Tissue Classification from Brain Perfusion MR Images Using Expectation-Maximization Algorithm Initialized by Hierarchical Clustering on Whitened Data

Expectation-maximization (EM) algorithm initialized by hierarchical clustering (HC) was applied on dynamic susceptibility contrast (DSC) MR images from the patients with unilateral internal carotid artery stenosis to segment out different brain tissue clusters depending on their own specific blood supply patterns. In comparison with the segmented normal and abnormal gray matter components demonstrated that difference in mean transit time (dMTT) and difference in time to peak (dTTP) can robustly reveal the hemodynamic change from pre-stenting to post-stenting state (p-values are 0.027 and 0.004, respectively). Additionally, change of local deficit before and after the placement of stent can be further investigated by the ratio of numbers of normal to abnormal gray-matter pixels within the territories of anterior cerebral artery (ACA), middle cerebral artery (MCA) and posterior cerebral artery (PCA) (p-values are 0.375, 0.037 and 0.020, respectively) in assistance to diagnosis and therapeutic assessment.

How far is arterial spin labeling MRI from a clinical reality? Insights from arterial spin labeling comparative studies in Alzheimer's disease and other neurological disorders.

Objective:Arterial spin-labelling (ASL) MRI uses intrinsic blood water to quantify the cerebral blood flow (CBF), removing the need for the injection of a gadolinium-based contrast agent used for conventional perfusion imaging such as dynamic susceptibility contrast (DSC). Owing to the non-invasive nature of the technique, ASL is an attractive option for use in paediatric patients. This work compared DSC and multi-timepoint ASL measures of CBF in paediatric brain tumours.Methods:Patients (n = 23; 20 low-grade tumours and 3 high-grade tumours) had DSC and multi-timepoint ASL with and without vascular crushers (VC). VC removes the contribution from larger vessel blood flow. Mean perfusion metrics were extracted from control and T1-enhanced tumour regions of interest (ROIs): arterial arrival time (AAT) and CBF from the ASL images with and without VC, relative cerebral blood flow (rCBF), relative cerebral blood volume, delay time (DT) and mean transit time (MTT) from the DSC images.Results:Significant correlations existed for: AAT and DT (r = 0.77, p = 0.0002) and CBF and rCBF (r = 0.56, p = 0.02) in control ROIs for ASL-noVC. No significant correlations existed between DSC and ASL measures in the tumour region. Significant differences between control and tumour ROI were found for MTT (p < 0.001) and rCBF (p < 0.005) measures.Conclusion:Significant correlations between ASL-noVC and DSC measures in the normal brain suggest that DSC is most sensitive to macrovascular blood flow. The absence of significant correlations within the tumour ROI suggests that ASL is sensitive to different physiological mechanisms compared with DSC measures.Advances in knowledge:ASL provides information which is comparable with that of DSC in healthy tissues, but appears to reflect a different physiology in tumour tissues.

In Hierarchical Clustering (HC), a set of patterns is partitioned into a sequence of groups and the results are repre- sented by a dendrogram. The dendrogram is a tree representation where each node is associated with merging of two (or more) partitions and hence each partition is nested into the next partition. Hierarchical representation has valuable properties that are useful for visualization and interpretation of the clustering results. On one hand, it is well known that hierarchical clustering results are sensitive to the algorithms that are used, thus different dendrograms are created. On the other hand, clustering com- bination methods have demonstrated promising results. This motivated us to consider different combination method for a set of dendrograms. In this paper, we compare several weighted combination methods in which the hierarchical clustering results are used in order to derive a consensus hierarchical clustering. The Cophenetic measures are used for comparison. The results illustrate that the proposed weighing combination of hierarchical clusterings perform better than other combinations such as averaging.

Increased calcium influx secondary to glutamate induced excitotoxicity initiates and potentiates devastating pathological changes following ischemic stroke. Pertussis toxin (PTx), a G-protein blocker, is known to suppress intracellular calcium accumulation. We hypothesize that PTx can protect against stroke by blocking calcium influx. In a permanent middle cerebral artery occlusion model, PTx (1000 ng) was given intraperitoneally 30 min after inducing stroke. Magnetic Resonance Imaging of perfusion and T2-weighted brain scans were obtained to evaluate cerebral blood flow (CBF) and infarct volume. Primary neuronal culture was used to test glutamate induced excitotoxicity and calcium influx. We established a non-linear exponential curve model to minimize variations in animal cerebrovasculature. A reduction of 40-60% in relative CBF was a critical window where infarct volume started to increase as rCBF reduced. PTx showed maximal effects in reducing infarct volume at this window. In vitro studies further demonstrated PTx increased neuronal cell survival by decreasing glutamate-induced calcium influx into neurons and preventing neurons from apoptosis. PTx salvages the ischemic penumbra by blocking calcium influx. This provides us a new mechanism upon which experimental therapies can be explored to treat ischemic stroke. In ischemic stroke, excessive glutamate binds to AMPA receptor that depolarizes calcium channel and/ or NMDA receptor. Both of them allow calcium to enter the cell. The overload of calcium triggers cellular cascade that includes Caspase activation and release, leading to pre-mature cell death. We have demonstrated that PTx, a G-protein inhibitor, blocks calcium entry which in turn prevents further cellular damage.

Given the limitations of conventional imaging in accurately grading gliomas, predicting molecular subtypes, and assessing tumor proliferation and angiogenesis, there is a growing need for advanced quantitative magnetic resonance imaging (MRI) biomarkers. This study aimed to compare the diagnostic performance of histogram features of dynamic contrast enhanced (DCE) and dynamic susceptibility contrast (DSC) imaging in predicting glioma grade and genotyping, as well as to explore the association between DCE and DSC with Ki-67 and microvascular density (MVD). Forty-six patients with gliomas were enrolled prospectively. The histogram features of DCE and DSC were extracted from the entire tumor and compared among subgroups based on the grades and status of isocitrate dehydrogenase (IDH) mutation and 1p/19q co-deletion. Pearson correlation analyses were employed to examine the associations among Ki-67, MVD, and histogram features of perfusion model. Receiver operating characteristic (ROC) analyses were used to evaluate the accuracy of DCE and DSC in differentiating glioma grades and genotypes, while Cox regression analyses identified prognostic factors associated with survival after surgical resection. A cohort of 46 patients with IDH-mutant (n=23) and IDH-wildtype (n=23) gliomas was analyzed. DCE-MRI histogram features demonstrated superior performance to DSC parameters in both IDH genotyping and tumor grading. For IDH discrimination, 90th percentile of volume fraction of the extravascular-extracellular space (Ve) [area under the curve (AUC) =0.849, sensitivity 91.3%, specificity 78.3%] and mean value of volume transfer constant (Ktrans) (AUC =0.847, sensitivity 87.0%, specificity 82.6%) showed optimal performance (both P<0.001). In contrast, DSC-derived relative cerebral blood volume (rCBV) (AUC =0.577) and relative cerebral blood flow (rCBF) (AUC =0.537) exhibited limited diagnostic value (P>0.05). In tumor grading, 90th percentile of Ve (AUC =0.923) and mean value of Ktrans (AUC =0.908) achieved near-perfect differentiation (sensitivity 83.9%, specificity 93.3%; P<0.001). While DSC-derived parameters showed improved performance compared to genotyping (rCBV: AUC =0.742, P=0.008; rCBF: AUC =0.688, P=0.040), DCE-based models remained significantly superior (DeLong's test P<0.05 for all comparisons). Significant correlations were observed between DCE parameters and both Ki-67 proliferation index (r=0.42-0.51) and MVD (r=0.35-0.56; all P<0.05). Survival analysis identified three independent prognostic factors: O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation [hazard ratio (HR) =4.15, 95% confidence interval (CI): 1.10-15.68, P=0.04], elevated Ktrans (HR =1.08, 95% CI: 1.01-1.17, P=0.03), and rCBV (HR =1.87, 95% CI: 1.17-2.97, P=0.009). The histogram features of DCE and DSC demonstrate exceptional diagnostic efficacy in glioma grading, with DCE providing deeper insights into the molecular characteristics of glioma.

Brain perfusion evaluated by perfusion-weighted magnetic resonance imaging before and after stenting internal carotid artery stenosis in asymptomatic and symptomatic patients

Alterations in density and morphology of the cerebral microvasculature have been reported to occur in Alzheimer's disease patients and animal models of the disease. In this study we compared magnetic resonance imaging (MRI) techniques for their utility to detect age-dependent changes of the cerebral vasculature in the arcAβ mouse model of cerebral amyloidosis. Dynamic susceptibility contrast (DSC)-MRI was performed by tracking the passage of a superparamagnetic iron oxide nanoparticle in the brain with dynamic gradient echo planar imaging (EPI). From this measurements relative cerebral blood volume [rCBV(DSC)] and relative cerebral blood flow (rCBF) were estimated. For the same animal maps of the relaxation shift index Q were computed from high resolution gradient echo and spin echo data that were acquired before and after superparamagnetic iron oxide (SPIO) nanoparticle injection. Q-values were used to derive estimates of microvessel density. The change in the relaxation rates obtained from pre- and post-contrast gradient echo data was used for the alternative determination of rCBV [rCBV()]. Linear mixed effects modeling found no significant association between rCBV(DSC), rCBV(), rCBF, and Q with genotype in 13-month old mice [compared to age-matched non-transgenic littermates (NTLs)] for any of the evaluated brain regions. In 24-month old mice there was a significant association for rCBV(DSC) with genotype in the cerebral cortex, and for rCBV() in the cerebral cortex and cerebellum. For rCBF there was a significant association in the cerebellum but not in other brain regions. Q-values in the olfactory bulb, cerebral cortex, striatum, hippocampus, and cerebellum in 24-month old mice were significantly associated with genotype. In those regions Q-values were reduced between 11 and 26% in arcAβ mice compared to age-matched NTLs. Vessel staining with CD31 immunohistochemistry confirmed a reduction of microvessel density in the old arcAβ mice. We further demonstrated a region-specific association between parenchymal and vascular deposition of β-amyloid and decreased vascular density, without a correlation with the amount of Aβ deposition. We found that Q mapping was more suitable than the hemodynamic read-outs to detect amyloid-related degeneration of the cerebral microvasculature.

ObjectiveTo assess the role of arterial spin-labeling (ASL) perfusion MRI in identifying cerebral perfusion changes after indirect revascularization in children with moyamoya disease.Materials and MethodsWe included pre- and postoperative perfusion MRI data of 30 children with moyamoya disease (13 boys and 17 girls; mean age ± standard deviation, 6.3 ± 3.0 years) who underwent indirect revascularization between June 2016 and August 2017. Relative cerebral blood flow (rCBF) and qualitative perfusion scores for arterial transit time (ATT) effects were evaluated in the middle cerebral artery (MCA) territory on ASL perfusion MRI. The rCBF and relative time-to-peak (rTTP) values were also measured using dynamic susceptibility contrast (DSC) perfusion MRI. Each perfusion change on ASL and DSC perfusion MRI was analyzed using the paired t test. We analyzed the correlation between perfusion changes on ASL and DSC images using Spearman's correlation coefficient.ResultsThe ASL rCBF values improved at both the ganglionic and supraganglionic levels of the MCA territory after surgery (p = 0.040 and p = 0.003, respectively). The ATT perfusion scores also improved at both levels (p < 0.001 and p < 0.001, respectively). The rCBF and rTTP values on DSC MRI showed significant improvement at both levels of the MCA territory of the operated side (all p < 0.05). There was no significant correlation between the improvements in rCBF values on the two perfusion images (r = 0.195, p = 0.303); however, there was a correlation between the change in perfusion scores on ASL and rTTP on DSC MRI (r = 0.701, p < 0.001).ConclusionRecognizing the effects of ATT on ASL perfusion MRI may help monitor cerebral perfusion changes and complement quantitative rCBF assessment using ASL perfusion MRI in patients with moyamoya disease after indirect revascularization.

Tumor heterogeneity complicates the quantification of tumor microvascular characteristics assessed by dynamic contrast-enhanced MRI (DCE-MRI). To address this issue a novel approach was developed that combines DCE-MRI with diffusion-based multispectral (MS) analysis to quantify the microvascular characteristics of specific tumor tissue populations. Diffusion-based MS segmentation (feature space: apparent diffusion coefficient, T(2) and proton density) was performed to identify tumor tissue populations and the DCE-MRI characteristics were determined for each tissue class. The ability of this MS DCE-MRI technique to detect microvascular changes due to treatment with an antibody (G6-31) to vascular endothelial growth factor-A (VEGF) was evaluated in a tumor xenograft mouse model. Anti-VEGF treatment resulted in a significant reduction in K(trans) for the MS viable tumor tissue class (-0.0034 +/- 0.0022 min(-1), P < 0.01) at 24 hr posttreatment that differ significantly from the change observed in the control group (0.0002 +/- 0.0025 min(-1)). Viable tumor K(trans) for the anti-VEGF group was also reduced 62% relative to the pretreatment values (P < 0.01). Necrotic tissue classes were found to add only noise to DCE-MRI estimates. This approach provides a means to measure physiological parameters within the viable tumor and address the issue of tumor heterogeneity that complicates DCE-MRI analysis.

Purpose: To prove the feasibility of arterial spin labeling (ASL) to explore brain tumors by comparing dynamic susceptibility contrast-enhanced MRI to ASL at 3T MR. Materials and Methods: Twenty-seven patients were included presenting 9 gliomas, 10 metastases and 8 meningiomas. All were explored by a pseudo-continuous ASL and dynamic susceptibility contrast-enhanced T2* perfusion sequence. Two neuroradiologists analyzed the cerebral blood flow (CBF) maps to assess feasibility, examination quality and quantitative comparison. The Spearman nonparametric correlation test and the Bland-Altman graphic test were used to analyze our quantitative results. Results: 92% of ASL CBF maps were informative. ASL detected all lesions as well as dynamic susceptibility contrast-enhanced MRI. Both sequences provide relative quantitative CBF values closely correlated. Conclusion: On a 3T MR unit, ASL is a good alternative to dynamic susceptibility contrast-enhanced MRI when contrast medium is contraindicated or intravenous injection is not possible. Its results on relative CBF parameters are similar to contrast-injected perfusion.

We report perfusion weighted imaging (PWI) findings of nonenhanced anaplastic astrocytoma in a 30-year-old woman. Brain magnetic resonance imaging showed a nonenhanced brain tumor with mild peritumoral edema on the right medial frontal lobe and right genu of corpus callosum, suggesting a low-grade glioma. However, PWI showed increased relative cerebral blood volume, relative cerebral blood flow, and permeability of nonenhanced brain tumor compared with contralateral normal brain parenchyma, suggesting a high-grade glioma. After surgery, final histopathological analysis revealed World Health Organization grade III anaplastic astrocytoma. This case demonstrates the importance of PWI for preoperative evaluation of nonenhanced brain tumors.

Accurately locating the epileptogenic focus in temporal lobe epilepsy (TLE) is important in clinical practice. Single-photon emission computed tomography (SPECT) and positron-emission tomography (PET) have been widely used in the lateralization of TLE, but both have limitations. Magnetic resonance perfusion imaging can accurately and reliably reflect differences in cerebral blood flow and volume. To investigate the diagnostic value of dynamic susceptibility contrast-enhanced (DSC) perfusion magnetic resonance imaging (MRI) in the lateralization of the epileptogenic focus in TLE. Conventional MRI and DSC-MRI scanning was performed in 20 interictal cases of TLE and 20 healthy volunteers. The relative cerebral blood volume (rCBV) and relative cerebral blood flow (rCBF) of the bilateral mesial temporal lobes of the TLE cases and healthy control groups were calculated. The differences in the perfusion asymmetry indices (AIs), derived from the rCBV and rCBF of the bilateral mesial temporal lobes, were compared between the two groups. In the control group, there were no statistically significant differences between the left and right sides in terms of rCBV (left 1.55 ± 0.32, right 1.57 ± 0.28) or rCBF (left 99.00 ± 24.61, right 100.38 ± 23.46) of the bilateral mesial temporal lobes. However, in the case group the ipsilateral rCBV and rCBF values (1.75 ± 0.64 and 96.35 ± 22.63, respectively) were markedly lower than those of the contralateral side (2.01 ± 0.79 and 108.56 ± 26.92; P < 0.05). Both the AI of the rCBV (AI(rCBV); 13.03 ± 10.33) and the AI of the rCBF (AI(rCBF); 11.24 ± 8.70) of the case group were significantly higher than that of the control group (AI(rCBV) 5.55 ± 3.74, AI(rCBF) 5.12 ± 3.48; P < 0.05). The epileptogenic foci of nine patients were correctly lateralized using the 95th percentile of the AI(rCBV) and AI(rCBF) of the control group as the normal upper limits. In patients with TLE interictal, both rCBV and rCBF of the ipsilateral mesial temporal lobe were markedly lower than that of healthy control subjects. DSC-MRI can provide lateralization for TLE.

Purpose To compare arterial spin labeling (ASL) data between low- and high-grade brain tumors in children to establish a cutoff to distinguish low- from high-grade neoplasms and to assess potential correlations between cerebral blood flow (CBF) and quantitative histologic microvascular data. Materials and Methods Approval was obtained from the regional review board. ASL data obtained in 129 children between 2011 and 2015 were retrospectively analyzed. CBF and relative CBF in the most perfused area of each neoplasm and contrast enhancement were quantified with a semiquantitative ratio. The correlation between CBF and microvascular density was analyzed in specimens stained with anti-CD34. Results were controlled in two validation cohorts with 1.5- and 3.0-T magnetic resonance (MR) imaging. Results Mean CBF was significantly higher for high-grade than for low-grade hemispheric (116 mL/min/100 g [interquartile range {IQR}, 73-131 mL/min/100 g] vs 29 mL/min/100 g [IQR, 23-35 29 mL/min/100 g], P < .001), thalamic (87 mL/min/100 g [IQR, 73-100 mL/min/100 g] vs 36 mL/min/100 g [IQR, 30-40 mL/min/100 g], P = .016), and posterior fossa (59 mL/min/100 g [IQR, 45-91 mL/min/100 g] vs 33 mL/min/100 g [IQR, 25-40 mL/min/100 g], P < .001) tumors. With a cutoff of 50 mL/min/100 g, sensitivity and specificity were 90% (95% confidence interval [CI]: 68, 100) and 93% (95% CI: 66, 100), respectively, for hemispheric tumors; 100% (95% CI: 48, 100) and 80% (95% CI: 28, 100), respectively, for thalamic tumors; and 65% (95% CI: 51, 78) and 94% (95% CI: 80, 99), respectively, for posterior fossa tumors. In posterior fossa tumors, additional use of the CBF-to-contrast enhancement ratio yielded sensitivity and specificity of 96% (95% CI: 87, 100) and 97% (95% CI: 84, 100), respectively. Use of a simple algorithm based on these values yielded an accuracy of 93% (95% CI: 87, 97). Validation sets yielded similar results, with grading accuracy of 88% (95% CI: 62, 98) with 1.5-T MR imaging and 77% (95% CI: 46, 95) with 3.0-T MR imaging. CBF was strongly correlated with microvascular density (R = 0.66, P < .001). Conclusion High-grade pediatric brain tumors display higher CBF than do low-grade tumors, and they may be accurately graded by using these values. CBF is correlated with tumor microvascular density. © RSNA, 2016 Online supplemental material is available for this article.

Background : Examination procedure of Magnetic resonance imaging (MRI) Brain Perfusion in patient with tumor disease at Radiology Department of National Brain Hospital used the MR Perfusion sequence with the Dynamic Susceptibility Contrast (DSC) method with the addition of Arterial Spin Labeling (ASL) method which was also used to determine CBF values. Nevertheless CBF values can also be obtained from the use of the DSC method. The application of these two roles in patient with tumor disease certainly has their respective roles in providing accurate information for the diagnosis. The purpose of this study was to determine the role of the application of MR perfusion with the Arterial Spin Labeling (ASL) method and the Dynamic Susceptibility Contrast (DSC) method for the diagnosis of patient with tumor disease at Radiology Department of National Brain Hospital.&#x0D; Methods :The type of this research is qualitative with case study approach. The data were collected from March – April 2019 at at Radiology Department of National Brain Hospital by using observation method, interview with radiology specialist doctor, sending doctor, FGD, and documentation. Data obtained from the study were analyzed by making the transcript then reduced in the form of categorization table and open coding, presented in quote form and then can be drawn conclusion.&#x0D; Result : Brain perfusion MRI examination procedures in tumor patients begin with examination preparation, preparation of tools and materials, carry out the examination using the MRI Brain protocol with contrast and without contrast. The use of contrast uses the Dynamic Susceptibility Contrast (DSC) method and non-contrast use using the Arterial Spin Labeling (ASL) method where both complement each other to inform perfusion parameter values.&#x0D; Conclusion : The conclution of this study indicate still needed the application of both MR perfusion methods, namely the ASL method and the DSC method. The ASL method will provide additional information in the form of absolute values of Cerebral Blood Flow that are not affected by permeability, besides that the DSC method is able to provide Cerebral Blood Volume values that are used as a parameter of Perfusion MR evaluation.&#x0D; Keyword : ASL, DSC, Tumor, CBF&#x0D;