EpiGe-App: An algorithm that facilitates the diagnosis of medulloblastoma

Medulloblastoma is a malignant brain tumour that mainly affects children and young adults, accounting for approximately 20% of all brain tumours in this population. Accurate classification of the molecular groups of medulloblastoma is crucial for oncologists to define the most appropriate treatment plan for each patient.

Molecular classification of medulloblastoma is increasingly important for clinical decision-making that defines patient management and treatment. Current classification systems for the molecular subgroups of medulloblastoma are expensive and take several weeks to obtain results. This makes them inaccessible for many centres around the world treating patients with brain tumours.

In this context, the team formed by researchers from CREB's B2SLab of the UPC, the Institut de Recerca Sant Joan de Déu (IRSJD) and l'Hospital Sant Joan de Déu, has developed a new methodology to classify the main molecular groups of medulloblastomas quickly, accurately and easily, using technology within the reach of most centres that treat children with brain tumours.

Initially, in 2018, a team from the Laboratory of Molecular Oncology of the Hospital Sant Joan de Déu and coordinator of the Translational Genomics group of the IRSJD, analysed genomic data from more than 1,500 medulloblastoma samples obtained thanks to the collaboration of international research groups for two years. This study identified the minimum number of markers that allowed the classification of medulloblastoma into clinically relevant molecular groups with a reliability of 96%. From these data, CREB researchers were able to develop a system that allowed the classification of medulloblastoma into the three subgroups of clinical interest (WNT, SHH and non-WNT/non-SHH) in an accurate, rapid and accessible manner.

With these results and to facilitate their application, the IRSJD and CREB's B2SLab group have created an interactive web application, EpiGe-App, which allows automated analysis and interpretation of the type of medulloblastoma. The advances achieved in this project have been rapidly transferred to care to support clinical decisions, to individualise patient therapy according to the genetic characteristics of the tumour and to offer more effective targeted therapies with less toxicity.

The EpiGe application allows the classification of medulloblastoma samples into the three molecular groups based on DNA methylation data obtained through the molecular technique of quantitative PCR (qPCR). Given the availability and accessibility of qPCR in most molecular biology laboratories, the tool developed makes the classification of these tumours an accessible test for most centres treating patients with central nervous system tumours.

To train the learning algorithm, genomic data from 4,800 samples, including 3,044 primary medulloblastomas and 1,644 non-medulloblastoma samples, were used. With these samples, a methylation status predictor algorithm was generated from raw qPCR data and from this prediction, an automatic classifier was generated into molecular subgroups of medulloblastoma. Through this learning, EpiGe-App is able to accurately classify WNT, SHH and non-WNT/non-SHH molecular subgroups of medulloblastoma, which can help medical teams define the most appropriate treatment.

EpiGe-App allows healthcare professionals around the world to upload the results obtained following the detailed protocol available on the website. In less than two minutes, the platform returns a report with the result, offering 96% reliability in identifying the medulloblastoma subgroup.

EpiGe-App has been funded with the support of the patient family associations of the Hospital Sant Joan de Déu, the TV3 Marathon and the Ministry of Science, Innovation and Universities with a budget of €298,625. The project has lasted three years.

    → Access to the EpiGe-App platform