definition of tumor response
Complete remission was defined as the disappearance of the primary tumor and return to normal values of VMA and HVA urinary excretion; partial response as a >50% decrease in the primary tumor; minor response as a 2550% decrease in the primary tumor; no response as a <25% decrease or <25% increase of the primary tumor; and progressive disease as a >25% increase in the primary tumor or the appearance of a new lesion(s). Duration of response was not included in the definition.
Evaluation of response to therapy
Tumor response was evaluated; (i) after induction chemotherapy, (ii) within 1 month after delayed surgery, (iii) 12 months after the end of therapy, and (iv) at least every 6 months for the following 5 years.
Statistical analysis
The proportion of patients belonging to the two risk groups in different strata of prognostic factors was compared using the 2 test or Fishers exact test. The following variables were investigated for their prognostic impact on overall survival (OS): gender, age at diagnosis (<1 year versus 115 years), site of the primary tumor (abdomen versus other sites), MYCN oncogene copy number (<3 versus
3 copies), VMA and HVA urinary excretion (<2.5 versus
2.5 SD), and serum values of LDH (<1000 versus
1000 IU/l), NSE (<100 versus
100 ng/ml) and ferritin (<150 versus
150 ng/ml).
OS was defined as the time between diagnosis and death, regardless of the cause. Event-free survival was defined as the time between diagnosis and relapse or progression. Follow-up times were truncated on 31 March 1999, 15 months after the last patient was enrolled. Estimates of OS and event-free survival were calculated according to the KaplanMeier product limit method. Comparisons of estimated survival curves were performed by means of the MantelHaenszel 2 test. The uni- and multivariate estimates of the hazards ratio among different strata were calculated by Coxs proportional hazards model, and appropriate 95% confidence intervals were also reported. All tests are two sided.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Patients characteristics
The main clinical characteristics of the 95 evaluable patients are listed in Table 2. There was a slight prevalence of females (57 versus 43%). Median age at diagnosis was 19 months (range 1 day to 13 years). The abdomen was by far the most common site of primary tumors (72 cases), followed by thorax (14), pelvis (five) and neck (four). Diagnosis of neuroblastoma was made on histological grounds in 88 cases, and on imaging studies, elevated urinary catecholamines and an appropriate clinical pattern in seven patients, five of whom subsequently underwent surgery confirming the clinical impression of neuroblastoma. The majority of patients had elevated urinary catecholamines but normal plasma levels of LDH, NSE and ferritin. Of the 80 patients who were evaluated for MYCN oncogene copy number, 12 (15%) had three or more copies of the gene.
|
Forty-eight patients >1 year of age with abdominal primaries made up the high-risk group. Comparing low-risk patients with high-risk patients, the latter showed a greater male to female ratio and a higher percentage of abnormal values for LDH and NSE concentrations, and MYCN oncogene copy numbers (Table 2).
Clinical course
Low-risk group
Among 47 patients, 40 underwent biopsy and four underwent partial tumor resection. The remaining three patients were not operated on at this time. With the exception of an infant with a cervical tumor who died early of locally progressive disease, all patients were evaluated for tumor response following the six cycles of induction therapy. Complete response was observed in 10 patients, partial in 28 and minor in eight.
Following induction chemotherapy 25 of 46 evaluable patients underwent surgery, which resulted in complete resection in 15, partial resection in eight, biopsy in one and lack of visible tumor in another. The remaining 21 patients were not operated on because they were in complete remission (10 cases), or because their tumors still appeared inoperable (11 cases).
In conclusion, after chemotherapy and surgery (if performed), 26 patients were disease free and 20 had residual tumors (17 having achieved partial remission and three minor remission). None of the 26 patients in complete remission received further therapy. Twenty-five of them persist in remission while one, who had an amplified MYCN gene, relapsed locally and died. Of the 20 patients with residual tumors, five received therapeutic doses of MIBG (no relapse) two underwent additional chemotherapy (no relapse) and 13 had no further treatment (three local relapses). The characteristics of patients who relapsed or progressed are summarized in Table 3.
|
High-risk group
Forty-four of 48 patients underwent surgery at onset, which consisted of complete tumor resection in one case (stage 3 for infiltrated controlateral lymph nodes), partial resection in four and biopsy in 39. After induction therapy, three patients were in complete remission, 29 had achieved partial response and 11 minor response. Of the five remaining patients, two experienced early tumor progression and died (Table 3), one died of chemotherapy-related toxicity, one was withdrawn from the study and one could not be evaluated. Twenty-eight patients subsequently underwent surgery, which resulted in complete tumor resection in 11, partial resection in 13 and a biopsy in only four. At this time, 15 patients were not operated on because they were either disease free (three cases), or still inoperable (12 cases). No information about surgery was obtained regarding one patient.
In conclusion, after induction chemotherapy and surgery (if performed), of 43 evaluable patients 14 were in complete remission, and 24 in partial remission, while five had achieved only minor response. Of the 14 patients in complete remission, 10 received no further therapy (four relapsed) and four received additional chemotherapy (two relapsed). Of the 24 patients who had achieved partial response, one died shortly after surgery (due to an electrolyte imbalance), nine received no further therapy (six relapsed), three received additional chemotherapy (two relapsed), two received chemotherapy associated with radiotherapy (one relapsed) and nine were treated with therapeutic doses of MIBG (one relapsed). Of the five patients who had achieved minor response, two received no further therapy and both of them underwent tumor progression, one received chemotherapy and relapsed, and two received therapeutic doses of MIBG (one relapsed). The patient whose surgery information is missing received high-dose chemotherapy and is still alive.
Overall, 23 of 48 patients are alive progression free at 887 months (median 49 months). In 22 patients the tumors progressed or relapsed either locally (11 cases), or in distant sites (four cases), or in both (seven cases). Of these 22 patients, two are still alive in second complete remission, five are alive disease free and 15 have died from their disease. Lastly, two patients died of toxicity (chemotherapy-related and surgery-related in one case each) and one was withdrawn from the study during the induction phase.
Analysis of survival
The estimated cumulative probabilities of OS and event-free survival at 5 years from diagnosis for the entire series of 95 patients are 75 and 60%, respectively. Table 4 refers to the univariate analysis according to the main prognostic factors. Low-risk patients had an OS of 91% compared with 60% of those with high-risk characteristics (Figure 1). OS was significantly better in the 35 infants compared with the 60 children 1 year of age (91 versus 65%; P = 0.02). As for tumor markers, normal VMA urinary excretion, high LDH, ferritin and NSE plasma levels, and amplified MYCN gene were all associated with poorer survival. In multivariate analysis, the association between survival and risk group is confirmed (P = 0.023) only when the MYCN gene copy number is excluded from the model. However, the model including both the risk group and MYCN gene amplification showed that the interaction between these two factors had a significant effect (P = 0.002), thus demonstrating that the presence of MYCN gene amplification increases mortality more significantly in low-risk patients than in the high-risk ones.
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
There are few large studies related to unresectable neuroblastoma. In 1993, Garaventa et al. [7] reported on 145 such patients, of whom 77 were treated early with a standard-dose protocol, and 68 later on with a high-dose one. Progression-free survival at 5 years was 50% for the entire series but was significantly better for those treated with the high-dose protocol (59 versus 43%). Infants and children with non-abdominal tumors had a better outcome compared with children >1 year of age with abdominal disease (70 versus 30%). Surprisingly, high-dose therapy did improve progression-free survival in poor-risk patients, but not in good-risk ones. In 1998 Rubie et al. [6] published the French experience on 130 such patients treated with two cycles of carboplatinetoposide followed by two cycles of vincristinedoxorubicincyclophosphamide (CADO regimen). The 5-year event-free survival of the overall population was 78%. MYCN oncogene amplification was the most powerful predictor of poor outcome. In fact, children with unresectable neuroblastoma and no MYCN amplification fared as well as children with resectable tumors (event-free survivals of 89 and 91%, respectively). In 1998 Matthay et al. [18] reported the Childrens Cancer Groups experience on 228 stage 3 patients stratified for risk factors (age, MYCN gene copy number, histology type and ferritin levels). Patients with one or more risk factors were treated with more intensive therapy. The event-free survival at 4 years was 100% for patients with no risk factors, 90% for infants with one or more risk factors, and 50% for patients >1 year with one or more risk factors. Only MYCN copy number and age were independent factors in the multivariate analysis. The remarkable results of this and the previous study suggest that the biological characteristics of the tumor play a crucial role in the outcome of stage 3 neuroblastoma, and should be used as a guide for the therapeutic choices.
The present study was designed on the basis of our previous experience [7]. Accordingly, for low-risk patients (who fared well regardless of the intensity of treatment) a standard-dose and therefore scarcely toxic chemotherapy protocol was considered adequate, whereas high-risk patients were thought to possibly benefit from a more aggressive regimen. Somewhat disappointingly, the overall complete response rate thus obtained was slightly lower than in the previous study (45 versus 50%), while the rate of radical tumor resection was comparable (20 versus 19% of cases). Stricter criteria for defining response to chemotherapy and evaluating the post-operative residual tumor may account in part for the lack of improvement of these two endpoints. In particular, the low rate of surgical success can be accounted for by the rather cautious guidelines given by Italian surgeons, which have resulted in a remarkably low number of surgery-related fatal events.
Although the above figures fail to show an improvement compared with the previous study [7], the overall results of this study do appear superior. The 5-year OS rate of the present series has in fact increased from 55 to 75%, and the event-free survival rate has increased from 50 to 60%. Improvement has occurred in both patient groups. OS in low-risk patients, all of whom avoided aggressive therapy, increased from 70 to 91%, and in high-risk patients from 30 to 60%. Therefore, we clearly confirm that infants and children with non-abdominal tumors have an excellent chance of being cured without being exposed to risky chemotherapy. Further reduction of treatment could possibly be applied to these patients.
The improvement that occurred in children with high-risk characteristics is more difficult to explain. Perhaps the more meticulous extensive procedures applied in the present study have avoided enrolling some stage 4 patients with minimal tumor spread. In addition, the somewhat stricter criteria used to evaluate response have, on the one hand, decreased the number of complete remissions, while on the other implied longer and more intense treatment for partial response patients. The availability of carboplatin, which now substitutes the more toxic cisplatin, may also have contributed favorably [6, 19]. Finally, metabolic radiotherapy administered to a number of patients with macroscopic residual tumor may have had an eradicating effect [20].
Most treatment failures (23 of 27) occurred at the primary tumor site, usually within the first 2 years after diagnosis. Out of 26 patients who underwent delayed radical surgery only three relapsed (one local, two both locally and in distant sites) while more than one-third (26 of 69) of patients failing a delayed radical surgery progressed. Interestingly, the characteristics at diagnosis were similar, but none of the 21 evaluated patients who underwent radical surgery had amplification of MYCN gene. Although no further therapy was administerd to patients in complete remission, this probably did not lead to an increase in relapses (nine of 42 patients). Patients with residual tumors after surgery had a much higher risk of failure (18 of 48), except for the 16 who received radiometabolic therapy, of whom only two relapsed.
However, after induction therapy and delayed surgery, patients were treated heterogeneously and the role that treatment played is unclear. Only a homogeneous and prospective study can identify the impact on survival of residual tumors and further treatment.
In agreement with others [18, 21], we found that MYCN gene amplification acted as a strong predictor of failure. In fact, nine of 12 patients with this anomaly relapsed and only one of them is currently alive. However, MYCN amplification occurs in 20% of cases and therefore identifies only a proportion of patients destined to have an unfavorable clinical course. Better knowledge of the biological characteristics predisposing patients to treatment failure might help in designing more effective therapies for unresectable neuroblastoma.
![]() |
Acknowledgements |
---|
The following institutions (and main investigators) participated in this study: Giannina Gaslini Childrens Hospital, Genova (Bruno De Bernardi, Alberto Garaventa, Massimo Conte, Antonino Rizzo, Paolo Magillo, Claudio Gambini); Bambino Gesù Childrens Hospital, Roma (Alberto Donfrancesco, Alessandro Jenker, Clementina De Laurentis); Department of Pediatrics, University of Napoli, Napoli (Maria Teresa Di Tullio, Fiorina Casale); Department of Pediatrics, University of Torino (Maurizio Bianchi); Department of Pediatrics, University of Padova (Modesto Carli, Gianni Bisogno); Department of Pediatrics, University of Bari (Giovanni Surico, Nicola Santoro); Third Department of Pediatrics, University of Bologna (Antonia Mancini, Patrizia Alvisi); Department of Pediatrics, University of Palermo (Margherita Lo Curto); Department of Pediatrics, University of Brescia (Fabio Schumacher); Burlo Garofalo Childrens Hospital, Trieste (Paolo Tamaro); Department of Pediatrics, University of Modena (Monica Cellini); Department of Pediatrics, University of Parma (Giancarlo Izzi); Department of Pediatrics, University of Catania (Andrea Di Cataldo); Department of Pediatrics, University of Firenze (Gabriella Bernini, Angela Tamburrini); Division of Pediatrics, Civic Hospital, Bergamo (Pier Emilio Cornelli); Division of Pediatric Surgery, Civic Hospital, Vicenza (Luciano Musi); Department of Pediatrics, University of Perugia, Perugia (Augusto Amici); Department of Pediatrics, University of Siena (Antonio Acquaviva); Department of Pediatrics, University of Cagliari (Roberto Targhetta); and Department of Pediatrics, University of Sassari (Domenico Gallisai).
![]() |
Footnotes |
---|
The institutions and main investigators that participated in this study are listed in the Acknowledgements.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2. Castleberry RP. Paediatric update: neuroblastoma. Eur J Cancer 1997; 33: 14301438.[Medline]
3. Hayes FA, Smith EI. Neuroblastoma. In Pizzo AP, Poplack GP (eds): Principles and Practice of Pediatric Oncology, 2nd edition. London, UK: Lippincott 1993; 607622.
4.
Matthay KK, Villablanca JG, Seeger RC et al. Treatment of high-risk neuroblastoma with intensive chemotherapy, radiotherapy, autologous bone marrow transplantation and 13-cis-retinoic acid. N Engl J Med 1999; 341: 11651173.
5. Castel V, Badal MD, Bezanilla JL et al. Treatment of stage III neuroblastoma with emphasis on intensive induction chemotherapy: a report from Neuroblastoma Group of the Spanish Society of Pediatric Oncology. Med Pediatr Oncol 1995; 24: 2935.[ISI][Medline]
6. Rubie H, Michon J, Plantaz D et al. Unresectable localized neuroblastoma: improved survival after primary chemotherapy including carboplatinetoposide. Br J Cancer 1998; 77: 23102317.[ISI][Medline]
7. Garaventa A, De Bernardi B, Pianca C et al. Localized but unresectable neuroblastoma and outcome of 145 cases. J Clin Oncol 1993; 11: 17701779.[Abstract]
8. Brodeur GM, Pritchard J, Berthold F et al. Revisions of the international criteria for neuroblastoma diagnosis, staging and response to treatment. J Clin Oncol 1993; 11: 14661477.[Abstract]
9. Shimada H, Ambros IM, Dehner LP et al. Terminology and morphologic criteria of neuroblastic tumors. Cancer 1999; 86: 349363.[ISI][Medline]
10. Tonini GP, Boni L, Pession A et al. MYCN oncogene amplification in neuroblastoma is associated with worse prognosis except in stage 4s. The Italian experience with 295 children. J Clin Oncol 1997; 15: 8593.[Abstract]
11. Lanino E, Parasole R, Garaventa A et al. Treatment of poor-risk neuroblastoma with intensive chemotherapy and recombinant human granulocytemacrophage colony-stimulating factor. In Freund M, Link H, Schmidt RE, Welte K (eds): Cytokines in Hemopoiesis, Oncology and AIDS II. Berlin, Germany: Springer-Verlag 1992; 537543.
12. Donfrancesco A, De Bernardi B, Carli M et al. Deferoxamine followed by cyclophosphamide, etoposide, carboplatin, thiotepa, induction regimen in advanced neuroblastoma: preliminary results. Eur J Cancer 1995; 31A: 612615.
13. Evans AE, Silber JH, Shipilsky A et al. Successful management of low stage neuroblastoma with adjuvant therapies: a comparison of two decades, 1972 through 1981 and 1981 through 1992 in a single Institution. J Clin Oncol 1996; 14: 25042510.[Abstract]
14. Evans AE, DAngio GJ, Koop C et al. The role of multimodal therapy in patients with local and regional neuroblastoma. J Pediatr Surg 1984; 19: 7780.[ISI][Medline]
15. West DC, Shamberger RC, Macklis RM et al. Stage III neuroblastoma over 1 year of age at diagnosis: improved survival with intensive multimodality therapy including multiple alkylating agents. J Clin Oncol 1993; 11: 8490.[Abstract]
16. Powis MR, Imerson JD, Homes SJ. The effect of complete excision on stage III neuroblastoma: a report of the European Neuroblastoma Study Group. J Pediatr Surg 1991; 26: 11191124.[ISI][Medline]
17. Ikeda H, Suzuki N, Takahashi A et al. Surgical treatment of neuroblastoma in infants under 12 months of age. J Pediatr Surg 1998; 33: 12461250.[ISI][Medline]
18. Matthay KK, Perez C, Seeger RC et al. Successful treatment of stage III neuroblastoma based on prospective biologic staging: a Childrens Cancer Group study. J Clin Oncol 1998; 16: 12561264.[Abstract]
19. Gayon PS. Carboplatin in pediatric malignancies. Semin Oncol 1994; 21: 6576.[Medline]
20. Garaventa A, Bellagamba O, Lo Piccolo MS et al. 131I-metaiodobenzylguanidine (131I-MIBG) therapy for residual neuroblastoma: a mono-institutional experience with 43 patients. Br J Cancer 1999; 81: 13781384.[ISI][Medline]
21.
Bowman LC, Castleberry RP, Cantor A et al. Genetic staging of unresectable or metastatic neuroblastoma in infants: a Pediatric Oncology Group study. J Natl Cancer Inst 1997; 89: 373380.