A low-cost methodology for classifying the various types of medulloblastoma, the most common malignant central nervous system tumor in children, has been developed by a group of researchers based in São Paulo State, Brazil, in collaboration with colleagues affiliated with institutions in Switzerland and Germany.
The accuracy of the new method is similar to that of the most advanced sequencing technologies, which are far more expensive. Moreover, it can provide input for decision making on the best treatment strategy even in countries with scarce resources.
The results of the research, which was supported by FAPESP, are published in the journal Acta Neuropathologica Communications.
The researchers analyzed tumor tissue samples from 92 patients aged between 1 and 24 undergoing treatment for medulloblastoma at two general and teaching hospitals of the University of São Paulo – one attached to Ribeirão Preto Medical School (FMRP-USP), the other in São Paulo City – and at Centro Infantil Boldrini, a pediatric hospital in Campinas.
The method of analysis they used was real-time quantitative polymerase chain reaction (qPCR), which requires equipment that costs US$30,000 on average and is common in genetics laboratories but affordable to few Brazilian hospitals, according to Gustavo Alencastro Veiga Cruzeiro, who led the study as part of his PhD research at FMRP-USP with a scholarship from FAPESP.
The scientists first verified the expression of 20 genes associated with medulloblastoma, two fewer than are typically analyzed using more expensive technologies such as NanoString nCounter. The average cost of this analysis was US$60 per sample, similar to that of the high-precision technologies available to classify tumors into one of the four medulloblastoma molecular subgroups.
However, the researchers went further, using qPCR to find that a set of only six differentially expressed genes was sufficient to distinguish between subgroups. This discovery lowered the cost to US$26 per sample.
They confirmed the results by applying a computer algorithm to a gene expression dataset for 763 medulloblastoma samples already classified in previous international research.
Finally, 11 samples chosen at random from the 92 collected in Brazil were sent to the University Children’s Hospital in Zurich, Switzerland, and the German Cancer Research Center (DKFZ) in Heidelberg for analysis with more expensive equipment in routine use. The sample donors authorized the analysis.
“The equipment used for classification in developed countries costs about US$280,000 per unit in South America. The inputs for the analysis are also expensive. As a result, it’s very costly to identify the subgroup to which a tumor belongs and select the most suitable treatment,” Cruzeiro said.
Change of protocol
The research was part of the Thematic Project “Interactions between emerging therapeutic targets and developmental pathways associated with tumorigenesis: emphasis on pediatric malignancies”, for which the principal investigator is Luiz Gonzaga Tone, Full Professor of Pediatrics at FMRP-USP.
“The purpose of this wider project is to obtain new knowledge about the molecular mechanisms involved in the formation of some pediatric tumors and the possible interactions in their molecular development pathways, as a basis for enhanced classification criteria and treatment approaches. In the case of medulloblastoma, we found that the molecular classification criterion is fundamental,” said Tone, who heads the pediatric molecular oncology research group (GPOMP) at FMRP-USP.
The standard protocol for the treatment of medulloblastoma, which can affect different areas of the cerebellum, consists of surgical removal of the tumor, followed by chemotherapy and radiotherapy.
Recently, however, scientists have described four tumoral varieties that require therapies with differing degrees of aggressiveness. Two of these respond better to treatment.
Among patients with tumors classified in the WNT subgroup, the five-year survival rate can reach 90%, which is considered a very good prognosis. These patients may receive a lower amount of radiation or even be exempted from radiotherapy, which often has long-term toxic side effects, such as developmental, cognitive, mobility and speech impairment.
The prognosis for the SHH subgroup is intermediate: some patients respond well to treatment, while others respond less well. Targeted therapy, involving specific inhibitors of a key protein, is typically prescribed for these patients. Research has shown, however, that some patients do not respond to this treatment because of cell population diversity in this type of tumor.
The other two varieties, Group 3 and Group 4, metastasize more often than the rest and therefore require the most aggressive treatment. The biology of these subgroups is poorly understood.
“The molecular approach used to verify the tumoral subgroup and choose more or less intensive therapy is common in Switzerland, Germany and Canada, among other countries, but isn’t used in Brazil,” said Cruzeiro, who is currently doing a postdoctoral research internship at Harvard Medical School’s Massachusetts General Hospital (MGH) in the United States with FAPESP’s support.
In Brazil, according to Cruzeiro, all patients with medulloblastoma basically undergo the same treatment protocol, involving resection, chemotherapy and radiation therapy, with exceptions being made sometimes for children under three.
Many patients with WNT tumors therefore receive the treatment prescribed for patients at risk for metastasis, although they may not need radiation, for example. Even after the tumor is removed, the treatment can affect the child’s quality of life forever.
Cruzeiro warned that qPCR is not always precise and may fail to classify a tumor in any group in 5%-10% of cases. These tumors constitute a minority that must be subjected to more expensive methods.
“Nevertheless, this low-cost method can satisfactorily classify most tumors and provide important information for clinical decision making in Latin America, Africa and India,” he said.
The article “A simplified approach using Taqman low-density array for medulloblastoma subgrouping” (doi: 10.1186/s40478-019-0681-y) by Gustavo Alencastro Veiga Cruzeiro, Karina Bezerra Salomão, Carlos Alberto Oliveira de Biagi Jr, Martin Baumgartner, Dominik Sturm, Régia Caroline Peixoto Lira, Taciani de Almeida Magalhães, Mirella Baroni Milan, Vanessa da Silva Silveira, Fabiano Pinto Saggioro, Ricardo Santos de Oliveira, Paulo Henrique dos Santos Klinger, Ana Luiza Seidinger, José Andrés Yunes, Rosane Gomes de Paula Queiroz, Sueli Mieko Oba-Shinjo, Carlos Alberto Scrideli, Suely Marie Kazue Nagahashi, Luiz Gonzaga Tone and Elvis Terci Valera can be read at actaneurocomms.biomedcentral.com/articles/10.1186/s40478-019-0681-y.