1200 www.thelancet.com/oncology Vol 14 November 2013

Articles

Recurrence patterns across medulloblastoma subgroups: an integrated clinical and molecular analysisVijay Ramaswamy, Marc Remke, Eric Bouff et, Claudia C Faria, Sebastien Perreault, Yoon-Jae Cho, David J Shih, Betty Luu, Adrian M Dubuc, Paul A Northcott, Ulrich Schüller, Sridharan Gururangan, Roger McLendon, Darell Bigner, Maryam Fouladi, Keith L Ligon, Scott L Pomeroy, Sandra Dunn, Joanna Triscott, Nada Jabado, Adam Fontebasso, David T W Jones, Marcel Kool, Matthias A Karajannis, Sharon L Gardner, David Zagzag, Sofi a Nunes, José Pimentel, Jaume Mora, Eric Lipp, Andrew W Walter, Marina Ryzhova, Olga Zheludkova, Ella Kumirova, Jad Alshami, Sidney E Croul, James T Rutka, Cynthia Hawkins, Uri Tabori, Kari-Elise T Codispoti, Roger J Packer, Stefan M Pfi ster, Andrey Korshunov, Michael D Taylor

SummaryBackground Recurrent medulloblastoma is a therapeutic challenge because it is almost always fatal. Studies have confi rmed that medulloblastoma consists of at least four distinct subgroups. We sought to delineate subgroup-specifi c diff erences in medulloblastoma recurrence patterns.

Methods We retrospectively identifi ed a discovery cohort of all recurrent medulloblastomas at the Hospital for Sick Children (Toronto, ON, Canada) from 1994 to 2012 (cohort 1), and established molecular subgroups using a nanoString-based assay on formalin-fi xed paraffi n-embedded tissues or frozen tissue. The anatomical site of recurrence (local tumour bed or leptomeningeal metastasis), time to recurrence, and survival after recurrence were assessed in a subgroup-specifi c manner. Two independent, non-overlapping cohorts (cohort 2: samples from patients with recurrent medulloblastomas from 13 centres worldwide, obtained between 1991 and 2012; cohort 3: samples from patients with recurrent medulloblastoma obtained at the NN Burdenko Neurosurgical Institute [Moscow, Russia] between 1994 and 2011) were analysed to confi rm and validate observations. When possible, molecular subgrouping was done on tissue obtained from both the initial surgery and at recurrence.

Results Cohort 1 consisted of 30 patients with recurrent medulloblastomas; nine with local recurrences, and 21 with metastatic recurrences. Cohort 2 consisted of 77 patients and cohort 3 of 96 patients with recurrent medulloblastoma. Subgroup affi liation remained stable at recurrence in all 34 cases with available matched primary and recurrent pairs (fi ve pairs from cohort 1 and 29 pairs from cohort 2 [15 SHH, fi ve group 3, 14 group 4]). This fi nding was validated in 17 pairs from cohort 3. When analysed in a subgroup-specifi c manner, local recurrences in cohort 1 were more frequent in SHH tumours (eight of nine [89%]) and metastatic recurrences were more common in group 3 and group 4 tumours (17 of 20 [85%] with one WNT, p=0·0014, local vs metastatic recurrence, SHH vs group 3 vs group 4). The subgroup-specifi c location of recurrence was confi rmed in cohort 2 (p=0·0013 for local vs metastatic recurrence, SHH vs group 3 vs group 4,), and cohort 3 (p<0·0001). Treatment with craniospinal irradiation at diagnosis was not signifi cantly associated with the anatomical pattern of recurrence. Survival after recurrence was signifi cantly longer in patients with group 4 tumours in cohort 1 (p=0·013) than with other subgroups, which was confi rmed in cohort 2 (p=0·0075), but not cohort 3 (p=0·70).

Interpretation Medulloblastoma does not change subgroup at the time of recurrence, reinforcing the stability of the four main medulloblastoma subgroups. Signifi cant diff erences in the location and timing of recurrence across medulloblastoma subgroups have potential treatment ramifi cations. Specifi cally, intensifi ed local (posterior fossa) therapy should be tested in the initial treatment of patients with SHH tumours. Refi nement of therapy for patients with group 3 or group 4 tumours should focus on metastases.

Funding Canadian Institutes of Health Research, National Institutes of Health, Pediatric Brain Tumor Foundation, Garron Family Chair in Childhood Cancer Research at The Hospital for Sick Children and The University of Toronto.

IntroductionMedulloblastoma is the most common malignant brain tumour of childhood.1,2 With multimodal therapy, consisting of surgery, craniospinal irradiation, and adjuvant chemotherapy, 5-year overall survival approaches 85% for average-risk disease and 70% for high-risk disease.3–5 However, recurrent medulloblastoma is a great challenge, because it is almost always fatal in previously irradiated patients despite a multitude of therapies including re-resection, re-irradiation, high-dose

chemotherapy with autologous stem-cell support, and enrolment in clinical trials.6 Integrative genomic studies have shown that medulloblastoma consists of at least four subgroups (WNT, SHH, group 3, and group 4) that are clinically, transcriptionally, and genetically distinct.2,7–18

Of these four subgroups, patients with WNT subgroup tumours have an excellent prognosis whereas patients with group 3 tumours have the worst prognosis and more commonly present with disseminated disease at diagnosis.1,2,12 Although these integrative genomic studies

Lancet Oncol 2013; 14: 1200–07

Published OnlineOctober 17, 2013

http://dx.doi.org/10.1016/S1470-2045(13)70449-2

See Comment page 1147

Division of Neurosurgery (V Ramaswamy MD,

M Remke MD, C C Faria MD, D J Shih MSc, A M Dubuc PhD,

Prof J T Rutka MD, Prof C Hawkins MD,

Prof M D Taylor MD), Division of Pediatric Hematology/Oncology (Prof E Bouff et MD, U Tabori MD),

Labatt Brain Tumour Research Centre (V Ramaswamy, M Remke,

Prof E Bouff et, C C Faria, D J Shih, B Luu BSc, A M Dubuc,

Prof S E Croul MD, Prof J T Rutka, Prof C Hawkins, U Tabori,

Prof M D Taylor), Hospital for Sick Children, Toronto, ON,

Canada; Department of Laboratory Medicine and

Pathobiology, University of Toronto, Toronto, ON, Canada

(V Ramaswamy, M Remke, D J Shih, A M Dubuc,

Prof J T Rutka, Prof M D Taylor); Division of Neurosurgery (C C Faria), Laboratory of

Neuropathology (J Pimentel MD), Hospital de

Santa Maria, Centro Hospitalar Lisboa Norte, EPE, Lisbon,

Portugal; Department of Neurology, Stanford University,

Palo Alto, CA, USA (S Perreault MD, Y-J Cho MD);

Division of Pediatric Neurooncology

(P A Northcott PhD, D T W Jones PhD, M Kool PhD, Prof S M Pfi ster MD), Clinical

Cooperation Unit Neuropathology

(Prof A Korshunov MD), German Cancer Research Center (DKFZ),

Heidelberg, Germany; Center for Neuropathology,

Ludwig-Maximilians-University, Munich, Germany

(U Schüller MD); The Preston Robert Tisch Brain Tumor

Center, Duke University Medical

Articles

www.thelancet.com/oncology Vol 14 November 2013 1201

Center, Durham, NC, USA (Prof S Gururangan MD, Prof R McLendon MD, Prof D Bigner MD, E Lipp MD); Division of Hematology/Oncology, Cincinnati Children’s Hospital, Cincinnati, OH, USA (Prof M Fouladi MD); Department of Pathology (K L Ligon MD), Department of Neurology (Prof S L Pomeroy MD), Harvard Medical School, Brigham and Women’s Hospital, Boston, MA, USA; Boston Children’s Hospital, Boston, MA, USA (K L Ligon); Department of Medical

have shown that there are signifi cant diff erences in survival between the four subgroups, little is known with respect to subgroup-specifi c anatomical and temporal characteristics of recurrence.

In glioblastoma, molecular subgroup affi liation can change at recurrence, partly because of intratumoural heterogeneity based on geographical location.19,20 Although medulloblastoma subgroups have been shown to arise from distinct cells of origin, the stability of subgroup affi liation at recurrence remains unknown.21–23

Moreover, the clinical behaviour of the individual subgroups at recurrence has yet to be established. As such, an understanding of subgroup-specifi c temporal and spatial details could help to develop treatment of recurrent medulloblastoma, because the next generation of subgroup-specifi c clinical trials will probably be based initially in the context of recurrent medulloblastoma.

We aimed to characterise the subgroup-specifi c clinical patterns of recurrence in medulloblastoma.

MethodsPatientsWe assembled a discovery cohort (cohort 1) and, to account for unobserved variables and potential bias due to diff erent subgrouping methods, two non-overlapping validation cohorts (cohorts 2 and 3). Cohort 1 consisted of all patients with medulloblastoma with either frozen or formalin-fi xed paraffi n-embedded (FFPE) material along with clinical variables and survival data, treated between 1994 and 2012, at the Hospital for Sick Children (Toronto, ON, Canada).

Cohort 2 consisted of samples from patients with recurrent medulloblastomas from 13 centres, obtained between 1991 and 2012 (Munich University Hospital, Munich, Germany; Hospital de Santa Maria in Lisbon, Lisbon, Portugal; Duke University Medical Center, Durham, NC, USA; Lucille Packard Children’s Hospital, Palo Alto, CA, USA; Children’s National Medical Center, Washington, DC, USA; British Columbia Children’s Hospital, Vancouver, BC, Canada; A I duPont Hospital for Children, Wilmington, DE, Canada; Brain Tumour Bank of Canada, London, ON, Canada; Hospital Sant Joan de Deu de Barcelona, Barcelona, Spain; McGill University Health Centre, Montreal, QC, Canada; Children’s Hospital Boston, Boston, MA, USA; New York University Langone Medical Center, New York City, NY, USA; and Cincinnati Children’s Hospital, Cincinnati, OH, USA).

Cohort 3 consisted of samples from patients with recurrent medulloblastoma obtained at the NN Burdenko Neurosurgical Institute (Moscow, Russia) between 1994 and 2011.

Worldwide, only a minority of patients with relapsed medulloblastoma undergo a second surgery at recurrence. Therefore, most patients do not have tissue available at recurrence. We obtained matched samples from diagnosis and recurrence whenever possible, and

assembled cohort 1 with the goal of acquiring as many of these paired samples as possible.

The research ethics boards at all participating centres approved the study and all samples and clinical information were obtained with consent in accordance with the research ethics board at the Hospital for Sick Children and collaborating centres.

ProceduresWe extracted RNA from fresh frozen tissue using the guanidinium thiocyanate-phenol-chloroform extraction method (TRIzol, Invitrogen, Carlsbad, USA), according to the manufacturer’s instructions. RNA from FFPE samples (fi ve to seven paraffi n sections per sample or the

Cohort 1 (n=30) Cohort 2 (n=77) Cohort 3 (n=96) p value

Male 20 (67%) 39 (70%)* 61 (64%) 0·76

Female 10 (33%) 17 (30%)* 35 (36%) ··

Age (years)

Median 5·4 (3·6–8·8) 7·0 (3·9–11·6) 7·0 (4·0–11·0) 0·31

<3 6 (20%) 10 (13%) 12 (13%) ··

3–16 24 (80%) 53 (69%) 81 (84%) ··

>16 0 14 (18%) 3 (3%) ··

Histology 0·19

LCA 6 (20%) 18 (30%)† 26 (27%) ··

Classic 21 (70%) 31 (52%)† 63 (66%) ··

Desmoplastic or MBEN 3 (10%) 11 (18%)† 7 (7%) ··

Metastases at diagnosis 8 (27%) 19 (32%)‡ 44 (46%) 0·11

Extent of resection 0·26

Gross-total 21 (75%)§ ·· 60 (63%) ··

Sub-total 7 (25%)§ ·· 36 (38%) ··

Treatment 0·49

CSI with or without chemotherapy 22 (80%)§ 50 (70%) 75 (78%) ··

Chemotherapy only 6 (20%)§ 21 (30%)¶ || 21 (22%) ··

Pattern of recurrence 0·20

Tumour bed 9 (30%) 26 (34%) 24 (25%) ··

Metastatic 18 (60%) 42 (54%) 49 (51%) ··

Mixed 3 (10%) 9 (12%) 23 (24%) ··

Subgroup 0·15

WNT 1 (3%) 0 2 (2%) ··

SHH 11 (37%) 30 (39%) 21 (22%) ··

Group 3 9 (30%) 22 (29%) 37 (39%) ··

Group 4 9 (30%) 25 (32%) 36 (38%) ··

Time to recurrence (months) 18·3 (11·4–44·8) 19·9 (10·9–32·6) 12·0 (8·0 –21·9) 0·0038

Overall follow-up time (months) 32·9 (13·7–72·3) 33·2 (17·0–57·8) 30·0 (18·0 –64·7) 0·91

Survival 0·38

Alive 6 (20%) 24 (33%) 32 (33%) ··

Dead 24 (80%) 49 (67%)** 64 (67%) ··

Data are n (%) or median (IQR). Some percentages do not total 100 because of rounding. p values are from Fisher’s exact test for categorical variables, and Kruskal-Wallis test for continuous variables. LCA=large cell/anaplastic histology. MBEN=medulloblastoma with extensive nodularity. CSI=craniospinal irradiation. *Sex unavailable for 21 cases. †Histology unavailable for 17 cases in validation cohort 1. ‡Metastases at diagnosis unavailable in 18 cases. §Data missing for two patients. ¶Two patients in validation cohort 1 received radiation therapy to the posterior fossa only followed by chemotherapy. ||Data missing for six patients. **Data missing for four patients.

Table 1: Demographics of the three independent cohorts of recurrent medulloblastoma

Articles

1202 www.thelancet.com/oncology Vol 14 November 2013

Oncology, Dana-Farber Cancer Institute, Boston, MA, USA

(K L Ligon); Broad Institute of Massachusetts Institute of

Technology and Harvard, Cambridge, MA, USA

(Prof S L Pomeroy); Division of Hematology/Oncology, British

Columbia Children’s Hospital, Vancouver, BC, Canada

(S Dunn PhD, J Triscott BSc); Division of Pediatric

Hematology/Oncology, Montreal Children’s Hospital,

Montreal, PQ, Canada (N Jabado MD, A Fontebasso BSc,

J Alshami BSc); Division of Pediatric Hematology/Oncology

(M A Karajannis MD, S L Gardner MD), Department of

Pathology (Prof D Zagzag MD), Department of Neurosurgery (Prof D Zagzag), NYU Langone Medical Center, New York, NY,

USA; Unidade de Neuro-Oncologia Pediátrica, Instituto

Português de Oncologia de Lisboa Francisco Gentil, Lisbon,

Portugal (S Nunes MD); Department of Oncology,

Hospital Sant Joan de Deu de Barcelona, Barcelona, Spain

(J Mora MD); A I duPont Hospital for Children, Wilmington, DE,

USA (A W Walter MD); Department of

Neuropathology, NN Burdenko Neurosurgical Institute,

Moscow, Russia (M Ryzhova MD); Department of

Pediatric Neurooncology, Dmitry Rogachev Federal

Research Center of Pediatric Hematology, Oncology and

Immunology, Moscow, Russia (O Zheludkova MD,

E Kumirova MD); Center for Neuroscience and Behavioral Medicine, Children’s National Medical Center, Washington,

equivalent in scrolls) was extracted using the RNeasy FFPE kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions.

Subgroup affi liation was established using nanoString targeted gene-expression profi ling, as previously described, in all cases from cohorts 1 and 2.14 Subgroup determination for cohort 3 was done in all samples by immunohistochemistry using the four antibody method, as described previously (WNT=nuclear β-catenin, SHH=SFRP1, group 3=NPR3, group 4=KCNA1).9,10,17 Cases were ascribed a subgroup if they were immunoreactive for only one marker.

For cohorts 1 and 3, results of neuroimaging to establish the pattern of recurrence were reviewed by a multidisciplinary tumour board, which included a neuroradiologist masked to molecular subgroup, at both the Hospital for Sick Children and the Burdenko Neurosurgical Institute. The pattern of recurrence in cohort 2 was provided by individual contributing institutions on the basis of a review of neuroimaging.

Statistical analysisWe established survival after recurrence between subgroups using the Kaplan-Meier method and used a log-rank test to test for diff erences between subgroups. We plotted time to recurrence using the Kaplan-Meier method and we used the generalised Wilcoxon test to look for diff erences between subgroups attributable to the absence of censoring in these data. We used Fisher’s exact test to compare binary and categorical characteristics of patients between subgroups and cohorts. We analysed continuous variables using the Kruskal-Wallis test or Mann-Whitney U test. We used StataSE (version 12) for all statistical analyses, except nanoString class prediction. Principle component analysis was done in the Partek Genomic Suite and negative matrix factorisation was done using Gene Pattern 2.0.24 We used R (version 2.15) for nanoString-based class prediction and normalisation of nanoString data, as previously described.14

Role of the funding sourceThe funding sources of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit for publication.

ResultsCohort 1 consisted of 131 patients, of whom 30 had recurrent medulloblastoma. Cohort 2 consisted of 77 patients and cohort 3 of 96 patients, all of whom had recurrent medulloblastoma. At fi rst diagnosis, the WNT subgroup was rare, occurring in one (3%) of 30 patients in cohort 1, no patients in cohort 2, and two (2%) of 96 patients in cohort 3. Specimens from the remaining patients were distributed roughly equally between SHH, group 3, and group 4 in all cohorts (table 1). Of the 96 cases in cohort 3 subgrouped by immunohistochemistry, none were immuno reactive for more than one marker, and 39 were also subgrouped using whole genome expression profi ling or nanoString without any reclassifi cations of subgroup.

Table 1 shows the demographics of all three cohorts. The adult age group (>16 years) was over-represented in cohort 2, but median age did not diff er between groups. Metastatic dissemination at diagnosis was higher in cohort 3 than in either of the other cohorts, although the diff erence was not signifi cant. Median time to recurrence was signifi cantly shorter in cohort 3 than in the other cohorts (table 2). Histological classifi cation did not diff er across the three cohorts. When available, treatment at diagnosis for all three cohorts are summarised in table 1 and the appendix (appendix pp 8–9). Adjuvant chemotherapy regimens were all cisplatin-based across the three cohorts. Treatment in cohort 3 was uniform as per the German HIT protocols, as previously described.25,26

To establish whether subgroup affi liation remains stable at recurrence, 34 paired samples (fi ve pairs from cohort 1, and 29 pairs from various centres as part of cohort 2) with FFPE tissue available from the initial surgery and recurrence were analysed by nanoString. At recurrence, tissue was obtained from the primary site in 16 tumours; 18 samples were from biopsies of leptomeningeal metastases. In all 34 paired samples, subgroup affi liation remained stable between diagnosis and the corresponding local or metastatic recurrence (15 SHH, fi ve group 3, 14 group 4; fi gure 1; appendix p 10). One local recurrence was subgrouped as a group 4 at diagnosis and SHH at recurrence; however, upon review of the histological specimen by a senior neuro pathogist (AK), the recurrence was established to be a radiation-induced secondary glioblastoma and this patient was therefore excluded from all other analyses (appendix p 12). To further confi rm the fi nding that subgroup affi liation does not change, an orthogonal technique of subgroup determination using immuno histochemistry was done on 17 paired samples from the initial surgery and recurrence as part of cohort 3, using a four antibody method as previously described.17,27

Cohort 1 Cohort 2* Cohort 3 p value

Time to recurrence

All subgroups 1·49 (1·09–1·90) 1·65 (1·25–1·92) 1·00 (0·91–1·09) 0·0038

SHH 0·96 (0·85–1·08) 1·34 (1·00–2·25) 1·00 (0·81–1·19) 0·58

Group 3 1·55 (0·00–3·21) 1·11 (0·56–1·51) 0·92 (0·78–1·06) 0·36

Group 4 3·88 (3·30–4·47) 2·18 (1·71–2·74) 1·17 (0·68–1·66) 0·00018

Survival after recurrence

All subgroups 0·70 (0·00–1·99) 1·78 (1·08–2·66) 1·67 (0·65–2·68) 0·47

SHH 0·14 (0·08–0·19) 0·98 (0·47–1·68) 0·92 (0·04–1·79) 0·0066

Group 3 0·28 (0·14–0·42) 1·52 (0·47–2·41) 1·75 (0·47–3·03) 0·012

Group 4 2·14 (0·99–3·29) 3·73 (2·33–7·26) 1·75 (0·50–2·99) 0·34

Data are median years (95% CI). p values calculated using the generalised Wilcoxon test. “All subgroups” includes WNT patients. *Data available for 73 of 77 patients.

Table 2: Subgroup-specifi c median time to recurrence and survival post-recurrence across all three cohorts

Articles

www.thelancet.com/oncology Vol 14 November 2013 1203

DC, USA (K-E T Codispoti MD, Prof R J Packer MD); Heidelberg University Hospital, Department of Pediatric Hematology and Oncology, Heidelberg, Germany (Prof S M Pfi ster); and Department of Neuropathology, University of Heidelberg, Heidelberg, Germany (Prof A Korshunov)

Correspondence to:Prof Michael D Taylor, Hospital for Sick Children, Toronto, ON, Canadamdtaylor@sickkids.ca

See Online for appendix

In all 17 paired samples, the initial pattern of immunoreactivity remained stable at recurrence, with one WNT, six SHH, four group 3, and six group 4 patients, demonstrating the same subgroup affi liation at recurrence as at diagnosis (appendix p 11). Therefore, we conclude that medulloblastoma does not change subgroup at the time of recurrence.

Median time to recurrence for cohort 1 was 1·49 years (95% CI 1·09–1·90; table 2). When time to recurrence in the cohort 1 was analysed in a subgroup-specifi c manner, group 4 tumours recurred signifi cantly later than both group 3 and SHH tumours (p=0·0080, generalised Wilcoxon; appendix p 13). Of the 30 recurrences in cohort 1, six patients were alive at the time of the study with a median overall survival after recurrence of 0·70 years (95% CI 0·00–1·99, appendix p 14), calculated as the time to death or last follow-up after fi rst recurrence. There were two long-term survivors (>5 years), a WNT subgroup patient who survived for 8·47 years after recurrence and a group 4 subgroup patient who survived for 13·75 years after recurrence. Both patients were initially treated with reduced-dose craniospinal irradiation and chemotherapy. The WNT long-term survivor relapsed 1·49 years after diagnosis and was

salvaged with re-irradiation and high-dose chemotherapy with autologous stem-cell support, followed by temozolomide plus etoposide at a second recurrence 1 year later. The group 4 long-term survivor was salvaged with re-resection followed by high-dose chemotherapy with autologous stem-cell support. When overall survival after recurrence in cohort 1 was analysed in a subgroup-specifi c manner, patients with group 4 tumours had a signifi cantly longer overall survival after recurrence than did those with group 3 and SHH tumours (fi gure 2A; p=0·013, log-rank; table 2).

The two fi ndings that patients with group 4 tumours recur later and have longer overall survival after recurrence than group 3 and SHH tumours were confi rmed in cohort 2 but not in cohort 3 (fi gure 2B, 2C; table 2). For all subgroups, time to recurrence was signifi cantly shorter for cohort 3 than for the cohort 1 or cohort 2 (table 2), suggesting that diff erences in therapy might account for this discrepancy; however, treatment regimens were much the same across cohorts (table 1). Therefore time to recurrence and overall survival after recurrence might still diff er signifi cantly between subgroups.

Survival data were available in 73 of 77 cases in cohort 2. Four patients were alive 5 years after recurrence: one

WIF1

TNC

GAD1

DKK2

EMX2

PDLIM3

EYA1

HHIP

ATOH

1SFRP1

IMPG2

GABRA5EYS

MAB21L2

NRL

NPR3

KCNA1

EOMES

KHDRBS2

RBM24

UNC5D

OAS1

WIF1

TNC

GAD1

DKK2

EMX2

PDLIM3

EYA1

HHIP

ATOH

1SFRP1

IMPG2

GABRA5EYS

MAB21L2

NRL

NPR3

KCNA1

EOMES

KHDRBS2

RBM24

UNC5D

OAS1

1SHH SHH2SHH SHH3SHH SHH4SHH SHH5SHH SHH6SHH SHH7SHH SHH8SHH SHH9SHH SHH

10SHH SHH11SHH SHH12SHH SHH13SHH SHH14SHH SHH15SHH SHH16Group 3 Group 317Group 3 Group 318Group 3 Group 319Group 3 Group 320Group 3 Group 321Group 4 Group 422Group 4 Group 423Group 4 Group 424Group 4 Group 425Group 4 Group 426Group 4 Group 427Group 4 Group 428Group 4 Group 429Group 4 Group 430Group 4 Group 431Group 4 Group 432Group 4 Group 433Group 4 Group 434Group 4 Group 4

Diagnosis Patient Recurrence

Figure 1: Subgroup affi liation at recurrenceHeatmap of relative gene expression of 22 nanoString probes normalised to three housekeeping genes (ACTB, GAPDH, LDHA) across 34 medulloblastoma samples collected at diagnosis, and 34 matched recurrent medulloblastoma samples horizontally aligned. At recurrence, tissue was obtained from the primary site in 12 of 15 SHH tumours and four of 14 group 4 tumours; the remaining samples are from biopsies of leptomeningeal metastases. Subgroup affi liation of both primary tumour and recurrence is shown to the left and right, respectively, for each sample. Relative gene expression is plotted on a blue-red gradient in which red indicates high expression and blue low expression.

Articles

1204 www.thelancet.com/oncology Vol 14 November 2013

patient each with an SHH tumour and group 3 tumour and two patients with a group 4 tumour. Two patients in cohort 2, both with group 4 tumours, died of disease at 6·2 years and 7·4 years after the initial recurrence. Survival data were available for all 96 cases in cohort 3. Long-term survivors after recurrence were more common in this cohort, with 17 patients surviving more than 5 years after recurrence, of whom 15 received radiation at initial therapy. The subgroup distribution of these 17 cases was one WNT, four SHH, six group 3, and six group 4. Two recurrent WNT cases were also identifi ed in cohort 3; one was a long-term survivor of 5·4 years after recurrence (time to recurrence 3·2 years), whereas the other died 0·67 years after recurrence (time to recurrence 1·25 years). When treatment information was available, we did not identify any consistent diff erences in treatment regimens in these long-term survivors compared with the remainder of the cohort.

To establish whether the location of recurrence diff ers between subgroups, we reviewed the location of fi rst recurrence in cohort 1. The pattern of recurrence was available in all 30 recurrent cases. All patients were followed up with serial MRI of the craniospinal axis after completion of therapy as per treatment protocol. Recurrences were divided into local recurrences (tumour bed only without involvement of the cerebellar lepto-meninges and when CSF examinations were available, no malignant cells in the CSF), metastatic recurrences (relapse at distant sites outside the tumour bed), and metastatic plus local recurrences (relapse at both distant sites and the tumour bed). We identifi ed nine local recurrences and 21 metastatic recurrences. When we re-analysed these data in a subgroup-specifi c manner, eight of nine local recurrences were SHH tumours, and 17 of 20 metastatic recurrences were either group 3 or group 4, with one WNT metastatic recurrence (p=0·0014 Fisher’s exact test; local vs metastatic recurrence, SHH vs group 3 vs group 4; table 3, appendix p 15).

We also identifi ed this pattern in the two independent, non-overlapping validation cohorts, and the results were signifi cant (cohort 2 p=0·0013, cohort 3 p<0·0001, Fisher’s exact test; table 3, appendix p 15). We identifi ed no signifi cant diff erence in the pattern of relapse between group 3 and group 4 (appendix p 15).

In an exploratory analysis combining all three cohorts, incidence of local recurrences tended to be higher in group 4 in non-irradiated patients, suggesting that non-irradiated group 4 patients might recur locally (p=0·031; table 4, appendix p 16). When comparing the location of recurrence of patients given chemotherapy only versus craniospinal irradiation (with or without adjuvant chemotherapy), we noted no diff erence between SHH and group 3 in any of the three cohorts, and we noted no diff erence in these two treatment regimens when combining all three cohorts (SHH p=0·83, group 3 p=0·34; appendix p 16). CSF examinations in fi ve of nine local recurrences in cohort 1 and in all local recurrences

B Cohort 2 (Multicentre)

A Cohort 1 (Toronto)

C Cohort 3 (Burdenko)

Group 4Group 3

SHHNumber at risk

Ove

rall

surv

ival

(%)

9 7 4 2 2 1 1 1 1 1 19 2 1 1 0 0 0 0 0 0 0

10 2 1 0 0 0 0 0 0 0 0

0 1 2 3 4 5 6 7 8 9 10

p=0·013

0

25

50

75

100

36 24 14 12 9 7 3 2 1 1 137 20 13 9 8 4 2 2 1 1 021 10 8 6 5 4 3 2 1 1 0

0 1 2 3 4 5 6 7 8 9 10Time to death (years)

p=0·70

24 19 12 7 5 4 3 2 0 0 019 9 6 3 3 1 1 0 0 0 030 11 6 4 2 1 1 1 0 0 0

0 1 2 3 4 5 6 7 8 9 10

p=0·00750

25

50

75

100

0

25

50

75

100

Ove

rall

surv

ival

(%)

Ove

rall

surv

ival

(%)

Group 4Group 3

SHHNumber at risk

Group 4Group 3

SHHNumber at risk

SHHGroup 3Group 4

Figure 2: Subgroup-specifi c survival after recurrence across three non-overlapping cohorts of recurrent medulloblastomaKaplan-Meier survival estimates of overall survival after recurrence for (A) cohort 1, (B) cohort 2, and (C) cohort 3. p values were calculated using the log-rank test across the three subgroups.

Articles

www.thelancet.com/oncology Vol 14 November 2013 1205

in cohort 3 showed no metastatic dissemination. In cohort 2, CSF examinations were not available for all patients; however, in one local group 4 recurrence a CSF examination showed distant metastases (M1 disease). Two extraneural metastases occurred in cohort 2. The fi rst case was an SHH tumour in an adult patient who received radiation and chemotherapy and the second occurred in a child given radiation only. Three WNT tumours recurred across all three cohorts: two recurred with distant metastases only and one recurred in the tumour bed only (appendix p 8). No signifi cant diff erences were noted in either time to recurrence, or overall survival when comparing by location of recurrence in any of the three subgroups. SHH tumours recurred with metastases in 29% (18 of 62) of cases across all three cohorts. To work out whether age was a factor in disseminated relapse in SHH tumours we established the presence of metastatic recurrences by age across all three cohorts and noted no signifi cant diff erence (appendix p 17).

Metastatic dissemination at diagnosis was associated with 21% (11 of 52) of SHH tumours across all three cohorts, and was not statistically signifi cant when comparing local versus metastatic recurrences (p=0·25, appendix p 8). 39% (23 of 59 of group 3 patients and 58% (33 of 57) of group 4 patients with metastatic dissemination at relapse did not have distant metastases (ie, had M0 disease) at diagnosis (appendix p 8). We therefore conclude that SHH tumours more frequently have isolated tumour bed recurrences, and that group 3 and group 4 tumours usually recur with metastases.

DiscussionOur results show that medulloblastoma does not change subgroup at recurrence. This fi nding was not dependent on location of recurrence, because subgroup affi liation remained stable in both local tumour bed samples and metastatic samples at recurrence. We also identifi ed signifi cant diff erences across subgroups with respect to the anatomical and temporal patterns of recurrence, specifi cally SHH tumours recur mostly in the local tumour bed and group 3 and group 4 tumours recur almost exclusively with metastases. Patients with group 4 tumours also have an increased time to death after recurrence. These fi ndings have important impli-cations in the care of patients with recurrent medullo-blastoma and provide insight into the planning of future clinical trials. As far as we are aware, this study represents the largest cohort of recurrent medullo-blastoma assembled so far (appendix p 3) and the largest cohort of recurrent medullo blastoma with molecular correlation (panel).

Retention of subgroup provides further evidence supporting the notion that medulloblastoma arises from distinct cells of origin within the posterior fossa, the characteristics of which are carried forward from ontogeny into oncology. Indeed, this fi nding is in agreement with murine models in which WNT, SHH, and group 3

tumours have unique cells of origin, specifi cally the lower rhombic lip, the external granule layer, and postnatal cerebellar progenitor cells, respectively. Our fi ndings suggest that group 4 tumours also arise from a yet to be identifi ed cell of origin, distinct from the other three subgroups.21–23 Our fi ndings contrast with some initial fi ndings in glioblastoma that suggestions that subgroup affi liation can change at recurrence.19,20

As in breast cancer and renal cell carcinoma, metastases from medulloblastoma have been shown to be highly genetically divergent from their matched primary tumour.16,29,30 Genetic divergence of medulloblastoma metastases from their primary tumour in the face of subgroup stability across both tumour types suggests that subgroup identity might be established in the cell of origin. Further investigation of paired samples from diagnosis and matched metastatic recurrences using next generation methods, such as RNA sequencing and

Local Mixed Metastatic

Cohort 1 (p=0·0014)

SHH (n=11) 8 (73%) 1 (9%) 2 (18%)

Group 3 (n=9) 1 (11%) 1 (11%) 7 (78%)

Group 4 (n=9) 0 1 (11%) 8 (89%)

Cohort 2 (p=0·0013)

SHH (n=30) 18 (60%) 2 (7%) 10 (33%)

Group 3 (n=22) 4 (18%) 5 (23%) 13 (59%)

Group 4 (n=25) 4 (16%) 19 (76%) 2 (8%)

Cohort 3 (p<0·0001)

SHH (n=21) 18 (86%) 2 (10%) 1 (5%)

Group 3 (n=37) 1 (3%) 10 (27%) 26 (70%)

Group 4 (n=36) 3 (8%) 11 (31%) 22 (61%)

p values from Fisher’s exact test; comparison of local versus mixed versus metastatic pattern of recurrence. One WNT recurrence in the discovery cohort and two WNT recurrences in validation cohort 2 are described in the Results.

Table 3: Anatomical patterns of recurrence across medulloblastoma subgroups

Local Mixed Metastatic

SHH (p=0·83)

Chemotherapy (n=20) 14 (70%) 1 (5%) 5 (25%)

CSI with or without chemotherapy (n=38) 26 (68%) 4 (11%) 8 (21%)

Group 3 (p=0·32)

Chemotherapy (n=20) 2 (10%) 7 (37%) 11 (58%)

CSI with or without chemotherapy (n=48) 4 (8%) 9 (19%) 35 (73%)

Group 4 (p=0·031)

Chemotherapy (n=8) 3 (38%) 3 (38%) 2 (25%)

CSI with or without chemotherapy (n=59) 4 (7%) 43 (73%) 12 (20%)

p values from Fisher’s exact test; comparison of local versus mixed versus metastatic pattern of recurrence. Two patients in validation cohort 1 received focal radiotherapy with chemotherapy, one SHH with a metastatic recurrence and one group 4 with a metastatic recurrence; both cases are included in the chemotherapy only category. CSI=craniospinal irradiation.

Table 4: Anatomical pattern of recurrence combining all three cohorts, stratifi ed by treatment with CSI, across medulloblastoma subgroups

Articles

1206 www.thelancet.com/oncology Vol 14 November 2013

whole genome sequencing, will be needed to establish the full spectrum of heterogeneity between primary and metastatic tumours.

Previous research has suggested that local recurrences are more common in younger children, and metastatic recurrences are more common in older children.31,32 Protocols for treatment of infants consisting of chemotherapy only have been associated with a higher propensity for local recurrence than those including irradiation. However, our data suggest a simple, proximate explanation: this outcome is probably driven by subgroup, because SHH tumours are more common in infants and group 4 tumours are more common in older children.7,12

Overall, we noted no association between treatment regimen and site of recurrence, with the exception of a possible association between chemotherapy-only approaches and local recurrences in group 4 tumours. This fi nding suggests that in younger children who are not irradiated, local group 4 recurrences should not be unexpected; however, this conclusion is limited by a small sample size of non-irradiated patients with group 4

tumours (n=8) and warrants further investigation in a prospective trial. Treatment was uniform across cohorts and subgroups, with infants younger than 3 years given chemotherapy-only approaches and children older than 3 years given radiation followed by cisplatin-based adjuvant chemotherapy. Taken together, our data suggest that subgroup affi liation, rather than treatment eff ect, seems to be the primary driver of location of recurrence, particularly in patients with group 3 or SHH tumours.

A limitation of our study is the scarcity of knowledge with respect to assessments of distant metastases at relapse and absence of detailed treatment information at relapse. Future prospective studies will need to rigorously show that the CSF is free of metastatic disease in local recurrences, specifi cally in the rare situations in which local recurrences occur in group 3 and group 4 tumours, to exclude metastatic dissemination. Prospective multicentre longitudinal studies of recurrent medullo blastoma in a subgroup-specifi c manner are needed to establish whether present or future salvage therapies confer any benefi t. Our fi nding that the pattern of recurrence is highly subgroup specifi c needs prospective validation in a multicentre cooperative study of homogeneously treated patients.

Previous cross-species genomic studies have shown that metastatic medulloblastoma is a bicompartmental disease in which the metastases are highly divergent from their matched primary tumour in both human beings and mice.16,28 This fi nding suggests that response and susceptibility to therapy are likely to be diff erent in the primary tumour and metastases. Our data show the excellent tumour control for group 3 and group 4 medulloblastoma in the posterior fossa with existing therapies. However, the worrying fi nding that most group 3 and group 4 recurrences are metastatic, in combination with the fact that metastases are genetically divergent from the primary tumour,16 and that medullo-blastoma metastases are understudied, suggests that future basic science and clinical trials of group 3 and group 4 tumours should be more highly focused on metastases. Most patients with group 3 and group 4 tumours of average risk in our cohort were given high-dose craniospinal irradiation, further reinforcing the importance of generating novel approaches to therapy in metastatic disease. Moreover, our fi nding that most group 3 and group 4 metastatic relapses do not relapse in the primary site suggests that microscopic lepto-meningeal metastases not visible by neuroimaging or CSF examination are resistant to existing therapy. As such, additional local therapies targeting the primary site in the posterior fossa are unlikely to increase the proportion of patients cured for patients with group 3 or group 4 medulloblastoma. Possible strategies aimed at the metastases include intrathecal consolidation regimens in addition to existing therapies, which achieve excellent local tumour bed control.

Currently clinical trials for previously irradiated relapsed metastatic medulloblastoma are very heterogeneous and

Panel: Research in context

Systematic reviewWe searched PubMed and Google Scholar with the terms “medulloblastoma”, “recurrent medulloblastoma”, “medulloblastoma recurrences”, and “medulloblastoma treatment”, for reports in English, without any date restrictions. We identifi ed several studies over the past 30 years reporting the anatomical pattern of recurrence in medulloblastoma, specifi cally that recurrences can occur either in the tumour bed, along the leptomeninges, or both; however, all these previous studies are limited by both small numbers and absence of biological correlation. Several integrated genomic studies have identifi ed four subgroups of medulloblastoma with distinct demographics, genetics, transcriptomes, and outcomes; however, little is known with regards to the clinical implications of these subgroups at the time of disease recurrence.

InterpretationAs far as we are aware, our study is the largest single study of recurrent medulloblastoma and the fi rst to assess recurrent medulloblastoma in a subgroup-specifi c manner. Many researchers in paediatric neuro-oncology have suggested that medulloblastomas might recur as more aggressive subgroups (ie, an SHH or group 4 tumour becoming a group 3 tumour). In this study we showed that at the time of recurrence, medulloblastomas maintain their subgroup affi liation and that the anatomical pattern of recurrence is highly subgroup specifi c. SHH medulloblastomas almost always recur locally, suggesting that additional therapies aimed at the posterior fossa might be most benefi cial. Importantly, we showed that almost all recurrences of group 3 and group 4 medulloblastoma are metastatic, with recurrence in the posterior fossa in radiated patients being very rare. Moreover we showed that this subgroup-specifi c pattern of relapse was highly consistent across three independent cohorts, and was not dependent on therapy at diagnosis. As such, subgroup affi liation, rather than treatment eff ect, seems to be the main driver of location of recurrence. Our data for metastatic recurrence, in combination with previous reports showing that metastases are clinically and genetically distinct from the primary tumour,16,28 suggest that the paediatric neuro-oncology community needs to radically shift the focus away from the primary tumour towards metastases because patients with group 3 or group 4 tumours are dying almost exclusively from metastatic disease .

Articles

www.thelancet.com/oncology Vol 14 November 2013 1207

many are at phase 2 stage. Since any future subgroup-specifi c clinical trial will probably begin with relapsed patients, trials for relapsed group 3 or group 4 medulloblastoma are poised to fail if they are based on the biology of the primary tumour. Because SHH tumours predominantly recur in the posterior fossa, consideration of a clinical trial intensifying treatment to the posterior fossa (to treat local recurrence) while simultaneously reducing craniospinal doses (because leptomeningeal failure is rare) should be considered. Trials that are being planned will help to elucidate the eff ects of inhibitors of SMO on local tumour control for patients with SHH tumours.

ContributorsVR, MR, EB, SMP, AK, and MDT designed the study. MDT procured

fi nancial support. VR, MR, BL, CCF, SP, Y-JC, US, SG, RM, DB, MF,

KLL, SLP, SD, JT, NJ, AF, DTWJ, MK, MAK, SLG, DZ, SN, JP, JM, EL,

AWW, MR, OZ, EK, JA, SEC, JTR, CH, UT, K-ETC, RJP, and AK collected

data and provided study materials. VR, MR, EB, DJS, AMD, PAN, SMP,

AK, and MDT analysed and interpreted the data. VR, MR, EB, SMP, and

MDT wrote the report. All authors approved the fi nal report.

Confl icts of interestWe declare that we have no confl icts of interest.

AcknowledgmentsMDT is supported by a CIHR Clinician Scientist Phase II award, funds

from the Garron Family Chair in Childhood Cancer Research at The

Hospital for Sick Children and The University of Toronto, and operating

funds from the Canadian Institutes of Health Research, the National

Institutes of Health (R01CA159859 and R01CA148699) and the Pediatric

Brain Tumor Foundation. VR is supported by a CIHR fellowship and an

Alberta Innovates-Health Solutions Clinical Fellowship. SG, RM, and

DB are supported by the Pediatric Brain Tumor Foundation. MK, SG,

DZ are supported in part by grant UL1TR000038 from the National

Center for Research Resources, National Institutes of Health and grant

5P30CA016087-32 from the National Cancer Institute. MR is supported

by a fellowship from the Mildred Scheel Cancer Foundation. SMP is

supported by a grant from the Deutsche Kinderkrebsstiftung. We thank

Susan Archer for technical writing.

References1 Ramaswamy V, Northcott PA, Ta ylor MD. FISH and chips: the

recipe for improved prognostication and outcomes for children with medulloblastoma. Cancer Genet 2011; 204: 577–88.

2 Northcott PA, Jones DT, Kool M, et al. Medulloblastomics: the end of the beginning. Nat Rev Cancer 2012; 12: 818–34.

3 Packer RJ, Gajjar A, Vezina G , et al. Phase III study of craniospinal radiation therapy followed by adjuvant chemotherapy for newly diagnosed average-risk medulloblastoma. J Clin Oncol 2006; 24: 4202–08.

4 Gajjar A, Chintagumpala M, As hley D, et al. Risk-adapted craniospinal radiotherapy followed by high-dose chemotherapy and stem-cell rescue in children with newly diagnosed medulloblastoma (St Jude Medulloblastoma-96): long-term results from a prospective, multicentre trial. Lancet Oncol 2006; 7: 813–20.

5 Tarbell NJ, Friedman H, Polki nghorn WR, et al. High-risk medulloblastoma: a pediatric oncology group randomized trial of chemotherapy before or after radiation therapy (POG 9031). J Clin Oncol 2013; 31: 2936–41.

6 Pizer B, Donachie PH, Robinso n K, et al. Treatment of recurrent central nervous system primitive neuroectodermal tumours in children and adolescents: results of a Children’s Cancer and Leukaemia Group study. Eur J Cancer 2011; 47: 1389–97.

7 Taylor MD, Northcott PA, Kors hunov A, et al. Molecular subgroups of medulloblastoma: the current consensus. Acta Neuropathol 2012; 123: 465–72.

8 Dubuc AM, Morrissy AS, Kloost erhof NK, et al. Subgroup-specifi c alternative splicing in medulloblastoma. Acta Neuropathol 2012; 123: 485–99.

9 Remke M, Hielscher T, Korshun ov A, et al. FSTL5 Is a marker of poor prognosis in non-WNT/non-SHH medulloblastoma. J Clin Oncol 2011; 29: 3852–61.

10 Remke M, Hielscher T, Northco tt PA, et al. Adult medulloblastoma comprises three major molecular variants. J Clin Oncol 2011; 29: 2717–23.

11 Jones DT, Jager N, Kool M, et al. Dissecting the genomic complexity underlying medulloblastoma. Nature 2012; 488: 100–05.

12 Kool M, Korshunov A, Remke M, et al. Molecular subgroups of medulloblastoma: an international meta-analysis of transcriptome, genetic aberrations, and clinical data of WNT, SHH, group 3, and group 4 medulloblastomas. Acta Neuropathol 2012; 123: 473–84.

13 Northcott PA, Shih DJ, Peacoc k J, et al. Subgroup-specifi c structural variation across 1000 medulloblastoma genomes. Nature 2012; 488: 49–56.

14 Northcott PA, Shih DJ, Remke M, et al. Rapid, reliable, and reproducible molecular sub-grouping of clinical medulloblastoma samples. Acta Neuropathol 2012; 123: 615–26.

15 Pugh TJ, Weeraratne SD, Archer TC, et al. Medulloblastoma exome sequencing uncovers subtype-specifi c somatic mutations. Nature 2012; 488: 106–10.

16 Wu X, Northcott PA, Dubuc A, et al. Clonal selection drives genetic divergence of metastatic medulloblastoma. Nature 2012; 482: 529–33.

17 Northcott PA, Korshunov A, Witt H, et al. Medulloblastoma comprises four distinct molecular variants. J Clin Oncol 2011; 29: 1408–14.

18 Robinson G, Parker M, Kranenburg TA, et al. Novel mutations target distinct subgroups of medulloblastoma. Nature 2012; 488: 43–48.

19 Sottoriva A, Spiteri I, Piccirillo SG, et al. Intratumor heterogeneity in human glioblastoma refl ects cancer evolutionary dynamics. Proc Natl Acad Sci USA 2013; 110: 4009–14.

20 Phillips HS, Kharbanda S, Chen R, et al. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 2006; 9: 157–73.

21 Pei Y, Moore CE, Wang J, et al. An animal model of MYC-driven medulloblastoma. Cancer Cell 2012; 21: 155–67.

22 Kawauchi D, Robinson G, Uziel T, et al. A mouse model of the most aggressive subgroup of human medulloblastoma. Cancer Cell 2012; 21: 168–80.

23 Gibson P, Tong Y, Robinson G, et al. Subtypes of medulloblastoma have distinct developmental origins. Nature 2010; 468: 1095–99.

24 Reich M, Liefeld T, Gould J, Lerner J, Tamayo P, Mesirov JP. GenePattern 2.0. Nat Genet 2006; 38: 500–01.

25 Pfi ster S, Remke M, Benner A, et al. Outcome prediction in pediatric medulloblastoma based on DNA copy-number aberrations of chromosomes 6q and 17q and the MYC and MYCN loci. J Clin Oncol 2009; 27: 1627–36.

26 Korshunov A, Remke M, Werft W, et al. Adult and pediatric medulloblastomas are genetically distinct and require diff erent algorithms for molecular risk stratifi cation. J Clin Oncol 2010; 28: 3054–60.

27 Ellison DW, Kocak M, Dalton J, et al. Defi nition of disease-risk stratifi cation groups in childhood medulloblastoma using combined clinical, pathologic, and molecular variables. J Clin Oncol 2011; 29: 1400–07.

28 Gerlinger M, Rowan AJ, Horswell S, et al. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med 2012; 366: 883–92.

29 Aparicio S, Caldas C. The implications of clonal genome evolution for cancer medicine. N Engl J Med 2013; 368: 842–51.

30 Warmuth-Metz M, Blashofer S, Bueren AO, et al. Recurrence in childhood medulloblastoma. J Neurooncol 2010; 103: 705–11.

31 Tabori U, Sung L, Hukin J, et al. Distinctive clinical course and pattern of relapse in adolescents with medulloblastoma. Int J Radiat Oncol Biol Phys 2006; 64: 402–07.

32 Korshunov A, Benner A, Remke M, Lichter P, von Deimling A, Pfi ster S. Accumulation of genomic aberrations during clinical progression of medulloblastoma. Acta Neuropathol 2008; 116: 383–90.