Advances in computational and translational approaches for malignant glioma

Abstract

Gliomas are the most common primary brain tumors in adults and carry a dismal prognosis for patients. Current standard-of-care for gliomas is comprised of maximal safe surgical resection following by a combination of chemotherapy and radiation therapy depending on the grade and type of tumor. Despite decades of research efforts directed towards identifying effective therapies, curative treatments have been largely elusive in the majority of cases. The development and refinement of novel methodologies over recent years that integrate computational techniques with translational paradigms have begun to shed light on features of glioma, previously difficult to study. These methodologies have enabled a number of point-of-care approaches that can provide real-time, patient-specific and tumor-specific diagnostics that may guide the selection and development of therapies including decision-making surrounding surgical resection. Novel methodologies have also demonstrated utility in characterizing glioma-brain network dynamics and in turn early investigations into glioma plasticity and influence on surgical planning at a systems level. Similarly, application of such techniques in the laboratory setting have enhanced the ability to accurately model glioma disease processes and interrogate mechanisms of resistance to therapy. In this review, we highlight representative trends in the integration of computational methodologies including artificial intelligence and modeling with translational approaches in the study and treatment of malignant gliomas both at the point-of-care and outside the operative theater in silico as well as in the laboratory setting.

Keywords: artificial intelligence; diagnostics; glioma; heterogeneity; modeling; personalized medcine; therapeutics.

FIGURE 1

Emerging methodologies for the study and treatment of glioma at the point-of-care. Rapid, intraoperative techniques that enable improved delineation of the tumor margin and patient-specific “fingerprinting” may facilitate the development of effective therapies. Intraoperative experimental paradigms such as real-time microdialysis has the potential to identify patient-specific biomarkers and improve our understanding of treatment response and recurrence. P, perfusate; D, dialysate.

FIGURE 2

Modeling the glioma-neural interface and its implications. Technological advances in computing have enabled the modeling of large-scale brain networks and methodology to study network disruption by glioma. This may help uncover and preserve higher order functions and has the potential to impact the current surgical paradigm. At the cellular and molecular level, understanding network disruption and glioma-neural integration may yield novel therapeutics targeting neural regulation of cancer and network remodeling.