Treatment of poor-risk neuroblastoma patients with high-dose chemotherapy and autologous peripheral stem cell rescue

Transcription

1 Bone Marrow Transplantation, (1997) 20, Stockton Press All rights reserved /97 $12.00 Treatment of poor-risk neuroblastoma patients with high-dose chemotherapy and autologous peripheral stem cell rescue SL Cohn 1, TJ Moss 2, M Hoover 1,3, HM Katzenstein 1, PR Haut 1, ER Morgan 1, AA Green 4 and M Kletzel 1 1 Department of Pediatrics, Northwestern University Medical School and Children s Memorial Hospital, Chicago, IL; 2 BIS Laboratories, Reseda, CA; and 4 Department of Pediatrics, Rush Presbyterian Medical Center, Chicago, IL, USA Summary: than 20% have long-term survival. 1,2 Because chemotherapy dose-intensity has been shown to correlate strongly A single institutional pilot study was conducted in which with response and progression-free survival in metastatic 12 poor-risk neuroblastoma (NB) patients were uni- NB, 3 many centers are currently treating poor-risk patients formly treated with multi-agent induction chemotherapy with consolidation regimens that include myeloablative followed by myeloablative consolidation chemo- chemotherapy or chemoradiotherapy followed by allo- therapy and unpurged peripheral blood stem cell geneic bone marrow transplant (BMT) or autologous BMT. (PBSC) rescue. In addition to using standard criteria While the efficacy of myeloablative vs nonmyeloablative for evaluating response to induction chemotherapy, consolidation therapy has not been determined, some of the tumor cell contamination of the peripheral blood and/or best long-term survival rates have been reported with highdose bone marrow was analyzed in seven patients by myeloablative chemotherapy with or without totalbone immunocytology using a panel of five anti-nb mono- body irradiation, and bone marrow rescue. 2,4 8 However, clonal antibodies. Seven patients had morphologic evidence the ability to provide bone marrow stem cells for patient of bone marrow disease at the time of diagnosis, rescue following intensive therapy can be limited by mar- and two additional patients had tumor cells detected in row contamination with tumor cells, and by the lack of bone marrow samples by immunocytology prior to the adequate marrow reserves to harvest sufficient numbers of second cycle of chemotherapy. After three cycles of stem cells. chemotherapy, two of the 12 patients continued to have Rescue with peripheral blood stem cells (PBSCs) has evidence of bone marrow disease. Samples from 29 several advantages over autologous bone marrow. The pro- PBSC harvests collected from nine patients were also genitor cell collection can be performed on an outpatient analyzed for the presence of contaminating tumor cells basis without general anesthesia, and in general PBSCs can by immunocytology. In each case, the stem cells were be successfully harvested in patients who do not have found to be free of tumor. Eleven of the 12 patients adequate marrow reserves. 9 In addition, rescue with PBSCs underwent myeloablative therapy and PBSC rescue; results in a significantly reduced cytopenic period after five patients remain alive without disease progression, myeloablative therapy Further, studies in adults have 28+ to 53+ months from diagnosis, and six patients have indicated that the peripheral blood is enriched with stem developed recurrent disease. We conclude that PBSCs cells during the recovery phase following chemotherapy, can be successfully harvested from children with NB, and with proper timing and growth factor support, higher and used for hematopoietic reconstitution following yields of stem cells can be obtained from peripheral stem myeloablative chemotherapy. However, more effective cell harvests than from conventional bone marrow harvests. therapy for poor-risk NB patients is still urgently 9,12 Although the experience with PBSC harvest and needed. rescue in children is limited, recent pediatric studies have Keywords: neuroblastoma; peripheral blood stem cells; shown that PBSC collection is feasible in children, and the tumor cell contamination; immunocytology procedure does not appear to be limited by age or weight of the patient. 11,13 16 Interestingly, in patients with NB, the incidence of tumor contamination has been reported to be lower in PBSC harvests than in bone marrow grafts. 13,17 Children diagnosed after the age of 1 year with widely dis- We report the results of a pilot study conducted at the seminated neuroblastoma (NB), remain a therapeutic chal- Children s Memorial Hospital in Chicago in which 12 poorlenge. Even with intensive multi-modality therapy, less risk NB patients were uniformly treated with multi-agent induction chemotherapy followed by myeloablative consolidation chemotherapy and unpurged PBSC rescue. Bone Correspondence: Dr M Kletzel, Division of Hematology/Oncology, Box marrow and peripheral blood samples from seven patients 30, Children s Memorial Hospital, 2300 Children s Plaza, Chicago, IL were examined for tumor contamination sequentially during 60614, USA 3 therapy using established sensitive immunologic tech- Current address: Children s Hospital of Orange County, 455 S Main St, Orange, CA, USA niques. 18,19 In addition, immunocytologic analysis was per- Received 29 January 1997; accepted 6 June 1997 formed on samples from 29 PBSC harvests collected from

2 544 nine patients to determine the incidence of tumor contami- therapy. An additional three cycles of chemotherapy, identical to the first three cycles, were administered to patients nation. Our studies demonstrate that neuroblasts are no longer detected in the bone marrow and peripheral blood with responsive tumors. in a subset of patients after three cycles of chemotherapy. Patients were re-evaluated after the sixth cycle of chemo- Furthermore, there was no evidence of tumor contamination therapy, and second-look surgery was performed on all in PBSC harvests collected at this time. patients who had responsive tumors and no morphological evidence of bone marrow disease. In the operating room an effort was made to remove all residual tumor. If resection was not possible, multiple biopsies of the tumor were Materials and methods performed. After recovery from surgery, patients received Patients myeloablative therapy consisting of thiotepa 300 mg/m 2 i.v. daily, day 7 to day 5, and cyclophosphamide 1500 From December 1991 to September 1994, 11 newly diag- mg/m 2 i.v. daily, day 5 to day 2. On day 0, the PBSCs nosed consecutive patients older than 1 year of age with were infused. Due to inadequate yields of PBSCs, one International Neuroblastoma Staging System (INSS) stage patient (patient 3) received a supplement of stem cells 4NB 20,21 were enrolled on a pilot study conducted at Chil- obtained from bone marrow harvest. Thirty to 40 days foldren s Memorial Hospital in Chicago designed to treat lowing the PBSC infusion, seven of the 11 patients received poor-risk NB. In addition, one child with recurrent NB was 2400 cgy to the primary tumor bed in 20 fractions over enrolled on the study. This child was initially diagnosed 10 days. with INSS stage 4 NB at 2. years of age, and relapsed in Routine supportive care included hospitalization in the blood and bone marrow 3 years following an allogeneic rooms with positive pressure undirectional airflow systems, bone marrow transplant. The protocol was approved by the prophylactic acyclovir 250 mg/m 2 /dose i.v. twice daily, Children s Memorial Institutional Review Board and writ- daily fluconazole mg p.o., gammaglobulin 250 ten informed consent was obtained for each patient. Diag- mg/kg i.v. weekly for the first 100 days and then monthly nosis was made according to international criteria, 20,21 and for 3 months, pentamidine 4 mg/kg i.v. monthly from day was based on unequivocal pathologic examination of tumor to +180, and broad spectrum antibiotics for treattissue or bone marrow and urine catecholamine excretion. ment of febrile neutropenia. All patients received GM-CSF Clinical data collected included age at diagnosis, sex of 1000 g/m 2 i.v. daily, until they achieved an ANC the child, and stage of disease according to the INSS and /l. Computed tomography scans, bone scans, outcome. Tumor extent at diagnosis was determined by and urine catecholamine assays were performed at +30 and computed tomography, bone scan, meta-iodobenzylguani days to determine disease status. dine (MIBG) scan, and bone marrow aspirate. Nine patients were simultaneously enrolled on the Pediatric Oncology Group (POG) NB Biology Study No. 9047, and prior to Peripheral blood stem cell collection treatment, tumor N-myc copy number was analyzed by Bone marrow aspirates were performed prior to each cycle Southern blot or fluorescence in situ hybridization at the of chemotherapy until two sequential bone marrow samples POG reference laboratory at St Jude Children s Research were morphologically free of tumor. Twenty-four hours fol- Hospital using standard methods Tumor tissues from lowing completion of each chemotherapy cycle all patients patients 2 and 10 were not available at the time of diag- received daily G-CSF to enhance stem cell mobilization. nosis. However, N-myc copy number was measured in PBSCs were collected approximately 7 to 10 days followrecurrent tumor by Southern blot analysis in the laboratory ing completion of a chemotherapy cycle, once the periphof one of the authors (SCL). One patient (3) did not have eral white blood cell count recovered to /l. tumor tissue available for N-myc analysis. The response to The harvest was performed through a double lumen cuffed induction chemotherapy was determined according to inter- central venous line (Hickman or Broviac, Salt Lake City, national criteria after cycles 3 and 6 using bone scan, com- UT, USA), and 1.5 to 2.0 blood volumes were processed puted tomography, bone marrow aspirate, and urine cat- either on a CS-3000 (Baxter, Deerfield, IL, USA) or Cobe echolamine metabolite determination. Following cycle 6, a Spectra pheresis machine (Cobe, Denver, CO, USA). The second MIBG scan was performed. pheresis was repeated on 2 to 4 consecutive days after cycles 4, 5 and 6 of induction chemotherapy until a minimum of mononuclear cells/kg were collected. The Treatment duration of each pheresis was approximately 4. h. All patients were initially treated with three cycles of combination chemotherapy delivered every 21 days. The chemotherapy cycle was delayed 1 week if the absolute neutrophil count (ANC) was /l or if the platelet count was /l. G-CSF (5 10 g/kg/day) was started 24 h after each cycle of chemotherapy was completed to mobilize stem cells, and it was continued until the patient achieved an ANC of /l. Patients were evaluated for response after the third cycle of chemo- Cell processing and analysis Samples from the individual daily harvests were obtained for total cell count, CD34 + content, hematopoietic progeni- tor assay, and bacterial culture; the PBSC harvests were suspended in a final dimethyl-sulfoxide (DMSO) concen- tration of 10%. Each PBSC harvest was cryopreserved individually in a controlled rate freezer to 70 C, and stored in a liquid nitrogen tank until the time of PBSC rescue

3 as previously described. 10 Cells expressing surface antigen Results CD34 were identified by flow cytometry analysis using fluoroscein isothiocyanate (FITC)-conjugated mouse anti- Patients and response to therapy CD34 monoclonal antibody (Becton Dickinson, San Jose, CA, USA). In addition, surface antigen expression of The clinical characteristics of the 12 high-risk NB patients CD33, CD38, and HLA-DR was determined using FITCchildren had newly diagnosed stage 4 disease, and one enrolled on this pilot protocol are shown in Table 1. Eleven conjugated mouse anti-cd33, anti-cd38, and anti-hlapatient had recurrrent NB. All 12 patients were older than DR antibodies (Becton Dickinson, San Jose, CA, USA). Hematopoietic progenitors (colony-forming cells (CFU)- 1 year of age at the time they were enrolled on study, and seven of the 12 children had N-myc amplified tumors. mix, erythroid burst-forming cells (BFU-E) and colony-for- Seven patients had morphologic evidence of bone marrow ming units granulocyte monocyte (CFU-GM) were assayed disease at the time of diagnosis (Table 2). However, using in a methylcellulose system containing phytohemagglutimore sensitive immunocytologic methods, tumor cells were nin-stimulated leukocyte conditioned medium and erythrodetected in bone marrow samples from two additional poietin using a kit from Stem Cell Technologies patients prior to the second cycle of chemotherapy (patients (Vancouver, Canada). On days 7 and 14, hematopoietic 5 and 7). progenitor colony number was determined. Of the five patients who did not have morphologic evidence of bone marrow disease at the time of diagnosis, two patients (3 and 8) were classified as stage 4 because of extensive disease in the chest and abdomen. Patient 2 had a Immunocytologic analysis large retroperitoneal primary tumor and lung nodules which were determined to be tumor metastases pathologically. Patient 5 had a primary adrenal mass with metastases to The immunocytologic analysis was performed in the labthe thoracic cavity, inguinal and cervical lymph nodes, and oratory of one of the authors (TJM). Bone marrow and peribone. In addition to disease in the retroperitoneum and pheral blood samples from seven patients were obtained mediastinum, patient 7 had metastatic disease in suprabefore cycles 2 and 4 of induction chemotherapy to evaluclavicular lymph nodes. ate whether contaminating neuroblasts were present. In The chemotherapy schema is shown in Table 3. All 12 addition, samples from 29 PBSC harvests collected from patients achieved a partial response (PR) after the first three nine patients were analyzed for tumor contamination. cycles of chemotherapy, and the bone marrow was morpho- Unfortunately, patient 1 was enrolled on this protocol prior logically tumor-free in all but one patient (patient 11). to the time arrangements were made to examine the Sequential immunocytologic analysis of peripheral blood samples for tumor contamination, and initially only PBSC and bone marrow samples from seven patients during the samples were collected for immunocytology. Patients 11 induction phase of chemotherapy demonstrated that five and 12 did not have PBSC samples analyzed because the patients had tumor-free peripheral blood and bone marrow immunocytology laboratory was closed during the period samples prior to the 4th cycle of chemotherapy, while two of time the samples were collected. patients continued to have detectable bone marrow disease Mononuclear cell fractions were isolated from each sam- (Table 2). Patient 11 had both morphologic and immunople by density gradient separation on Ficol Hypaque logic evidence of bone marrow disease after five cycles of (Pharmacia, Uppsala, Sweden), tested for cell viability by chemotherapy, and was removed from our study. The Trypan Blue (Sigma Chemical Co, St Louis, MO, USA), remaining 11 patients completed the induction chemoand resuspended in Liebovitz L15 medium (GIBCO, Long therapy and underwent second-look surgery. Following Island, NY, USA) supplemented with 15% fetal bovine completion of induction therapy, three patients were serum (GIBCO) at a cell volume of cells/ml as determined to be in a complete response (CR), five had a described. 19 The cell suspension was incubated at room very good partial response (VGPR), and three had a partial temperature for 45 min with a cocktail of antibodies gener- response (PR) (Table 1). All 11 received myeloablative ously provided by Dr John Kemshead (UJ13A, 5.1. H11, chemotherapy followed by PBSC rescue (Table 4). M340, UJ127.11, and anti-thy 1). Cells were then spun onto glass slides using a Shandon cytocentrifuge (Shandon, Philadelphia, PA, USA), and stored overnight at 4 C. Slides Peripheral stem cell collection and rescue were fixed in a paraformaldehyde solution at 4 C for 30 The 11 patients underwent a total of 50 pheresis. Two to min, washed in phosphate-buffered saline, and then placed four phereses were completed following each cycle of on a Biotek automated immunostainer (Biotek, Walkerville, chemotherapy in an out-patient setting. The procedure was MD, USA) and incubated with Biotek immunoreagents for well tolerated. No patient developed hypocalcemia or citalkaline phosphatase staining as described. 19 Positive con- rate toxicity. However, one patient did become hypotensive trol slides were prepared with cells from the neuroblastoma during the pheresis, and required fluid administration. Most cell line HTB-10. Non-immune mouse serum at 5 g/ml patients developed thrombocytopenia and received platelet was used as a negative control. If control studies did not transfusions. The total number of phereses performed for each patient was two to nine (mean 4.5) (Table 5). Three patients underwent their first harvest 8 to 10 weeks following diagnosis; seven patients had their first harvest perfor- yield the expected result, the assays were repeated with a fresh cocktail of antibodies. A minimum of cells per sample were examined. 545

4 546 Table 1 Patient characteristics and response to treatment Patient Age Sex INSS N-myc Disease status prior Disease Post-PBSC Recurrent disease Post-relapse Outcome stage amplified to PBSC rescue status at rescue treatment (time) treatment (follow-up) day mo M 4 Yes CR CR None 1, LN, Bone (17 mo) Hydroxyurea Died: 19 mo 2 19 yrs, 10 mo M 4 No VGPR VGPR None 1 (9 mo) Topotecan Died: 9 mo (3 cycles) 3 2 yrs, 6 mo M 4 ND VGPR CR None 1 (11 mo) None Died: 12 mo 4 2 yrs 9 mo F 4 No CR CR None BM (15 mo) None Died: 16 mo 5 12 mo M 4 Yes VGPR CR Rt to site of 1 None Alive: NED (40+ mo) 6 3 yrs 5 mo M 4 Yes PR PR Rt to site of 1, Brain (12 mo) Brain RT Alive: NED anti-gd2 (39+ mo) 7 5 yrs, 0 mo M 4 No PR PR Rt to site of 1, None Alive: stable retinoic acid disease (36+ mo) 8 3 yrs 2 mo M 4 Yes PR PR RT to site of 1, None Alive: NED radiolabeled MIBG, (36+ mo) surgery, retinoic acid 9 3 yrs, 4 mo F 4 Yes VGPR PR RT to site of 1, Bone, BM, intracranial Brain RT Died: 11 mo RT to knee (11 mo) 10 5 yrs 9 mo M relapsed No VGPR CR RT to site of 1 None Alive: NED (53+ mo) 11 3 yrs, 7 mo F 4 Yes Persistent disease in NA NA NA NA Died: 14 mo BM after 5 cycles of CT; off study 12 3 yrs 4 mo F 4 Yes CR CR RT to site of 1 None Alive: NED (28+ mo) mo = months; yrs = years; M = male; F = female; ND = not done; CR = complete response; VGPR = very good partial response; PR = partial response; BM = bone marrow; CT = chemotherapy; NA = not applicable; RT = radiotherapy; LN = lymph node; NED = no evidence of disease; 1 =primary.

5 Table 2 PBSC rescue for poor-risk neuroblastoma Morphologic and immunologic analysis of neuroblastoma tumor cell contamination in sequential bone marrow and peripheral blood samples 547 Patient At Prior to cycle 2 Prior to cycle 4 PBSC collections diagnosis No. negative samples/total BM BM BM PB BM BM PB no. analyzed (morph) (morph) (imm) (imm) (morph) (imm) (imm) 1 + ND ND ND ND ND ND 2 ND ND ND ND 2/2 3 ND ND ND ND ND 3/3 4 + ND ND ND ND ND 7/ / /1 7 + ND 4/4 8 2/2 9 + a + + 2/ ND ND ND ND 3/ ND ND 12 + ND Morph = morphology; imm = immunology; BM = bone marrow; PB = peripheral blood. a Dilute sample. Table 3 Chicago pilot No. 1 treatment schema Chemotherapy agent Week 0 Week 3 Week 6 Week 9 Week 12 Week 15 Week 18 Cisplatinum/etoposide X X (40 mg/m 2 /day; 150 mg/m 2 /day 5 days) Cyclophosphamide (1 g/m 2 /day 2 days) X X Doxorubicin (35 mg/m 2 ) X X Carboplatin (500 mg/m 2 ) X X Ifosfamide/etoposide (1.8 g/m 2 day; 150 X X mg/m 2 /day 5 days) PBSC harvest X a X a X a Surgery X a PBSC harvest was performed approximately 7 10 days following completion of chemotherapy cycle. Table 4 Chicago pilot No. 1 conditioning regimen Chemotherapy agent Day 7 Day 6 Day 5 Day 4 Day 3 Day 2 Day 1 Day 0 Thiotepa (300 mg/m 2 ) X X X Cyclophosphamide (1500 mg/m 2 ) X X X X Stem cell infusion X range 0 685) respectively. The mononuclear cell dose/kg collected in each apheresis ranged from to (Table 5). No significant difference in cell yields was observed after the different cycles of chemotherapy. The median percentage of CD34 + cells in each apheresis ranged from 0.2 to 3.5% ( to CD34 + cells/kg). Twenty-nine PBSC harvests collected from nine patients were analyzed by immunocytology for the pres- ence of contaminating tumor cells, and all samples were determined to be negative (Table 2). Mononuclear cell dose/kg infused ranged from to with a median mononuclear cells/kg (Table 6). Patient 3 received /kg bone marrow stem cells in addition med at week 11 to 13 after diagnosis, and one patient underwent initial PBSC harvest at week 15. Ten patients had additional aphereses performed following cycle 5 of chemotherapy, 2 to 5 weeks after the first harvest. Four patients also had PBSCs collected after cycle 6 of chemotherapy (weeks 13 to 19 following diagnosis). Hematopoietic progenitor colony number was determined on each harvest for quality control. In each experiment mononuclear cells were plated. On day 7 the mean CFU-GM was 70.5 (median 28; range 0 586), and the mean BFU-E count was 15.6 (median 5; range 0 135). On day 14 the mean CFU-CM and BFU-E counts were 200 (median 139; range 0 879) and 106 (median 52;

6 548 Table 5 Peripheral blood stem cell collection data Patient Weight Number of Collection vol TNC collected MNC collected MNC 10 8 /kg % CD34 + median CD /kg (kg) collections processed median 10 8 median 10 8 median median (range) (range) median (range) (range) (c.c.) (range) (range) (50 50) 3.3 (0.7 28) 1.7 (0.3 23) 0.17 ( ) 1.4 (1 2.5) 0.3 ( ) (56 56) 360 ( ) 106 (91 120) 1.5 ( ) 1.7 ( ) 0.8 (0.05 1) (50 100) 2.2 ( ) 2.8 (2 3.7) 0.15 ( ) 1.8 ( ) 0.2 ( ) (50 100) 2.1 ( ) 2.4 (0.15 4) 0.14 ( ) 1.2 ( ) 0.1 ( ) (90 171) 25.6 (2.2 67) 9.5 (2 21) 0.77 ( ) 3.5 ( ) 2.2 ( ) ( ) 12.8 (4.5 39) 10.8 (4.1 21) 0.5 ( ) 1.4 ( ) 0.7 ( ) (98 130) 30 (11 63) 16.5 (9 18) 0.94 ( ) 2.5 ( ) 1.4 ( ) (80 139) 2.4 (1.7 25) 2.3 (1.6 24) 0.2 ( ) 1.7 ( ) 0.2 ( ) (91 166) 11.7 (3.7 61) 9.7 (2.7 21) 0.6 ( ) 2.6 (2 3) 1.5 ( ) (40 53) 8.3 ( ) 1.5 ( ) 0.1 ( ) 0.2 ( ) 0.1 ( ) ( ) 8.6 (4.8 19) 6.9 (4.2 12) 0.46 ( ) 0.7 (0.3 1) 0.6 (0.2 1) mean/apheresis median/apheresis TNC = total nucleated cells; MNC = mononuclear cells; vol = volume.

7 Table 6 Engraftment results Of the seven patients who received radiotherapy post- PBSC rescue, two have relapsed; one patient (patient 6) Patient MNC dose ANC Platelet relapsed in the brain, and the other (patient 9) developed infused /l /l recurrent disease in the region of the ethmoid sinus, the 10 8 /kg (days) (days) skeletal bones, and the bone marrow (Table 1). Patient 6 remains alive with no evidence of disease after receiving radiotherapy to the brain and 8 months of treatment with a monoclonal antibody anti-gd2. Patient 9 received radio therapy to the brain, but died with progressive disease months following diagnosis. The remaining five patients have no evidence of disease progression at the time of writ ing. Three (patients 5, 8 and 12) are currently alive with no evidence of disease at 40+ months, 36+ months, and months from diagnosis, respectively. Patient remains disease-free 53+ months following his relapse. Mean Patient 7 remains alive with stable disease 36+ months Median Range ( ) (9 25) (11 64) from diagnosis. 549 MNC = mononuclear cells; ANC = absolute neutrophil count. a Patient 3 received /kg MNC obtained from bone marrow due to low PBSC yields. to /kg PBSCs. The median time to absolute neutrophil count of /l was 14 days (range 9 33 days), and platelet count recovery ( /l) was observed at a median of 27 days (range days). Toxicity and patient outcome Discussion Although peripheral blood stem cells have been harvested from adult cancer patients and used for hematopoietic reconstitution following myeloablative therapy for several years, only recently have stem cells been successfully harvested from the peripheral blood of children weighing less than 20 kg. We report our experience with intensive multiagent induction chemotherapy followed by myeloablation and peripheral blood stem cell rescue in 12 poor-risk NB patients. Eleven of the 12 patients achieved cytologic bone marrow remissions with the multi-agent induction chemotherapy, and received myeloablative chemotherapy followed by PBSC rescue. These 11 patients underwent PBSC harvesting without complication, and in all but one case at least mononuclear cells/kg were collected from PBSC harvests. The cell yields are similar to that reported in other pediatric series. 15,16 There was no transplant-related mortality, although seven episodes of bacteremia or fungal infection were documented following PBSC infusion. Engraftment of neutrophils and platelets was observed after a median of 14 and 27 days, respectively, which is similar to what has been observed in adult clinical trials and the few published pediatric studies In addition to using the standard international criteria for evaluating response to the induction chemotherapy, we also analyzed samples of peripheral blood and bone marrow from seven patients sequentially during therapy using sensitive immunocytologic methods to determine if tumor cells were present. 18,19 Although NB cells from individual patients were not tested for reaction with the antibodies used in our studies, Moss and colleagues 25 have previously demonstrated that virtually all NB cells react with this cocktail of antibodies. These studies demonstrated that tumor cells were not detected by morphology or immunocytology in the bone marrow or peripheral blood of five of Five patients had documented bacteremia and two patients developed fungal infections following the PBSC rescue, prior to hematopoietic recovery. One patient developed a urinary tract infection. Additional transplant-related morbidity included cardiac insufficiency which resolved (one patient), and mild skin sloughing which healed without complication (one patient). There was no transplantrelated mortality. At day +100, all 11 patients were alive without evidence of progressive disease (Table 1). Six were determined to be in CR, one had a VGPR, and four had a PR. The first three patients enrolled in the study developed recurrent disease at their primary tumor site 9 to 17 months following completion of therapy. Recurrent metastatic disease was also observed in one of these three patients. The fourth patient relapsed in her bone marrow 15 months following PBSC rescue, and was not further evaluated to determine if there were additional sites of recurrent disease per the request of his family. Post-relapse, two patients received no further therapy. The other two patients were treated on phase I or II studies (Table 1). All four patients died of disease 9 to 19 months following diagnosis. In an effort to decrease the high rate of local failure, the next seven patients enrolled on the study were treated with radiation to the site of their primary disease approximately 30 to 40 days after the PBSC rescue, once they achieved the seven patients analyzed following three cycles of intensive multi-agent chemotherapy. Furthermore, our study an ANC /l and a platelet count of /l. The four patients who were in PR at day +100 indicates that tumor-free PBSC harvests can be collected (patients 6, 7, 8 and 9) received additional therapy following completion of the treatment outlined in the pilot study of the PBSC harvests we analyzed by immunocytology after four cycles of conventional chemotherapy, as all 29 as shown in Table 1. showed no evidence of tumor contamination. Unfortu-

8 550 nately, for logistical reasons, the PBSC harvests from some lowing myeloablative therapy. In a subset of patients, tumor of the patients were not analyzed by immunocytology. cells are no longer detected morphologically or by immuno- It is possible that a small number of neuroblasts may cytology in the peripheral blood or bone marrow after three have been present in the PBSC harvests which were not cycles of conventional chemotherapy. Immunocytologic detected by immunocytology. Others have shown that analysis of PBSC grafts harvested at this time demonstrate clonogenic neuroblastoma tumor cells can be present in no evidence of tumor contamination. While the number of peripheral blood samples that are negative by immuno- patients in our cohort is small, the outcome appears to be cytology. 25 It is also possible that more sensitive assays better than the historical results at this institution when, such as analysis of tyrosine hydroxylase expression using similar to other centers, less than 15% of our poor-risk the reverse-transcription polymerase chain reaction 17 may patients survived 2 years with conventional doses of have demonstrated minimal residual disease. However, chemotherapy. However, the high incidence of relapse seen even when minimal residual disease was measured by other in this and other studies emphasizes the urgent need for the methods, other investigators have reported similar results. development of more effective therapy. We are currently Using the sensitive reverse-transcription polymerase chain conducting another pilot study for poor-risk NB patients in reaction assay, Miyajima and colleagues 17 sequentially which the consolidation phase of therapy consists of three examined both peripheral blood and bone marrow samples repetitive PBSC rescues following cycles of very high-dose for neuroblast contamination in 15 patients with stage IV multi-agent chemotherapy. Early results suggest that this disease. These investigators observed that after several approach is feasible, and that most of the treatment can cycles of conventional chemotherapy, a subset of these be given in an outpatient setting. Further follow-up will poor-risk patients cleared their bone marrow and peripheral determine if intensifying therapy in this manner will lead blood of tumor cells. Interestingly, at the time of PBSC to improved long-term survival for children with poorrisk harvesting, these investigators demonstrated that the incidence NB. of tyrosine hydroxylase mrna positivity was significantly lower in PBSC samples than in bone marrow samples obtained concurrently, suggesting that tumor cell Acknowledgements contamination is less likely to occur if stem cells are collected from the peripheral blood than from the bone marthe The authors thank Helen Salwen and Susan Rowe for performing row. Similarly, Saarinen and colleagues 26 have reported N-myc analyses, Christina Rigby and Marie Balderas for that it is possible to purge bone marrows in vivo in poor- assistance with the data collection, Marie Olszewski and Carmen risk NB patients using conventional chemotherapy. These Co for assistance with processing the PBSC harvests, and Karina Danner-Koptik and Maureen Hubbell for outstanding nursing investigators examined bone marrow samples for tumor cell care. contamination by routine cytology as well as by immunofluorescence using the anti-gd2 monoclonal antibody 3A7. Of the 18 patients who initially had evidence of bone marrow disease, tumor cells were not detected in the bone marrow References of five patients after 2 months of conventional therapy, 1 Matthay KK. Neuroblastoma. In: Pochedly C (ed). Neoplastic and an additional eight patients cleared their marrows after Diseases of Childhood. Harwood Academic: Chur, Switzer- 2 to 6.5 months of treatment. Five patients never achieved land, 1994, pp bone marrow remissions. 2 Philip T, Zucker JM, Bernard JL et al. Improved survival at Although our results and review of the literature suggest 2 and 5 years in the LMCE1 unselected group of 72 children that it is possible to obtain PBSC grafts that have no evinosis: with stage IV neuroblastoma older than 1 year of age at diag- dence of tumor contamination from patients with poor-risk is cure possible in a small subgroup? J Clin Oncol 1991; NB, the outcome still remains poor for this group of chil- 9: dren. In fact, regardless of the source of the graft or the 3 Cheung NK, Heller G. Chemotherapy dose intensity correlates method of purging, a relapse rate of 40 50% following strongly with response, median survival, and median pro- gression-free survival in metastatic neuroblastoma. J Clin myeloablative therapy and stem cell transplant has been Oncol 1991; 9: reported in most clinical trials. 1,2,4,5,7,27 33 NB most fre- 4 Stram DO, Matthay KK, O Leary M et al. Consolidation quently recurs post-transplant at the primary site, or in sites chemoradiotherapy and autologous bone marrow transplanof previous disease. 27 This pattern of relapse suggests that tation vs continued chemotherapy for metastatic neuroblasin many cases the recurrent disease is due to a failure to toma: a report of two concurrent Children s Cancer Group eradicate the patient s original disease fully, rather than Studies. J Clin Oncol 1996; 14: from infusion of tumor-cell contaminated grafts. In our sertherapy 5 Pole JG, Casper J, Elfenbein G et al. High-dose chemoradioies of patients, five of the initial 12 patients enrolled on supported by marrow infusions for advanced neuro- this pilot protocol remain alive, without disease-pro- blastoma: a Pediatric Oncology Group study. J Clin Oncol gression, 28+ to 53+ months from diagnosis. However, six 1991; 9: Zucker JM, Philip T, Bernard JL et al. Single or double conpatients developed recurrent disease following myeloablasolidation treatment according to remission status after initial tive chemotherapy and PBSC rescue, five of whom relapsed therapy in metastatic neuroblastoma: first results of LMCE 3 in known sites of previous disease. study in 40 patients. Bone Marrow Transplant 1991; 7 (Suppl. 7 Hartmann O, Benhamou E, Beaujean F et al. Repeated highdose chemotherapy followed by purged autologous bone mar- The results of this pilot study suggest that PBSCs can 2): 91. be successfully collected in small children and used to hematopoietically reconstitute patients with poor-risk NB fol-

9 row transplantation as consolidation therapy in metastatic 22 Shapiro DN, Valentine MB, Rowe ST et al. Detection of N- neuroblastoma. J Clin Oncol 1987; 5: myc gene amplification by fluorescence in situ hybridization. 8 McCowage GB, Vowels MR, Shaw PJ et al. Autologous bone Diagnostic utility for neuroblastoma. Am J Pathol 1993; 142: marrow transplantation for advanced neuroblastoma using teniposide, doxorubicin, melphalan, cisplatin, and total-body 23 Brodeur GM, Seeger RC, Schwab M et al. Amplification of irradiation. J Clin Oncol 1995; 13: N-myc in untreated human neuroblastomas correlates with 9 Lee JH, Klein HG. Collection and use of circulating hemato- advanced disease stage. Science 1984; 224: poietic progenitor cells. Hematol Oncol Clin N Am 1995; 9: 24 Look AT, Hayes FA, Shuster JJ et al. Clinical relevance of tumor cell ploidy and N-myc gene amplification in childhood 10 Kletzel M, Longino R, Danner K et al. Peripheral blood stem neuroblastoma: a Pediatric Oncology Group study. J Clin cell rescue in children with advanced stage IV neuroblastoma. Oncol 1991; 9: Prog Clin Biol Res 1994; 389: Moss TJ, Cairo M, Santana VM et al. Clonogenicity of circul- 11 Takaue Y, Kawano Y, Abe T et al. Collection and transplan- ating neuroblastoma cells: implications regarding peripheral tation of peripheral blood stem cells in very small children blood stem cell transplantation. Blood 1994; 83: weighing 20 kg or less. Blood 1995; 86: Saarinen UM, Wikstrom S, Makipernaa A et al. In vivo purg- 12 Gray III TF, Shea TC. Current status of peripheral blood proing of bone marrow in children with poor-risk neuroblastoma genitor cell transplantation. Semin Oncol 1994; 21: for marrow collection and autologous bone marrow transplan- 13 Di Caro A, Bostrom B, Moss TJ et al. Autologous peripheral tation. J Clin Oncol 1996; 14: blood cell transplantation in the treatment of advanced neuro- 27 Matthay KK, Atkinson JB, Stram DO et al. Patterns of relapse blastoma. Am J Pediatr Hematol Oncol 1994; 16: Lasky LC, Bostrom B, Smith J et al. Clinical collection and after autologous purged bone marrow transplantation for use of peripheral blood stem cells in pediatric patients. Transneuroblastoma: a Children s Cancer Group pilot study. J Clin plantation 1989; 47: Oncol 1993; 11: Demeocq F, Kanold J, Chassagne J et al. Successful blood 28 Seeger RC, Matthay KK, Villablanca JG et al. Intensive stem cell collection and transplant in children weighing less chemoradiotherapy and autologous bone marrow transplan- than 25 kg. Bone Marrow Transplant 1994; 13: tation (ABMT) for high risk neuroblastoma. Proc Am Soc Clin 16 Mitchell PL, Shepherd VB, Proctor HM et al. Peripheral blood Oncol 1991; 10: 310. stem cells used to augment autologous bone marrow transplanautologous purged bone marrow transplantation for neuroblas- 29 Matthay KK, Seeger RC, Reynolds CP et al. Allogeneic vs tation. Arch Dis Child 1994; 70: Miyajima Y, Horibe K, Fukuda M et al. Sequential detection toma: a report from the Children s Cancer Group. J Clin of tumor cells in the peripheral blood and bone marrow of Oncol 1994; 12: patients with stage IV neuroblastoma by the reverse-transcripand 30 Dini G, Lanino E, Garaventa A et al. Myeloablative therapy tion polymerase chain reaction for tyrosine hydroxylase unpurged autologous bone marrow transplantation for mrna. Cancer 1996; 77: poor-prognosis neuroblastoma: report of 34 cases. J Clin 18 Moss TJ, Sanders DG, Lasky LC, Bostrom B. Contamination Oncol 1991; 9: of peripheral blood stem cell harvests by circulating neuro- 31 Dini G, Lanino E, Garaventa A et al. Unpurged ABMT for blastoma cells. Blood 1990; 76: neuroblastoma: AIEOP-BMT experience. Bone Marrow 19 Moss TJ, Reynolds CP, Sather HN et al. Prognostic value of Transplant 1991; 7 (Suppl. 2): 92. immunocytologic detection of bone marrow metastases in 32 Kushner BH, O Reilly RJ, Mandell LR et al. Myeloablative neuroblastoma. New Engl J Med 1991; 324: combination chemotherapy without total body irradiation for 20 Brodeur GM, Seeger RC, Barrett A et al. International criteria neuroblastoma. J Clin Oncol 1991; 9: for diagnosis, staging, and response to treatment in patients 33 Kushner BH, Gulati SC, Kwon JH et al. High-dose melphalan with neuroblastoma. J Clin Oncol 1988; 6: with 6-hydroxydopamine-purged autologous bone marrow 21 Brodeur GM, Pritchard J, Berthold F et al. Revisions of the transplantation for poor-risk neuroblastoma. Cancer 1991; 68: international criteria for neuroblastoma diagnosis, staging, and response to treatment. J Clin Oncol 1993; 11: