Friismose, Ancuta Ioana

Abstract Glioblastoma progression is spatially heterogeneous, but conventional imaging provides limited information about where subsequent tumor progression is likely to occur. We developed a directional magnetic resonance elastography (MRE) framework to test whether local post-treatment tumor–brain interface mechanics are associated with later spatial tumor progression. In a secondary analysis of a prospectively acquired glioblastoma cohort, wedge-level viscoelastic instability features were extracted from the first post-treatment MRE scan and related to novel tumor burden on the second post-treatment scan after excluding tumor already present on pretreatment or first post-treatment imaging. Nine patients had longitudinal imaging suitable for spatial comparison; six lesions showed net interval growth and were included in the primary wedge-level directional analysis, while three non-growing lesions were retained for descriptive comparison. In growing lesions, several directional mechanical features were descriptively associated with later novel tumor burden. In cluster-aware models accounting for within-patient dependence among wedges, mean Δtanδ showed the most consistent association with later wedge-level novel tumor fraction across mixed-effects and generalized estimating equation analyses. Associations were directionally stable across wedge-width sensitivity analyses. These findings provide proof of principle that post-treatment glioblastoma interface mechanics contain spatially resolved information related to where later tumor emergence occurs, supporting further validation of directional MRE as a framework for longitudinal mapping of progression geometry.

Abstract Mechanical properties of biological tissues vary across spatial scales, yet radiomics typically relies on fixed, heuristic choices for neighbourhood size, kernel geometry, and spectral content - choices that can silently reshape the feature space before any modelling begins. We introduce a label-free, information-theoretic framework for selecting extraction parameters in multi-frequency MRE radiomics. For each configuration θ - neighbourhood radius r, kernel geometry k (sphere or shell), and frequency subset f - we extract a radiomics feature matrix and score it using an objective J(θ) that integrates distributional richness (Shannon entropy), cross-frequency coherence (canonical correlation), inter-feature redundancy (Spearman correlation), and bootstrap stability. We evaluate 121 configurations per tissue in multi-frequency MRE (30-60 Hz) of human brain, liver, and a calibrated phantom, and test robustness using 10,000 Dirichlet-sampled objective weightings. Across tissues, neighbourhood aggregation is consistently preferred over voxel-wise extraction, outperforming the no-neighbourhood baseline in 98.4-100% of weightings. External validation in 100 independent brain scans acquired with a different protocol and wider frequency range (20-90 Hz) confirms a reproducible mesoscopic plateau at r = 3-5 (9-15 mm), with a modal optimum at r = 4; omitting neighbourhood analysis reduces J(θ) by 38% relative to each subject’s optimum. Frequency-subset preferences replicate across datasets, with lower frequencies most frequently selected for brain. By turning ad hoc extraction choices into an outcome-free optimisation step, this framework improves reproducibility, reduces sensitivity to heuristic parameter choices, and generalises across acquisition protocols and imaging sites.

Abstract Background Conventional MRI cannot reliably distinguish radiation necrosis (RN) from recurrent metastasis after cranial radiotherapy, as both can show similar enhancement despite different biology. We tested whether these entities are mechanically non-equivalent in vivo and separable by MRE-derived viscoelastic metrics and perilesional interface-instability features. Methods In a prospective, histopathology-anchored cohort, 11 post-radiotherapy enhancing lesions were classified as RN (n=3) or recurrent/progressive tumor (n=8). MRE was acquired at 3.0 T with single-frequency 60-Hz excitation to derive storage modulus (G′), loss modulus (G″), and complex shear modulus magnitude (|G*|). Co-primary endpoints were median tumor G′ and |G*|, each tested one-sided (RN > tumor) with Holm correction across the two co-primary tests. Median tumor G″ was tested two-sided. A prespecified secondary 6-endpoint family (absolute and tumor/NAWM-normalized G′, G″, and |G*|) was analyzed with Benjamini–Hochberg FDR control. Exploratory instability mapping in a 0–6 mm peritumoral shell generated interface-topology metrics, including convexity. Results Absolute tumor-core medians were higher in RN than tumor for |G*| (1.79 vs 1.32 kPa; Cliff’s δ=0.67; q=0.10), G′ (1.62 vs 1.09 kPa; δ=0.50; q=0.14), and G″ (0.81 vs 0.46 kPa; δ=0.75; q=0.10). NAWM normalization improved separation: tumor/NAWM |G*| (2.26 vs 1.41; δ=0.92; q=0.04) and tumor/NAWM G″ (2.67 vs 0.87; δ=1.00; q=0.04) were FDR-significant. Convexity also differentiated RN from tumor (0.49 vs 0.36; δ=1.00; MWU p=0.01). Conclusions Tumor/NAWM G″, tumor/NAWM |G*|, convexity, and tumor G″ emerged as the strongest candidate features, indicating that RN is mechanically harder and more dissipative than recurrent metastasis. Signal strength was high (Cliff’s δ up to 1.00) but should be interpreted cautiously given sample size. Exploratory analyses further suggest that instability mapping captures biologically relevant interface behavior. These findings support a mechanics-based RN-versus-recurrence framework and justify prespecified, preregistered external validation.

ABSTRACT Background Magnetic resonance elastography (MRE) has demonstrated potential in characterizing intracranial tumors, including glioblastoma. The influence of pathology on measurement consistency and interobserver agreement needs evaluation to promote MRE implementation as a quantitative biomarker. Purpose To assess repeatability and interobserver agreement of absolute and normalized magnitude of the complex shear modulus (| G *|), storage modulus ( G ′), and loss modulus ( G ″) in glioblastoma. Study Type Prospective. Population Thirteen adults (5 male, 8 female, mean age 66.23 years) with histopathologically confirmed glioblastoma. Field Strength/Sequence 3 T MRI, modified single‐shot spin‐echo echo‐planar imaging. Assessment Two same‐session MRE acquisitions were performed with patient repositioning. The solid tumor component was independently segmented by 2 observers on contrast‐enhanced T1‐weighted images co‐registered to MRE maps. Statistical Tests Repeatability was assessed using the repeatability coefficient, coefficient of variation, and Bland–Altman plots, and interobserver agreement by the intraclass correlation coefficient. The Wilcoxon signed‐rank test compared parameters and coefficients of variation between tumor‐normal‐appearing white matter and absolute and normalized measurements. Significance was set at 0.05. Results Tumor | G *| and G ′ showed repeatability coefficients of 0.07–0.08 kPa and 0.05 kPa, compared with 0.13 kPa (| G *|) and 0.15–0.16 kPa ( G ′) for normalized measurements. Coefficients of variation in tumor regions were 1.42% ± 1.12%–1.60% ± 1.41% for | G *| and 1.19% ± 0.96%–2.08% ± 2.22% for G ′, significantly lower than normalized values (4.82% ± 4.49%–4.21% ± 4.27% for | G *|; 5.12% ± 5.04%–5.45% ± 4.53% for G ′). Tumor | G *| and G ′ showed excellent interobserver agreement (intraclass correlation coefficients 0.97 and 0.95). Tumor G ″ demonstrated higher variability than | G *| and G ′ (coefficients of variation 8.58% ± 7.69%–7.51% ± 6.73%), with no significant difference between absolute and normalized measurements ( p = 0.14). Data Conclusion Tumor | G *| and G ′ are the most repeatable metrics in glioblastoma. Normalization reduces measurement repeatability due to normal‐appearing white matter variability. The small sample size ( n = 13) limits generalizability. Evidence Level 2. Technical Efficacy Stage 2.

ABSTRACTBackgroundMagnetic resonance elastography (MRE) quantifies tissue viscoelasticity, offering insights into intracranial meningiomas. MRE‐derived parameters, including the storage modulus (G') and the magnitude of the complex modulus (|G*|), may aid neurosurgical planning, but their repeatability remains unexplored in brain tumors.Purpose/HypothesisTo evaluate the repeatability of G' and |G*| in meningiomas, the influence of tumor volume and spatial location on variability, and ranking stability across paired elastograms.Study TypeProspective.Population/SubjectsSeventeen paired MRE scans from 16 patients with meningiomas (mean age 64 ± 12 years, 10 females) scheduled for resection.Field Strength/SequenceMRE was performed on a 3 T system using a single‐shot spin‐echo echo‐planar imaging (SE‐EPI) sequence.AssessmentTumors were segmented on contrast‐enhanced T1‐weighted images; distance and volume metrics were extracted. Regions of interest were applied to G' and |G*| elastograms. Repeatability was evaluated using the coefficient of variation (CV%). Ranking stability was assessed by comparing median G' or |G*| values.Statistical TestsThe difference in CV% was assessed using the Wilcoxon signed rank test. Linear regression assessed the effect of tumor volume and distance from the external occipital protuberance (EOP) on CV%. The Wilcoxon signed‐rank test evaluated ranking stability and scan‐rescan CV%. Statistical significance was defined as p < 0.05.ResultsThe difference in CV% G' (14.89 ± 12.32) and CV% |G*| (13.88 ± 12.60) was not significant (p = 0.61). Tumor volume inversely correlated with CV% (G': β = −0.75 |G*|: β = −0.59). Tumor‐EOP distance had no significant effect (G': p = 0.58, |G*|: p = 0.23). Rankings remained stable (G': p = 0.82, |G*|: p = 0.71).Data ConclusionMRE‐derived parameters are repeatable in intracranial meningiomas, with G' and |G*| being equally stable. Larger tumors yielded more consistent estimates. Preserved ranking stability supports the use of MRE in preoperative planning.Evidence Level2.Technical EfficacyStage 2.