Search

Purpose To develop and validate a deep learning (DL) method to detect and segment enhancing and nonenhancing cellular tumor on pre- and posttreatment MRI scans in patients with glioblastoma and to predict overall survival (OS) and progression-free survival (PFS). Materials and Methods This retrospective study included 1397 MRI scans in 1297 patients with glioblastoma, including an internal set of 243 MRI scans (January 2010 to June 2022) for model training and cross-validation and four external test cohorts. Cellular tumor maps were segmented by two radiologists on the basis of imaging, clinical history, and pathologic findings. Multimodal MRI data with perfusion and multishell diffusion imaging were inputted into a nnU-Net DL model to segment cellular tumor. Segmentation performance (Dice score) and performance in distinguishing recurrent tumor from posttreatment changes (area under the receiver operating characteristic curve [AUC]) were quantified. Model performance in predicting OS and PFS was assessed using Cox multivariable analysis. Results A cohort of 178 patients (mean age, 56 years ± 13 [SD]; 116 male, 62 female) with 243 MRI timepoints, as well as four external datasets with 55, 70, 610, and 419 MRI timepoints, respectively, were evaluated. The median Dice score was 0.79 (IQR, 0.53-0.89), and the AUC for detecting residual or recurrent tumor was 0.84 (95% CI: 0.79, 0.89). In the internal test set, estimated cellular tumor volume was significantly associated with OS (hazard ratio [HR] = 1.04 per milliliter; P < .001) and PFS (HR = 1.04 per milliliter; P < .001) after adjustment for age, sex, and gross total resection (GTR) status. In the external test sets, estimated cellular tumor volume was significantly associated with OS (HR = 1.01 per milliliter; P < .001) after adjustment for age, sex, and GTR status. Conclusion A DL model incorporating advanced imaging could accurately segment enhancing and nonenhancing cellular tumor, distinguish recurrent or residual tumor from posttreatment changes, and predict OS and PFS in patients with glioblastoma. Keywords: Segmentation, Glioblastoma, Multishell Diffusion MRI Supplemental material is available for this article. © RSNA, 2024.

Background: Diffusion-weighted imaging (DWI) is commonly used to detect prostate cancer, and a major clinical challenge is differentiating aggressive from indolent disease., Purpose: To compare 14 site-specific parametric fitting implementations applied to the same dataset of whole-mount pathologically validated DWI to test the hypothesis that cancer differentiation varies with different fitting algorithms., Study Type: Prospective., Population: Thirty-three patients prospectively imaged prior to prostatectomy., Field Strength/sequence: 3 T, field-of-view optimized and constrained undistorted single-shot DWI sequence., Assessment: Datasets, including a noise-free digital reference object (DRO), were distributed to the 14 teams, where locally implemented DWI parameter maps were calculated, including mono-exponential apparent diffusion coefficient (MEADC), kurtosis (K), diffusion kurtosis (DK), bi-exponential diffusion (BID), pseudo-diffusion (BID*), and perfusion fraction (F). The resulting parametric maps were centrally analyzed, where differentiation of benign from cancerous tissue was compared between DWI parameters and the fitting algorithms with a receiver operating characteristic area under the curve (ROC AUC)., Statistical Test: Levene's test, P < 0.05 corrected for multiple comparisons was considered statistically significant., Results: The DRO results indicated minimal discordance between sites. Comparison across sites indicated that K, DK, and MEADC had significantly higher prostate cancer detection capability (AUC range = 0.72-0.76, 0.76-0.81, and 0.76-0.80 respectively) as compared to bi-exponential parameters (BID, BID*, F) which had lower AUC and greater between site variation (AUC range = 0.53-0.80, 0.51-0.81, and 0.52-0.80 respectively). Post-processing parameters also affected the resulting AUC, moving from, for example, 0.75 to 0.87 for MEADC varying cluster size., Data Conclusion: We found that conventional diffusion models had consistent performance at differentiating prostate cancer from benign tissue. Our results also indicated that post-processing decisions on DWI data can affect sensitivity and specificity when applied to radiological-pathological studies in prostate cancer., Level of Evidence: 1 TECHNICAL EFFICACY: Stage 3., (© 2021 The Authors. Journal of Magnetic Resonance Imaging published by Wiley Periodicals LLC on behalf of International Society for Magnetic Resonance in Medicine.)

The goal of this study is to spatially discriminate tumor from treatment effect (TE), within the contrast-enhancing lesion, for brain tumor patients at all stages of treatment. To this end, the diagnostic accuracy of MRI-derived diffusion and perfusion parameters to distinguish pure TE from pure glioblastoma (GBM) was determined utilizing spatially-correlated biopsy samples. From July 2010 through June 2015, brain tumor patients who underwent pre-operative DWI and DSC-MRI and stereotactic image-guided biopsy were considered for inclusion in this IRB-approved study. MRI-derived parameter maps included apparent diffusion coefficient (ADC), normalized cerebral blood flow (nCBF), normalized and standardized relative cerebral blood volume (nRCBV, sRCBV), peak signal-height (PSR) and percent signal-recovery (PSR). These were co-registered to the Stealth MRI and median values extracted from the spatially-matched biopsy regions. A ROC analysis accounting for multiple subject samples was performed, and the optimal threshold for distinguishing TE from GBM determined for each parameter. Histopathologic diagnosis of pure TE (n = 10) or pure GBM (n = 34) was confirmed in tissue samples from 15 consecutive subjects with analyzable data. Perfusion thresholds of sRCBV (3575; SN/SP% = 79.4/90.0), nRCBV (1.13; SN/SP% = 82.1/90.0), and nCBF (1.05; SN/SP% = 79.4/80.0) distinguished TE from GBM (P < 0.05), whereas ADC, PSR, and PH could not (P > 0.05). The thresholds for CBF and CBV can be applied to lesions with any admixture of tumor or treatment effect, enabling the identification of true tumor burden within enhancing lesions. This approach overcomes current limitations of averaging values from both tumor and TE for quantitative assessments.

Purpose: To examine the effect of low b-values (0 < b < 50 s/mm(2)) on the calculation of the intravoxel incoherent motion (IVIM) derived pseudodiffusion parameter in the normal liver., Methods: Simulations were performed to examine the effects of adding low b-values on the pseudodiffusion parameter. Low b-values were cumulatively added to the distribution and the IVIM signal was generated with varying pseudodiffusion values. The signal was fit with the IVIM model after the addition of Gaussian noise, and the simulated values were compared with the true values. In addition, the livers of eight control subjects were imaged using respiratory-triggered DWI. Pseudodiffusion was calculated with and without low b-values and compared., Results: Pseudodiffusion tended to be underestimated when low b-values were not included in the b-value distribution as predicted by simulations and confirmed with in vivo imaging. The number of outlier values was also reduced as more low b-values were added., Conclusion: In conclusion, this study showed pseudodiffusion in the liver tended to be underestimated when too few low b-values (0 < b < 50 s/mm(2)) were included in the distribution. Therefore, it is recommended to include at least two low b-values when performing liver IVIM studies., (© 2014 Wiley Periodicals, Inc.)

Background: Recent conflicting reports have found both brain tumor hypercellularity and necrosis in regions of restricted diffusion on MRI-derived apparent diffusion coefficient (ADC) images. This study precisely compares ADC and cell density voxel by voxel using postmortem human whole brain samples., Methods: Patients with meningioma were evaluated to determine a normative ADC distribution within benign fluid attenuated inversion recovery (FLAIR) T2/hyperintensity surrounding tumor. This distribution was used to calculate a minimum ADC threshold to define regions of ADC-FLAIR mismatch (AFMM), where restricted diffusion presented in conjunction with T2/FLAIR hyperintensity. Contrast-enhancing voxels were excluded from this analysis. AFMM maps were generated using imaging acquired prior to death in 7 patients with high-grade glioma who eventually donated their brains upon death. Histological samples were taken from numerous regions of abnormal FLAIR and AFMM. Each sample was computationally processed to determine cell density. Custom software was then used to downsample coregistered microscopic histology to the more coarse MRI resolution. A voxel-by-voxel evaluation comparing ADC and cellularity was then performed., Results: An ADC threshold of 0.929 × 10(-3) mm(2)/s was calculated from meningioma-induced edema and was used to define AFMM. Regions of AFMM showed significantly greater cell density in 6 of 7 high-grade glioma cases compared with regions of hyperintense FLAIR alone (P < .0001). Two patients had small regions of diffusion-restricted necrosis that had significantly lower ADC than nearby hypercellularity., Conclusions: Regions of AFMM contain hypercellularity except for regions with extremely restricted diffusion, where necrosis is present., (© The Author(s) 2014. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

Purpose: To characterize the influence of perfusion on the measurement of diffusion changes over time when ADC is computed using standard two-point methods., Materials and Methods: Functional diffusion maps (FDMs), which depict changes in diffusion over time, were compared with rCBV changes in patients with brain tumors. The FDMs were created by coregistering and subtracting ADC maps from two time points and categorizing voxels where ADC significantly increased (iADC), decreased (dADC), or did not change (ncADC). Traditional FDMs (tFDMs) were computed using b = 0,1000 s/mm(2). Flow-compensated FDMs (fcFDMs) were calculated using b = 500,1000 s/mm(2). Perfusion's influence on FDMs was determined by evaluating changes in rCBV in areas where the ADC change significantly differed between the two FDMs., Results: The mean ΔrCBV in voxels that changed from iADC (dADC) on the tFDM to ncADC on the fcFDM was significantly greater (less) than zero. In addition, mean ΔrCBV in iADC (dADC) voxels on the tFDM was significantly higher (lower) than in iADC (dADC) voxels on the fcFDM., Conclusion: The ability to accurately identify changes in diffusion on traditional FDMs is confounded in areas where perfusion and diffusion changes are colocalized. Flow-compensated FDMs, which use only non-zero b-values, should therefore be the standard approach., (Copyright © 2013 Wiley Periodicals, Inc.)

Diffusion-weighted imaging (DWI) is a powerful MRI method, which probes abnormalities of tissue structure by detecting microscopic changes in water mobility at a cellular level beyond what is available with other imaging techniques. Accordingly, DWI has the potential to identify pathology before gross anatomic changes are evident on standard anatomical brain images. These features of tissue characterization and earlier detection are what make DWI particularly appealing for the evaluation of gliomas and the newer therapies where standard anatomical imaging is proving insufficient. This article focuses on the basic principles and applications of DWI, and its derived parameter, the apparent diffusion coefficient, for the purposes of diagnosis and evaluation of glioma, especially in the context of monitoring response to therapy.

Diffusion imaging has shown promise as a predictive and prognostic biomarker in glioma. We assessed the ability of graded functional diffusion maps (fDMs) and apparent diffusion coefficient (ADC) characteristics to predict overall survival (OS) in recurrent glioblastoma multiforme (GBM) patients treated with bevacizumab. Seventy-seven patients with recurrent GBMs were retrospectively examined. MRI scans were obtained before and approximately 6 weeks after treatment with bevacizumab. Graded fDMs were created by registering datasets to each patient's pretreatment scan and then performing voxel-wise subtraction between post- and pretreatment ADC maps. Voxels were categorized according to the degree of change in ADC within pretreatment fluid-attenuated inversion recovery (FLAIR) and contrast-enhancing regions of interest (ROIs). We found that the volume of tissue showing decreased ADC within both FLAIR and contrast-enhancing regions stratified OS (log-rank, P < .05). fDMs applied to contrast-enhancing ROIs more accurately predicted OS compared with fDMs applied to FLAIR ROIs. Graded fDMs (showing voxels with decreased ADC between 0.25 and 0.4 µm(2)/ms) were more predictive of OS than traditional (single threshold) fDMs, and the predictive ability of graded fDMs could be enhanced even further by adding the ADC characteristics from the fDM-classified voxels to the analysis (log-rank, P < .001). These results demonstrate that spatially resolved diffusion-based tumor metrics are a powerful imaging biomarker of survival in patients with recurrent GBM treated with bevacizumab.

The purpose of this study was to develop a voxel-wise analytical solution to a glioma growth model using serial diffusion MRI. These cell invasion, motility, and proliferation level estimates (CIMPLE maps) provide quantitative estimates of microscopic tumor growth dynamics. After an analytical solution was found, noise simulations were performed to predict the effects that perturbations in apparent diffusion coefficient values and the time between apparent diffusion coefficient map acquisitions would have on the accuracy of CIMPLE maps. CIMPLE maps were then created for 53 patients with gliomas with WHO grades of II-IV. MR spectroscopy estimates of the choline-to-N-acetylaspartate ratio were compared to cell proliferation estimates in CIMPLE maps using Pearson's correlation analysis. Median differences in cell proliferation and diffusion rates between WHO grades were compared. A strong correlation (R(2) = 0.9714) and good spatial correspondence were observed between MR spectroscopy measurements of the choline-to-N-acetylaspartate ratio and CIMPLE map cell proliferation rate estimates. Estimates of cell proliferation and diffusion rates appear to be significantly different between low- (WHO II) and high-grade (WHO III-IV) gliomas. Cell diffusion rate (motility) estimates are highly dependent on the time interval between apparent diffusion coefficient map acquisitions, whereas cell proliferation rate estimates are additionally influenced by the level of noise present in apparent diffusion coefficient maps., (Copyright © 2010 Wiley-Liss, Inc.)

Anti-angiogenic agents targeting brain tumor neovasculature may increase progression-free survival in patients with recurrent malignant gliomas. However, when these patients do recur it is not always apparent as an increase in enhancing tumor volume on MRI, which has been the standard of practice for following patients with brain tumors. Therefore alternative methods are needed to evaluate patients treated with these novel therapies. Furthermore, a method that can also provide useful information for the evaluation of conventional therapies would provide an important advantage for general applicability. Diffusion-weighted magnetic resonance imaging (DWI) has the potential to serve as a valuable biomarker for these purposes. In the current study, we explore the prognostic ability of functional diffusion maps (fDMs), which examine voxel-wise changes in the apparent diffusion coefficient (ADC) over time, applied to regions of fluid-attenuated inversion recovery (FLAIR) abnormalities in patients with malignant glioma, treated with either anti-angiogenic or cytotoxic therapies. Results indicate that the rate of change in fDMs is an early predictor of tumor progression, time to progression and overall survival for both treatments, suggesting the application of fDMs in FLAIR abnormal regions may be a significant advance in brain tumor biomarker technology.

Diffusion-weighted MRI is an intrinsically low signal-to-noise ratio application due to the application of diffusion-weighting gradients and the consequent longer echo times. The signal-to-noise ratio worsens with increasing image resolution and diffusion imaging methods that use multiple and higher b-values. At low signal-to-noise ratios, standard magnitude reconstructed diffusion-weighted images are confounded by the existence of a rectified noise floor, producing poor estimates of diffusion metrics. Herein, we present a simple method of rectified noise floor suppression that involves phase correction of the real data. This approach was evaluated for diffusion-weighted imaging data, obtained from ethanol and water phantoms and the brain of a healthy volunteer. The parameter fits from monoexponential, biexponential, and stretched-exponential diffusion models were computed using phase-corrected real data and magnitude data. The results demonstrate that this newly developed simple approach of using phase-corrected real images acts to reduce or even suppress the confounding effects of a rectified noise floor, thereby producing more accurate estimates of diffusion parameters.

Diffusion-weighted magnetic resonance imaging (DWI) is a sensitive imaging biomarker for tumor cellularity. Functional diffusion maps (fDMs), which examine voxel-by-voxel changes in the apparent diffusion coefficient (ADC) calculated from serial DWIs, have previously been applied to regions of contrast-enhancement; however, application of fDMs to non-enhancing brain tumors has not been pursued. In this case study we demonstrate the utility of applying fDMs to regions of abnormal FLAIR signal intensity in a patient diagnosed with gliomatosis cerebri: a relatively rare, infiltrative, non-enhancing brain tumor. The absolute volume of hypercellularity extracted from fDMs was useful in tracking tumor growth, which correlated in time with a progressive decline in neurological status despite no change in traditional magnetic resonance images. Results of this study demonstrate the value of fDMs, applied to regions of FLAIR abnormal signal intensity, for localizing regions of hypercellularity and for monitoring overall tumor status.

Purpose: To present comprehensive examinations of the assumptions made in functional diffusion map (fDM) analyses and provide a biological basis for fDM classification., Materials and Methods: Sixty-nine patients with gliomas were enrolled in this study. To determine the sensitivity of apparent diffusion coefficients (ADCs) to cellularity, cell density from stereotactic biopsy specimens was correlated with preoperative ADC maps. For definition of ADC thresholds used for fDMs, the 95% confidence intervals (CI) for changes in voxel-wise ADC measurements in normal appearing tissue was analyzed. The sensitivity and specificity to progressing disease was examined using both radiographic and neurological criteria., Results: Results support the hypothesis that ADC is inversely proportional to cell density with a sensitivity of 1.01 x 10(-7) [mm(2)/s]/[nuclei/mm(2)]. The 95% CI for white matter = 0.25 x 10(-3) mm(2)/s, gray matter = 0.31 x 10(-3) mm(2)/s, a mixture of white and gray matter = 0.40 x 10(-3) mm(2)/s, and a mixture of white matter, gray matter, and cerebrospinal fluid = 0.75 x 10(-3) mm(2)/s. Application of these measurements as ADC thresholds produce varying levels of sensitivity and specificity to disease progression, which were all significantly better than chance., Conclusion: This study suggests fDMs are valid biomarkers for brain tumor cellularity.

The alpha diffusion-weighted imaging (DWI) method was developed to study heterogeneous water diffusion in the human brain using magnetic resonance imaging (MRI). An advantage of this model is that it does not require an assumption about the shape of the intravoxel distribution of apparent diffusion rates, and it has a calculable relationship to this distribution. The alpha-DWI technique is useful for detecting microstructural tissue changes associated with brain tumor invasion, and may be useful for directing therapy to invading tumor cells. In previous work, alpha-DWI was performed with magnetic field gradients applied along a single direction in order to avoid artificially introducing a source of heterogeneity to the decay. However, it is known that restricted diffusion is anisotropic in the brain, and the alpha-DWI method must take this into account to be complete. In this work the relationship between the applied magnetic field gradients and the fitted stretched-exponential model parameters was studied in the human brain. It was found the distributed diffusion coefficient (DDC) varies with the direction of applied gradients, while the heterogeneity index alpha is relatively direction-insensitive. It is proposed that in clinical use, maps of alpha can be created using diffusion-weighting gradients applied in a single direction that reflect the tissue heterogeneity.

The hypothesis was tested that the intravoxel distribution of water diffusion rates, as measured with a stretched-exponential model of diffusion-weighted imaging (DWI), is a marker of brain tumor invasion. Eight rats underwent intracerebral inoculation of C6 glioma cells. In three rats, cells were labeled with a fluorescent dye for microscopy. One rat was inoculated with a saline solution, and five more rats were imaged without inoculation as controls. Five healthy uninoculated rats were also imaged. DWI was performed 14-15 days after inoculation, with diffusion-weighting factor b = 500 to 6500 sec/mm2, and the resulting signal attenuation was fitted with the stretched-exponential model. The heterogeneity index values were significantly lower (P < 0.05) in the peritumor ROI than in normal gray matter and significantly higher than in normal white matter. The distributed diffusion coefficient values were significantly lower than in normal white matter or normal gray matter. Fluorescence microscopy confirmed the presence of tumors in the peritumor region that could be histologically distinguished from the main tumor mass. There was no change in proton density or T2-weighted images in the peritumor region, making vasogenic edema unlikely as a source of contrast. It is therefore thought that the heterogeneity parameter alpha is a marker of brain tumor invasion., ((c) 2004 Wiley-Liss, Inc.)

Experience with diffusion-weighted imaging (DWI) shows that signal attenuation is consistent with a multicompartmental theory of water diffusion in the brain. The source of this so-called nonexponential behavior is a topic of debate, because the cerebral cortex contains considerable microscopic heterogeneity and is therefore difficult to model. To account for this heterogeneity and understand its implications for current models of diffusion, a stretched-exponential function was developed to describe diffusion-related signal decay as a continuous distribution of sources decaying at different rates, with no assumptions made about the number of participating sources. DWI experiments were performed using a spin-echo diffusion-weighted pulse sequence with b-values of 500-6500 s/mm(2) in six rats. Signal attenuation curves were fit to a stretched-exponential function, and 20% of the voxels were better fit to the stretched-exponential model than to a biexponential model, even though the latter model had one more adjustable parameter. Based on the calculated intravoxel heterogeneity measure, the cerebral cortex contains considerable heterogeneity in diffusion. The use of a distributed diffusion coefficient (DDC) is suggested to measure mean intravoxel diffusion rates in the presence of such heterogeneity., (Copyright 2003 Wiley-Liss, Inc.)