GliMR: Cross-Border Collaborations to Promote Advanced MRI Biomarkers for Glioma

Abstract

Purpose There is an annual incidence of 50,000 glioma cases in Europe. The optimal treatment strategy is highly personalised, depending on tumour type, grade, spatial localization, and the degree of tissue infiltration. In research settings, advanced magnetic resonance imaging (MRI) has shown great promise as a tool to inform personalised treatment decisions. However, the use of advanced MRI in clinical practice remains scarce due to the downstream effects of siloed glioma imaging research with limited representation of MRI specialists in established consortia; and the associated lack of available tools and expertise in clinical settings. These shortcomings delay the translation of scientific breakthroughs into novel treatment strategy. As a response we have developed the network “Glioma MR Imaging 2.0” (GliMR) which we present in this article. Methods GliMR aims to build a pan-European and multidisciplinary network of experts and accelerate the use of advanced MRI in glioma beyond the current “state-of-the-art” in glioma imaging. The Action Glioma MR Imaging 2.0 (GliMR) was granted funding by the European Cooperation in Science and Technology (COST) in June 2019. Results GliMR’s first grant period ran from September 2019 to April 2020, during which several meetings were held and projects were initiated, such as reviewing the current knowledge on advanced MRI; developing a General Data Protection Regulation (GDPR) compliant consent form; and setting up the website. Conclusion The Action overcomes the pre-existing limitations of glioma research and is funded until September 2023. New members will be accepted during its entire duration.

Imaging Biomarkers of Glioblastoma Treatment Response: A Systematic Review and Meta-Analysis of Recent Machine Learning Studies

“…Second, another promising approach is to use advanced imaging techniques (67). The last three decades have seen considerable technical developments in MRI (for example, those related to perfusion, permeability and diffusion), 1H-MR spectroscopic imaging, and positron emission tomography (for example using radiolabelled amino acids).…”

A systematic review on the use of quantitative imaging to detect cancer therapy adverse effects in normal-appearing brain tissue

“…osipi. org) and Glioma MR Imaging 2.0 (GliMR, COST Action CA18206) [20], aim at making knowledge and tools for acquisition and analysis of quantitative MRI widely available to advance the field. Additionally, pooling datasets might be indispensable to investigate subtle therapy effects and their regional variations, and to disentangle the interaction of different treatments, possibly with the help of machine learning methods.…”

“…While specific study design, like head-and-neck tumors or proton vs photon radiochemotherapy, can help to separate the tumor from radiation effects or radio- from chemotherapy, respectively, more evidence from pooled large studies will be needed. Several community-driven initiatives, for example, Open Source Initiative for Perfusion Imaging (OSIPI— www.osipi.org ) and Glioma MR Imaging 2.0 (GliMR, COST Action CA18206) [ 20 ], aim at making knowledge and tools for acquisition and analysis of quantitative MRI widely available to advance the field. Additionally, pooling datasets might be indispensable to investigate subtle therapy effects and their regional variations, and to disentangle the interaction of different treatments, possibly with the help of machine learning methods.…”

“…To summarize the recommendations, we advise that future studies investigating the normal tissue changes following cancer therapy should: Provide a complete description of the study population, treatment details and timeline, as well as details of acquisition and processing—the modified-QUADAS-2 criteria in supplementary materials provide guidance for this; adhere to the latest recommendations for image acquisition and processing—ASL [ 127 , 128 ], DSC [ 129 , 134 , 135 ], MRS [ 98 ]; avoid mixed cohorts with inhomogeneous tumor and treatment types; correct for multiple comparisons and perform a power analysis to avoid spurious findings and to help interpret negative findings; support community efforts for data sharing and collaboration (e.g. OSIPI, BIDS [ 131 ], OpenNeuro [ 136 ], GliMR [ 20 ], or ENBIT – www.enbit.ac.uk ) to enable future data pooling in larger studies. …”

“…The common reasons for this are a lack of insight into the biological specificity of these advanced sequences, the limited available scanning time beyond the standard protocol, and the fact that the complexity of some of these protocols makes them unfit for some of the patients. Insufficient clinical implementation [ 19 ], due to difficult acquisition and analysis, often leads to low levels of evidence in this context [ 20 ]. To trigger change, more attention needs to be devoted to these techniques, starting by reviewing the current body of literature.…”