by Edward C Schwalbe, Janet C Lindsey, Sirintra Nakjang, Stephen Crosier, Amanda J Smith, Debbie Hicks, Gholamreza Rafiee, Rebecca M Hill, Alice Iliasova, Thomas Stone, Barry Pizer, Antony Michalski, Abhijit Joshi, Stephen B Wharton, Thomas S Jacques, Simon Bailey, Daniel Williamson, Steven C Clifford.
Originally published in The Lancet Oncology, May 22nd 2017
Medulloblastoma, the most common CNS tumor of childhood, is the leading cause of cancer deaths in young people. What was once thought of as a single disease is, since an International consensus meeting in 2012, now recognized as four distinct molecular entities (WNT, SHH, Grp3, Grp4), each defined by differing disease features and outcomes. However, the studies that defined this consensus were based on small cohorts of tumors, and were performed using low-resolution approaches.
It has long been recognized that there is considerable variability in disease outcomes within these consensus molecular subgroups. In this study, we looked at molecular marks known as DNA methylation (small chemical tags on our DNA sequence that can influence whether a gene is switched on), in much larger cohorts (>700 patient samples), at a higher resolution across the whole genome, to investigate whether we could identify additional molecular subgroups of medulloblastoma that were better able to explain the differences that we see in disease outcomes.
We found that the consensus WNT subgroup remained as a distinct group, whereas the non-WNT subgroups each split into two. We were able to validate these findings in an external dataset. SHH tumors split into age-dependent groups, with an age cutoff of approximately 4.3 years. This might mean that children with SHH tumors aged between 3 and 4.3 years of age at diagnosis could be better treated on protocols that aim to delay radiotherapy, as is currently the case for patients under 3 years of age. For the older SHH group (MBSHH-Child), we found that the treatment classification currently being implemented in the European PNET5 clinical trial for non-infant patients could classify patients into favorable (96% 5-year survival) and very high (29% 5-year survival) risk groups.
Grp3 and Grp4 medulloblastoma subgroups each split into two, defined by the presence or absence of a shared biological signature. This signature identified low-risk variants of Grp3 and Grp4, with a much improved survival relative to their high-risk counterparts.
We identified new stratification schemes for Grp3 and Grp4 medulloblastomas, based upon these new low-risk subgroups, and other, previously identified disease features, to define favorable, standard, high and very high risk variants of Grp3 and Grp4, each with differential outcomes.
Putting all this together, we were able to reclassify 25% of childhood medulloblastoma patients as favorable risk (>90% survival at 5 years after diagnosis), compared to 10% of patients using current classification schemes. Of course, this work needs to be validated in a clinical trial, but in future, up to a quarter of patients classified as favorable risk by this scheme could be eligible for a reduction in the intensity of their therapies, ensuring that they remain cured of their disease, but with an accompanying improvement in their quality of life after treatment due to less damaging treatments.
At the same time, for the high (42% 5-year survival) and very-high risk patient groups (28% 5-year survival), it is clear that current therapies for these patient groups are inadequate, and that alternative upfront treatment approaches need to be considered for these patients, as well as optimization of current therapies, to maximize their chances of survival. This work therefore has important implications for future clinical management of the disease.
Dr. Ed Schwalbe
Ed studied Genetics at Leicester University in the UK, before undertaking a Master’s degree in Bioinformatics, also at Leicester, which focused on the application of computers to investigate complex biological questions. Subsequently, he carried out his PhD at Newcastle University with Professor Steve Clifford, on the molecular sub-classification of the pediatric brain tumor, medulloblastoma.
In his current role as a Senior Lecturer in Bioinformatics and Biostatistics at Northumbria University, Ed has continued working on the genome-wide analysis of medulloblastoma tumors in projects led by Professor Steve Clifford at Newcastle University. In their most recent study, which was published in May 2017 in The Lancet Oncology, Professor Clifford and Dr Schwalbe describe novel molecular subgroups of medulloblastoma with clear clinical application. Their major funders include Cancer Research UK, the Brain Tumor Charity and the INSTINCT project.