Abstract
We propose a statistical multiscale mapping approach to identify microscopic and molecular heterogeneity across a tumor microenvironment using multiparametric MR (mp-MR). Twenty-nine patients underwent pre-surgical mp-MR followed by MR-guided stereotactic core biopsy. The locations of the biopsy cores were identified in the pre-surgical images using stereotactic bitmaps acquired during surgery. Feature matrices mapped the multiparametric voxel values in the vicinity of the biopsy cores to the pathologic outcome variables for each patient and logistic regression tested the individual and collective predictive power of the MR contrasts. A non-parametric weighted k-nearest neighbor classifier evaluated the feature matrices in a leave-one-out cross validation design across patients. Resulting class membership probabilities were converted to chi-square statistics to develop full-brain parametric maps, implementing Gaussian random field theory to estimate inter-voxel dependencies. Corrections for family-wise error rates were performed using Benjamini-Hochberg and random field theory, and the resulting accuracies were compared. The combination of all five image contrasts correlated with outcome (P < 10−4) for all four microscopic variables. The probabilistic mapping method using Benjamini-Hochberg generated statistically significant results (α ≤ 0.05) for three of the four dependent variables: (1) IDH1, (2) MGMT, and (3) microvascular proliferation, with an average classification accuracy of 0.984 ± 0.02 and an average classification sensitivity of 1.567% ± 0.967. The images corrected by random field theory demonstrated improved classification accuracy (0.989 ± 0.008) and classification sensitivity (5.967% ± 2.857) compared with Benjamini-Hochberg. Microscopic and molecular tumor properties can be assessed with statistical confidence across the brain from minimally-invasive, mp-MR.
Highlights
Emerging targeted therapies interfere with specific molecules that promote tumor growth and infiltration based on patient-specific predictive cellular and molecular biomarkers[1]
We evaluate our hypothesis in three separate sub-steps: Sub-hypothesis 1) significant relationships (α ≤ 0.05) between macroand micro-scale properties can be identified using elementary statistical testing when surgical pathology results are localized to the pre-surgical image space; Sub-hypothesis 2) non-parametric machine learning can classify microscopic properties from macroscopic images with high accuracy (≥ 95%) when traditional corrections for family-wise error rates are employed; and Sub-hypothesis 3) clinically-useful multiscale classification across the entire image space can be accomplished when the parametric images are treated as Gaussian random fields
We evaluated the use of non-parametric machine learning to predict four clinically relevant properties: IDH1 mutation status, MGMT promoter methylation, cellular necrosis, and microvascular proliferation
Summary
Emerging targeted therapies interfere with specific molecules that promote tumor growth and infiltration based on patient-specific predictive cellular and molecular biomarkers[1]. Significant efforts are underway to develop tumor heterogeneity mapping techniques using minimally-invasive imaging including texture analysis[8,9,10] proton[11,12] and hyperpolarized 13C13 spectroscopy; and most recently MR fingerprinting[14] These methods classify tumor properties at one of two levels: (1) volumetrically, by segmenting adjacent voxels together into classes, or (2) on a voxel-wise basis, treating each voxel independently. Volumetric segmentation techniques leverage spatial correlations in adjacent voxels that may be associated with tumor biology and/or the physical attributes of the acquisition process to improve SNR and classification accuracy These improvements are balanced by a decrease in the theoretical spatial resolution of the parametric images, limiting the assessment of heterogeneity. We evaluated the images qualitatively by clinical experts and quantitatively by classification accuracy in the biopsy sample volume
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.