Medulloblastoma Possible Benefits of High-Dose Chemotherapy

Bone Marrow Transplantation (2002) 30, 565–569  2002 Nature Publishing Group All rights reserved 0268–3369/02 $25.00 www.nature.com/bmt Medulloblastoma Possible beneﬁts of high-dose chemotherapy and autologous stem cell transplantation for adults with recurrent medulloblastoma

MI Zia1,2, P Forsyth1,3, A Chaudhry1,4, J Russell1,2,4 and DA Stewart1,2,4

1Department of Oncology, University of Calgary, Calgary, Alberta, Canada; 2Department of Medicine, University of Calgary, Calgary, Alberta, Canada; 3Department of Neurosciences, University of Calgary, Calgary, Alberta, Canada; and 4Alberta Blood and Marrow Transplant Program, Tom Baker Cancer Centre, Calgary, Alberta, Canada

Summary: The prognosis of adults with recurrent medulloblastoma remains poor, with median survival rates reported to be as In an attempt to improve the dismal prognosis of adults low as 5 months.1 Currently, there is no generally accept- with recurrent medulloblastoma, six patients were able therapy for tumor recurrences. The most active chemo- treated with aggressive salvage therapy including high- therapeutic agents include cisplatin or carboplatin,2–4 low- dose chemotherapy (HDCT) and autologous stem cell dose oral etoposide,5 PCV (procarbazine, CCNU, transplantation (ASCT). At relapse, all patients vincristine),6 MOPP (nitrogen mustard, vincristine, procar- underwent surgical debulking followed by HDCT/ASCT bazine, prednisone),7 and intraventricular and intrathecal and then radiotherapy when possible. The treatment therapy with Ara-C, methotrexate and thiotepa.8 Despite plan included two cycles of HDCT/ASCT; first with reasonably high response rates of 40–75% with these cyclophosphamide, etoposide and carboplatin (CECb) agents, no long-term disease-free survival has been and then 2 months later with cyclophosphamide and thi- reported.1–9 Given that these agents are amenable to sub- otepa (CT). Three of the six patients received the stantial dose-escalation, evaluation of high-dose chemo- planned therapy. One patient experienced severe tox- therapy (HDCT) and autologous stem cell transplantation icity requiring life-sustaining therapy. This patient (ASCT) has been suggested.9 Promising results with developed multi-organ dysfunction including multiple HDCT/ASCT prompted us to use this form of therapy for enhancing lesions in both cerebral hemispheres that young adults with recurrent medulloblastoma, who were slowly resolved over several months. Two other patients otherwise healthy and had no possibility of cure with did not mobilize sufficient stem cells for two ASCT pro- conventional dose therapy.10–15 cedures. They received one ASCT conditioned with Here, we report the outcomes of six adults who cyclophosphamide, thiotepa and carboplatin (CTCb). underwent HDCT (cyclophosphamide, carboplatin, thi- Three of six patients had a complete response (CR); the otepa, etoposide) with ASCT for medulloblastoma relapse. other three had a partial response (PR). Following the first ASCT, median duration of response was 13.5 months (range 9–29 months) and median survival was Patients and methods 21.5 months (range 12–42 months). There was no treatment-related mortality. We conclude that HDCT/ASCT Patients with CECb-CT or CTCb is active against recurrent medulloblastoma in adults and may be associated with Table 1 lists the clinical profile of patients 1–6. Between prolonged remissions. Multiple enhancing cerebral March 1995 and November 1999, these patients were lesions on brain MRI early post-HDCT/ASCT may be treated with this regimen at the Tom Baker Cancer Centre a consequence of the treatment rather than metastatic in Calgary, Alberta. This report includes follow-up of these disease. patients as of May 2001. The patient sample included five Bone Marrow Transplantation (2002) 30, 565–569. males and one female with a median age of 25 years (range: doi:10.1038/sj.bmt.1703725 8–33 years) at the time of initial diagnosis. All patients Keywords: medulloblastoma; autologous stem cell trans- were initially treated with surgery and radiation Ϯ pre-radi- plantation; neurotoxicity ation chemotherapy (three courses of vincristine and carboplatin). All patients had histologically documented recurrent medulloblastoma. The median age at relapse was 29.5 years (range: 21–35 years). The median time to relapse from initial diagnosis was 40 months (range: 19–204 months). Tumor staging at the time of relapse was done Correspondence: Dr D Stewart, Tom Baker Cancer Centre, 1331-29th using the Chang system. M0 indicates no evidence of meta- Street, NW, Calgary, Alberta, Canada, T2N 4N2 static disease, M1 is classified as positive CSF cytology, Received 30 January 2002; accepted 21 May 2002 M2 as gross nodular seeding of cerebral or cerebellar subar- HDCT and ASCT for recurrent medulloblastoma MI Zia et al 566 Table 1 Patient clinical profiles

Patient Gender/Age Surgery RT (Gy) Age at Time to M Time from Histology No. at Initial at Initial craniospinal/ relapse relapse stage at relapse to Dx Dx primary (months) relapse protocol at initial Dx (months)

1 M/33 Total resection 36/54 35 24 2 2 DM 2 M/8 Subtotal resection 20/55 21 168 0 2 DM 3 M/10 Total resection 40/35 27 204 2 16 M 4 F/30 Total resection 36/50 (+ Cx) 32 19 0 22 M 5 M/31 Total resection Ϫ/55 35 50 1 12 DM 6 M/20 Subtotal resection 36/56 (+ Cx) 22 30 3 2 M

Dx ϭ diagnosis; M ϭ male; F ϭ, female; RT ϭ radiation therapy; Cx ϭ chemotherapy (vincristine and carboplatin); DM ϭ desmoplastic medulloblastoma; M ϭ ‘classic’ medulloblastoma.

achanoid space or in the third or lateral ventricles, M3 as Response assessment gross nodular seeding of the spinal subarachnoid space, and M4 as extraneuroaxial metastases. At the time of relapse, Patients were assessed by the use of brain and spine MRI the M stage was M0 in two patients, M1 in one patient, after the first and second cycles of HDCT/ASCT, along M2 in two patients, M3 in one patient and M4 in no patient. with CSF cytology if there were prior evidence of metasta- Histologically, three patients had the classic form of medul- sis. Standard radiographic criteria were used to determine loblastoma, whereas the other three had the desmoplastic response. Complete response (CR) was defined as complete variant. resolution of all lesions and no clinical progression. Partial After relapse, three of the six patients underwent surgical response (PR) was defined as a >50% reduction in the pro- resection, two had a combination of surgery and chemo- duct of the maximum perpendicular diameters of all lesions therapy, and one patient had chemotherapy alone. None of and no clinical progression. Stable disease (SD) was the six patients received radiation therapy. Despite these defined as <50% reduction and <25% increase in size with efforts, all patients had symptomatic disease at the time of no clinical progression, and progressive disease (PD) was protocol. The median time from radiographic relapse to the defined as >25% increase in size. The patients’ Karnofsky first cycle of HDCT/ASCT was 7 months (range: 2–22 scores at the time of relapse and shortly (<30 days) after months). All patients had a Karnofsky score of 60% or completing the course of chemotherapy were also recorded. greater before the administration of the protocol.

Results

Protocol Protocol delivery Three out of six patients (Nos 2, 3 and 5) received the The original protocol consisted of two cycles of high-dose planned two cycles of HDCT/ASCT with CECb followed chemotherapy as follows. Leukapheresis was performed in by CT. Two patients (Nos 4 and 6) did not mobilize suf- all patients after chemotherapy and G-CSF administration, ficient stem cells for two ASCT procedures (only 1.3 and leading to the collection and cryopreservation of stem cells. 3.2 ϫ 106 CD34 cells/kg collected). Both these patients had On days Ϫ6, Ϫ5, and Ϫ4 patients received carboplatin (600 received craniospinal radiotherapy in the past. Conse- mg/m2 over 1 h), etoposide (600 mg/m2 over 3 h) and quently, these two patients received only one cycle of cyclophosphamide (50 mg/kg over 1 h). Approximately HDCT using cyclophosphamide, carboplatin and thiotepa 96 h after completion of chemotherapy, the cryopreserved (same doses as above). One patient (No. 1) experienced stem cells were reinfused. On day +1, granulocyte colony- severe toxicity from CECb/ASCT and declined the second stimulating factor (G-CSF) was administered at a dose of + transplant. For all six patients, the CD34 cell dose ranged 300 ␮g (if less than 70 kg) or 480 ␮g (if over 70 kg) via from 1.3 to 7.5 ϫ 106 cells/kg for the first cycle of the subcutaneous route once daily, until the neutrophils + HDCT/ASCT. The CD34 cell dose for the three patients engrafted. who received the second ASCT ranged from 3.6 to 15 ϫ For the patients who had initially mobilized sufficient 106 CD34 cells/kg. stem cells and established hematopoietic recovery, the second cycle of chemotherapy was administered. On days Ϫ Ϫ Ϫ 2 7, 6, 5, thiotepa (250 mg/m over 24 h) and cyclopho- Toxicity sphamide (2 g/m2 over 24 h) were administered. About 120 h later, the remaining stem cells from the initial harvest Hematological recovery post-ASCT was prompt for all were reinfused. Once again, G-CSF was administered as patients. No patient experienced treatment-related mor- per the first cycle. tality. Only one patient experienced severe life-threatening,

Bone Marrow Transplantation HDCT and ASCT for recurrent medulloblastoma MI Zia et al 567 regimen-related toxicity (RRT). This patient developed a systemic inflammatory syndrome with multisystem dysfunction of renal, pulmonary, cardiovascular, gastrointes- tinal and central nervous systems. His carboplatin dose was based on the Calvert formula using a measured 24-h creatinine clearance of 230 ml/min and an area-under-the-curve (AUC) of 21. Despite the fact that his body surface area was 2.3 m2, he received a relatively high carboplatin dose of 1800 mg i.v. daily for 3 days; equivalent to a total dose of 2347 mg/m2. On the evening of the third chemotherapy day, well before he became neutropenic, he developed fever of 39.4°C and experienced persistent fever until day +7 post-transplant. He developed progressive azotemia over the first few days following HDCT with a peak creatinine of 244 ␮mol/l on day +2 post-transplant. On day +3he became delirious and by day +5 he was very drowsy, non- communicative, and bedridden. He did not develop focal neurological signs. A lumbar puncture revealed normal CSF cell count, protein, glucose, culture/sensitivity, viral studies and cytology. Neither a brain CT scan on day +3 nor a brain MRI scan on day +4 revealed any evidence of abscess, hemorrhage, stroke or tumor. On day +4he developed intermittent atrial fibrillation and cardiomegaly on chest X-ray. He was diagnosed with a systemic inflammatory syndrome and was treated with high-dose corticos- teroid therapy from day +4to+9, initially with dexame- thasone 20 mg i.v. followed by solumedrol 60 mg twice a day. His fever, azotemia and delirium improved fairly rap- idly over the next few days. His steroids were tapered over 3 days and then discontinued on day +13. On day +16 his fever returned, again with no focus of infection and nega- tive cultures. He was started on prednisone 50 mg daily on day +18 and defervesed once again. The patient was discharged on day +24 post-ASCT with ataxia and gen- eralized weakness requiring a walking aid. These symptoms resolved over the ensuing month after discharge, despite being tapered off steroids. Six weeks after discharge, this patient was noted for the first time to have multiple enhancing foci with surrounding edema in the cerebral hemispheres on a follow-up MRI (Figure 1). On T1-weighted images there was a grouping of lesions in the left corona radiata, two small enhancing lesions in the right superior Figure 1 (a) MRI brain showing multiple lesions in the right external frontal region, single lesions in the right inferior frontal, capsule following high-dose CECb/ASCT. (b) Follow-up MRI brain 15 right external capsule, and left medial temporal lobe. A months later shows resolution of lesions. small lesion was seen along the lateral aspect of the right middle cerebellar peduncle. The T2-weighted images dem- Response and survival onstrated significant edema surrounding the larger lesions. Table 2 summarizes the response data from the six patients. There was no particular pattern to the lesions such as periv- All patients eventually relapsed post-ASCT. The median entricular or perivascular. These abnormalitites were progression-free interval from the first ASCT was 13.5 initially thought to represent metastatic tumors. The lesions months and from the time of radiographic relapse was 23 gradually decreased in size over several months until they months (range: 15–41 months). For all patients, the median virtually disappeared some 15 months later. There were no total survival time from the time of the first cycle of abnormalities in renal function, liver enzymes, serum elec- HDCT/ASCT was 21.5 months and from the time of radio- trolytes, nor similar MRI findings in any other patient after graphic relapse was 26.5 months (range: 21–54 months). the HDCT/ASCT. Three patients (Nos 2, 3 and 5) proceeded to receive a second cycle of HDCT/ASCT at a median of 6 weeks Discussion (range: 4–7 weeks) after the first ASCT. No patient experienced grade 3 or 4 RRT with the second cycle of Despite advances in the treatment of medulloblastoma at HDCT/ASCT. initial diagnosis, the prognosis is dismal for patients that

Bone Marrow Transplantation HDCT and ASCT for recurrent medulloblastoma MI Zia et al 568 Table 2 Summary of responses to protocol

Patient Age Protocol Adjuvant CSF KPS (%) Best response/ Current No. Rx post cytology pre/post response status/ protocol pre/post protocol duration survival protocol (months) (months)

1 35 C1: Cyclo, Eto, Car Ϫ +/Ϫ 70/90 CR/13 Deceased/20 C2: None 2 21 C1: Cyclo, Eto, Cis RT (focal) Ϫ/Ϫ 60/70 PR/14 Deceased/23 C2: Cyclo, Thio 3 28 C1: Cyclo, Eto, Cis Microsurgical +/Ϫ 60/60 CR/9 Deceased/12 C2: Cyclo, Thio resection 4 34 C1: Cyclo, Thio, Car RT (focal) Ϫ/Ϫ 90/90 PR/12 AWSD/24* C2: None 5 36 C1: Cyclo, Eto, Car RT (CS) +/Ϫ 80/90 CR/29 Deceased/42 C2: Cyclo, Thio 6 22 C1: Cyclo, Thio, Car RT (CS) + Thio +/Ϫ 70/70 PR/19 AWSD/19a C2: None (it) ϫ 7 courses

C1 ϭ cycle 1; C2 ϭ cycle 2; Cyclo ϭ cyclophosphamide; Eto ϭ etoposide; Car ϭ carboplatin; Cis ϭ cisplatin; Thio ϭ thiotepa; Rx ϭ treatment; RT ϭ radiation therapy; CS ϭ craniospinal; CR ϭ complete response; KPS ϭ Karnofsky performance score; PR ϭ partial response; AWSD ϭ alive with stable disease. aOngoing survival.

relapse. Recurrent disease treated with surgery, radiation or population (median age at time of protocol of 31 years) as conventional chemotherapy has been shown to have a compared to previous studies. median survival time of 5 months with no long-term sur- Our results support the findings of Dunkel et al that the vivors.1 We present a small case series of six adults who use of high-dose chemotherapy in conjunction with stem underwent HDCT with ASCT for medulloblastoma relapse. cell rescue is associated with a favorable response with a The rationale for using HDCT with stem cell rescue for potential for long-term survival in some patients. Despite medulloblastoma is based on the fact that brain tumors are their advanced age, all of our patients responded to treat- sensitive to chemotherapeutic agents that have hematopo- ment with a minimum 50% reduction in tumor size. This ietic dose-limiting toxicity.16,17 By employing the use of a is an improvement in comparison to response rates ranging stem cell transplant to increase the tolerable dose of chemo- from 40 to 75% with conventional chemotherapy.2–9 therapy, the potential dose-dependent sensitivity of medul- Among patients who had a second recurrence, we report loblastoma can be fully exploited.10,18 an even longer median time to relapse post high-dose The rationale for specifically using cyclophosphamide, chemotherapy as compared to Dunkel et al (13.5 vs 7 carboplatin, etoposide and thiotepa in our regimen stems months), with two patients surviving with stable disease. It from their efficacy demonstrated in previous studies.4,11–13 is interesting to note that patients with metastatic disease Thiotepa in particular is a polyfunctional alkylating agent (>M0) had a comparable median time to relapse as com- that has shown significant activity against human medullo- pared to those with no metastatic disease (16 vs 13 months). blastoma cell lines both in vitro and in xenograft mod- This shows that this regimen has activity against both els.18,19 Both thiotepa and its active metabolite tepa cross aggressive and non-aggressive forms of the disease. Both the blood–brain barrier efficiently, making it a very promis- the classic form of medulloblastoma and the desmoplastic ing agent in the setting of CNS tumors.14,20–22 variant responded equally with a median time to relapse of One case report of two patients has shown that high-dose 12 and 14 months, respectively. chemotherapy consisting of busulfan and thiotepa followed One patient developed severe multi-organ dysfunction by infusion of autologous hematopoietic stem cells pro- early post-CECb/ASCT. A follow-up brain MRI (Figure 1) duced disease-free survival times of 20 and 27 months after done 6 weeks later showed peculiar enhancing lesions that bone marrow relapse of medulloblastoma.15 High-dose car- were initially thought to be metastases or toxicity from boplatin, thiotepa and etoposide with autologous stem-cell CECb. While peripheral neuropathy has been reported to rescue have also shown some activity against recurrent occur in as many as 6% of patients receiving carboplatin, medulloblastoma.11 High-dose chemotherapy is still CNS toxicity with enhancing foci has not been docu- thought to be an experimental procedure that has yet to mented.23 The gradual resolution of these lesions over sev- show a clear advantage over conventional chemotherapy in eral months was as mysterious as their initial appearance. adult recurrent medulloblastoma. He did not receive long-term antibiotics of any sort post- Our study parallels the report by Dunkel et al11 in 1998 ASCT, not even bactrim for PCP prophylaxis. It seems very that uses a similar regimen. We employed the use of two unlikely to us that his neurological complications could cycles followed by stem cell rescue with a total etoposide have been due to a severe infection that spontaneously dose of 1800 mg/m2 and a total thiotepa dose of 750 mg/m2. resolved without any specific therapy, particularly in the In contrast, Dunkel et al administered a total etoposide dose setting of immunosuppression post-ASCT. We feel that a of 750 mg/m2 and a total thiotepa dose of 900 mg/m2. Our more likely explanation was that the neurotoxicity was part study is also unique because we treated an older patient of the multisystem regimen-related toxicity experienced by

Bone Marrow Transplantation HDCT and ASCT for recurrent medulloblastoma MI Zia et al 569 the patient. Specifically, we feel the carboplatin was the 7 Cangir A, Ragab A, Steuber P et al. Combination chemo- likely culprit because of its known association with periph- therapy with vincristine, procarbazine, prednisone with or eral neurotoxicity, and because our patient received a parti- without nitrogen mustard (MOPP versus OPP) in children cularly high dose of carboplatin relative to the other with recurrent brain tumors. Med Ped Oncol 1984; 12:1–3. patients we have treated. Nevertheless, we cannot prove our 8 Levin V, Vestnys P, Edwards M et al. Improvement in survival produced by sequential therapies in the treatment of opinion because the lesions were never biopsied. The case recurrent medulloblastoma. Cancer 1983; 51: 1364–1370. does illustrate, however, the important fact that enhancing 9 Brandes AA, Palmisano V, Monfardini S. Medulloblastoma in lesions seen on brain MRI scans early post-ASCT may not adults: clinical characteristics and treatment. Cancer Treat Rev be due to metastatic cancer. 1999; 25:3–12. The limitations of this study include inconsistencies in 10 Lundberg JH, Weissman DE, Beatty PA et al. Treatment of regimen administration. One patient did not receive thi- recurrent metastatic medulloblastoma with intensive chemo- otepa at all and two did not receive any etoposide. How- therapy and allogenic bone marrow transplantation. J Neu- ever, all patients received at least one of the above. In rooncol 1992; 13: 151–155. addition, four patients received adjuvant radiation therapy 11 Dunkel IJ, Boyett JM, Yates A et al. High-dose carboplatin, and one patient had a microsurgical resection after the thiotepa, and etoposide with autologous stem-cell rescue for administration of high-dose chemotherapy with stem cell patients with recurrent medulloblastoma. J Clin Oncol 1998; 16: 222–228. rescue. However, the patient who did not receive any adju- 12 Graham ML, Herndon JE II, Casey S et al. High-dose chemo- vant treatment had a comparable median time to relapse as therapy with autologous stem-cell rescue in patients with the others (13 vs 14 months). Another limitation is that recurrent and high-risk pediatric brain tumors. J Clin Oncol this is a small retrospective study that does not allow us to 1997; 15: 1814–1823. determine which patient population would benefit the most 13 Finlay JL, Goldman S, Wong MC et al. Pilot study of high- from this type of treatment. dose thiotepa and etoposide with autologous bone marrow res- Although HDCT/ASCT does not cure relapsed medullo- cue in children and young adults with recurrent CNS tumors. blastoma in adults, it does result in high response rates and J Clin Oncol 1996; 14: 2495–2503. possibly some improvement in survival. Encouraging 14 Ascensao J, Ahmed T, Feldman E et al. High-dose thiotepa results from a pilot study of high-dose thiotepa and etopo- with autologous bone marrow transplantation for patients with side with ASCT in children and young adults with recurrent astrocytoma grade III–IV (glioma): A promising approach. high-grade gliomas has merited the evaluation of such Proc Am Soc Clin Oncol 1989; 8: 90. 15 Millot F, Delval O, Giraud C et al. High-dose chemotherapy agents as part of first-line treatment in newly diagnosed 13 with hematopoietic stem cell transplantation in adults with patients with these tumors. Similarly, since HDCT/ASCT bone marrow relapse of medulloblastoma: report of two cases. is very active even in the setting of recurrent medulloblas- Bone Marrow Transplant 1999; 24: 1347–1349. toma, perhaps the efficacy and tolerability would be 16 Kessinger A. High-dose chemotherapy with autologous bone enhanced if this regimen were to be employed as part of marrow rescue for high grade gliomas of the brain: a potential front-line treatment. The concern in this regard would be for improvement in therapeutic results. Neurosurgery 1984; subjecting patients to excess toxicity who might have other- 15: 747–750. wise been cured by less intensive treatment. However, 17 Fine HA, Antman KH. High-dose chemotherapy with autolog- long-term follow-up suggests frequent late relapses and ous bone marrow transplantation in the treatment of high overall low cure rates for most patients with medullo- grade astrocytomas in adults: therapeutic rationale and clinical blastoma who receive conventional treatment.1,9 experience. Bone Marrow Transplant 1992; 10: 315–321. 18 Friedman HS, Colvin OM, Skapek SX et al. Experimental chemotherapy of human medulloblastoma cell lines and trans- References plantable xenografts with bifunctional alkylating agents. Can- cer Res 1988; 48: 4189–4195. 1 Torres CF, Rebsamen S, Silber JH et al. Surveillance scanning 19 Friedman HS, Colvin OM, Ludeman SM et al. Experimental of children with medulloblastoma. New Engl J Med 1994; 330: chemotherapy of human medulloblastoma with classical 892–895. alkylators. Cancer Res 1986; 46: 2827–2833. 2 Walker R, Allen J. Cisplatin in the treatment of recurrent 20 Heideman RL, Cole DE, Balis F et al. Phase I and pharmacok- childhood primary brain tumors. J Clin Oncol 1988; 6:62–66. inetic evaluation of thiotepa in the cerebrospinal fluid and 3 Khan AB, D’Souza BJ, Wharam MD. Cisplatin therapy in plasma of pediatric patients: evidence for dose-dependent recurrent childhood brain tumors. Cancer Treat Rep 1982; 66: plasma clearance of thiotepa. Cancer Res 1989; 49: 736–741. 2013–2020. 21 Tartaglia RL, Cole DE, Heideman R et al. Chemosensitivity 4 Allen J, Walker R, Luks E et al. Carboplatin in recurrent of central nervous system tumors to thiotepa and tepa. Proc childhood brain tumors. J Clin Oncol 1987; 5: 459–463. Am Assoc Cancer Res 1989; 30: 461. 5 Ashley DM, Meier L, Kerby T et al. Response of recurrent 22 Heideman RL, Packer RJ, Reaman GH et al. A phase II evalu- medulloblastoma to low-dose oral etoposide. J Clin Oncol ation of thiotepa in pediatric central nervous system malig- 1996; 14: 1922–1927. nancies. Cancer 1993; 72: 271–275. 6 Crafts D, Levin V, Edwards M et al. Chemotherapy of recur- 23 Screnci D, McKeage MJ. Platinum neurotoxicity: clinical pro- rent medulloblastoma with combined procarbazine, CCNU, files, experimental models and neuroprotective approaches. J and vincristine. J Neurosurg 1978; 49: 589–592. Inorg Biochem 1999; 77: 105–110.

Bone Marrow Transplantation