Tumour heterogeneity in glioblastoma assessed by MRI texture analysis: a potential marker of survival

Abstract

Objective: The main objective of this retrospective work was the study of three-dimensional (3D) heterogeneity measures of post-contrast pre-operative MR images acquired with T 1 weighted sequences of patients with glioblastoma (GBM) as predictors of clinical outcome. Methods: 79 patients from 3 hospitals were included in the study. 16 3D textural heterogeneity measures were computed including run-length matrix (RLM) features (regional heterogeneity) and co-occurrence matrix (CM) features (local heterogeneity). The significance of the results was studied using Kaplan-Meier curves and Cox proportional hazards analysis. Correlation between the variables of the study was assessed using the Spearman's correlation coefficient.Results: Kaplan-Meyer survival analysis showed that 4 of the 11 RLM features and 4 of the 5 CM features considered were robust predictors of survival. The median survival differences in the most significant cases were of over 6 months. Conclusion: Heterogeneity measures computed on the post-contrast pre-operative T 1 weighted MR images of patients with GBM are predictors of survival. Advances in knowledge: Texture analysis to assess tumour heterogeneity has been widely studied. However, most works develop a two-dimensional analysis, focusing only on one MRI slice to state tumour heterogeneity. The study of fully 3D heterogeneity textural features as predictors of clinical outcome is more robust and is not dependent on the selected slice of the tumour. INTRODUCTIONGlioblastoma (GBM) is the most frequent malignant brain tumour in adults and the most lethal type, with a median survival of 14.6 months for patients receiving the standard of care, i.e. maximal safe surgery plus radiotherapy and chemotherapy.1 Pre-operative MRI is routinely used for diagnosis, treatment planning, response evaluation and follow-up.The typical GBM appearance upon diagnosis on MRI consists of an enhanced ring with a central non-enhanced core of necrosis observed mainly on contrast-enhanced T 1 weighted images. Recently, there has been an increased use of advanced imaging techniques alone or in combination with conventional MRI modalities to characterize the connection of the so-called radio phenotype with the tumour genotype, the so-called radiogenomics (e.g. recent reviews for GBM 2,3 ). However, the use of those techniques requires further research and validation to achieve a broad clinical applicability. Thus, in clinical practice and trials, T 1 contrast-enhanced and T 2 /fluid-attenuated inversionrecovery images are still the gold standard for diagnosis and treatment planning. 4 One of the most important characteristics of GBM is its marked intratumour heterogeneity. In fact, these tumours are usually referred to as "glioblastoma multiforme", in order to highlight their heterogeneous nature. GBM tumours are formed by tumour cells which differ in their morphology, genetics and biological behaviour and may underlie the inability of conventional therapies to significantly impact patient outcomes.5 It is well known that ...

Applied mathematics and nonlinear sciences in the war on cancer

“…Mathematical "dynamical" models have already been the basis of many theoretical proposals including: tumor control [13][14][15], adaptive therapies [16], metronomic [17,18] or protracted [19] therapies, implementing concepts from evolutionary dynamics [20][21][22], non-Darwinian dynamics [23], therapy personalization [24][25][26][27][28], to cite a few of very many examples. In addition, the definition of quantitative measures of the disease such as novel imaging biomarkers may benefit from the use of mathematical ideas [27,[29][30][31].…”

Optimizing Texture Retrieving Model for Multimodal MR Image‐Based Support Vector Machine for Classifying Glioma

“…Generally, a set of first‐ and second‐order texture attributes based on varied analyzing strategies were applied to reflect specific image heterogeneity. For example, Global texture attributes, as the first‐order image texture attributes, generally refer to those statistical histogram indictors and reflect the gray‐level distribution within the whole tumor region . Second‐order texture attributes were usually based on different analysis matrices including the gray‐level co‐occurrence matrix (GLCM), gray‐level run‐length matrix (GLRLM), and gray‐level size‐zone matrix (GLSZM), which reflect different spatial gray‐level distribution from different aspects.…”

Textural features of dynamic contrast‐enhanced MRI derived model‐free and model‐based parameter maps in glioma grading

“…Among these radiomic features, textural features are one of the most commonly used, which could quantify spatial variation of gray‐level intensity and characterize the underlying heterogeneity of the image under evaluation . Until recently, textural features, extracted from routine MR images, have successfully been associated with glioma grade, overall survival, and isocitrate dehydrogenase (IDH) genotype . However, dynamic contrast‐enhanced MRI (DCE‐MRI), which makes the heterogeneity of vascular physiology and structure quantifiable data and reveals dramatic imaging heterogeneity, was seldom analyzed by texture analysis.…”

“…16 Until recently, textural features, extracted from routine MR images, have successfully been associated with glioma grade, overall survival, and isocitrate dehydrogenase (IDH) genotype. [17][18][19][20] However, dynamic contrast-enhanced MRI (DCE-MRI), which makes the heterogeneity of vascular physiology and structure quantifiable data and reveals dramatic imaging heterogeneity, was seldom analyzed by texture analysis. However, heterogeneity of microvascular proliferation has been recognized as one of the most important signatures of glioma.…”