Radiologic Image-Based Statistical Shape Analysis of Brain Tumours

Abstract

Summary. We propose a curve-based Riemannian geometric approach for general shape-based statistical analyses of tumours obtained from radiologic images. A key component of the framework is a suitable metric that enables comparisons of tumour shapes, provides tools for computing descriptive statistics and implementing principal component analysis on the space of tumour shapes and allows for a rich class of continuous deformations of a tumour shape. The utility of the framework is illustrated through specific statistical tasks on a data set of radiologic images of patients diagnosed with glioblastoma multiforme, a malignant brain tumour with poor prognosis. In particular, our analysis discovers two patient clusters with very different survival, subtype and genomic characteristics. Furthermore, it is demonstrated that adding tumour shape information to survival models containing clinical and genomic variables results in a significant increase in predictive power.

Shape Features of the Lesion Habitat to Differentiate Brain Tumor Progression from Pseudoprogression on Routine Multiparametric MRI: A Multisite Study

“…We presented the first approach at assessing the effectiveness of 3D shape and surface radiomic features extracted from the tumor habitat (enhancing lesion, and T 2w /FLAIR hyperintense peri-lesional regions) to differentiate tumor progression from PsP on conventional MR scans (Gd-c T 1w , T 2w , FLAIR). Our work was based on the rationale that there are observable 3-dimenstional shape and surface irregularities encountered on the enhancing lesion as well as the perilesional boundaries in patients with tumor recurrence (potentially due to aggressive tumor infiltration and disruption) as compared to those with benign pseudoprogression, which will likely have more regular boundaries [16, 19, 20].…”

Intracranial glioma xenograft model rapidly reestablishes blood–brain barrier integrity for longitudinal imaging of tumor progression using fluorescence molecular tomography and contrast agents

“…Application of a threshold based on naïve controls or shape analysis of tumor ROI fluorescence may help to further distinguish tumor from background. 34 We also demonstrated that FUS-assisted delivery of LS301 can significantly enhance tumor contrast at time points in tumor progression when BBB is intact, preventing access of intravenously administered molecular probes to brain tumors. This finding illustrates how a combination of tumor-targeting molecules and transient BBB disruption can additively augment both tumor uptake and retention in the brain.…”

Analysis of shape data: From landmarks to elastic curves

“…One can reexpress the data using coordinates of this subspace via the principal coefficients in ℝ r computed as . One can then use these principal coefficients for further modeling, for example, PC regression (Bharath et al, ).…”

“…(a) Laga et al [2014] (b) Srivastava et al [2009] (c) O'Higgins and Dryden [1992] (d) Bharath et al [2018] (e) Kurtek et al [2012] (g) Kaziska and Srivastava [2006] (f) Joshi and Srivastava [2009] F I G U R E 1 Applications of statistical shape analysis include (a) leaf shape classification, (b) facial recognition, (c) anthropology, (d) tumor shape modeling, (e) clustering of diffusion tensor magnetic resonance imaging fibers, (f) military defense, and (g) gait recognition…”

“…An important goal in medical imaging is to assess the morphology of anatomical structures for the purposes of disease detection or monitoring. While most research focuses on 3D shape analysis in this setting (Kurtek et al, ; Samir, Kurtek, Srivastava, & Canis, ), curve‐based approaches have been used to model shapes of diffusion tensor magnetic resonance imaging fiber tracts (Kurtek, Srivastava, Klassen, & Ding, ), to assess variability in manual segmentation of medical images (Kurtek et al, ), and to model survival based on glioblastoma multiforme tumor shapes (Bharath, Kurtek, Rao, & Baladandayuthapani, ). Other popular applications of statistical shape analysis include biometrics (Kaziska & Srivastava, ; Samir, Srivastava, Daoudi, & Klassen, ; Srivastava, Samir, Joshi, & Daoudi, ), military (Joshi & Srivastava, ), activity recognition and modeling (Su, Kurtek, Klassen, & Srivastava, ), and anthropology (O'Higgins & Dryden, ), among others.…”