Artificial intelligence-based locoregional markers of brain peritumoral microenvironment.

In malignant primary brain tumors, cancer cells infiltrate into the peritumoral brain structures which results in inevitable recurrence. Quantitative assessment of infiltrative heterogeneity in the peritumoral region, the area where biopsy or resection can be hazardous, is important for clinical decision making. Here, we derive a novel set of Artificial intelligence (AI)-based markers capturing the heterogeneity of tumor infiltration, by characterizing free water movement restriction in the peritumoral region using Diffusion Tensor Imaging (DTI)-based free water volume fraction maps. We leverage the differences in the peritumoral region of metastasis and glioblastomas, the former consisting of vasogenic versus the latter containing infiltrative edema, to extract a voxel-wise deep learning-based peritumoral microenvironment index (PMI). Descriptive characteristics of locoregional hubs of uniformly high PMI values are then extracted as AI-based markers to capture distinct aspects of infiltrative heterogeneity. The proposed markers are utilized to stratify patients' survival and IDH1 mutation status on a population of 275 adult-type diffuse gliomas (CNS WHO grade 4). Our results show significant differences in the proposed markers between patients with different overall survival and IDH1 mutation status (t test, Wilcoxon rank sum test, linear regression; p < 0.01). Clustering of patients using the proposed markers reveals distinct survival groups (logrank; p < 10-5, Cox hazard ratio = 1.82; p < 0.005). Our findings provide a panel of markers as surrogates of infiltration that might capture novel insight about underlying biology of peritumoral microstructural heterogeneity, providing potential biomarkers of prognosis pertaining to survival and molecular stratification, with applicability in clinical decision making.

The University of Pennsylvania glioblastoma (UPenn-GBM) cohort: advanced MRI, clinical, genomics, & radiomics.

Glioblastoma is the most common aggressive adult brain tumor. Numerous studies have reported results from either private institutional data or publicly available datasets. However, current public datasets are limited in terms of: a) number of subjects, b) lack of consistent acquisition protocol, c) data quality, or d) accompanying clinical, demographic, and molecular information. Toward alleviating these limitations, we contribute the "University of Pennsylvania Glioblastoma Imaging, Genomics, and Radiomics" (UPenn-GBM) dataset, which describes the currently largest publicly available comprehensive collection of 630 patients diagnosed with de novo glioblastoma. The UPenn-GBM dataset includes (a) advanced multi-parametric magnetic resonance imaging scans acquired during routine clinical practice, at the University of Pennsylvania Health System, (b) accompanying clinical, demographic, and molecular information, (d) perfusion and diffusion derivative volumes, (e) computationally-derived and manually-revised expert annotations of tumor sub-regions, as well as (f) quantitative imaging (also known as radiomic) features corresponding to each of these regions. This collection describes our contribution towards repeatable, reproducible, and comparative quantitative studies leading to new predictive, prognostic, and diagnostic assessments.

Toxicity after AAV delivery of RNAi expression constructs into nonhuman primate brain.

RNA interference (RNAi) for spinocerebellar ataxia type 1 can prevent and reverse behavioral deficits and neuropathological readouts in mouse models, with safety and benefit lasting over many months. The RNAi trigger, expressed from adeno-associated virus vectors (AAV.miS1), also corrected misregulated microRNAs (miRNA) such as miR150. Subsequently, we showed that the delivery method was scalable, and that AAV.miS1 was safe in short-term pilot nonhuman primate (NHP) studies. To advance the technology to patients, investigational new drug (IND)-enabling studies in NHPs were initiated. After AAV.miS1 delivery to deep cerebellar nuclei, we unexpectedly observed cerebellar toxicity. Both small-RNA-seq and studies using AAVs devoid of miRNAs showed that this was not a result of saturation of the endogenous miRNA processing machinery. RNA-seq together with sequencing of the AAV product showed that, despite limited amounts of cross-packaged material, there was substantial inverted terminal repeat (ITR) promoter activity that correlated with neuropathologies. ITR promoter activity was reduced by altering the miS1 expression context. The surprising contrast between our rodent and NHP findings highlight the need for extended safety studies in multiple species when assessing new therapeutics for human application.

Enhanced Fiber Tractography Using Edema Correction: Application and Evaluation in High-Grade Gliomas.

BACKGROUND: A limitation of diffusion tensor imaging (DTI)-based tractography is peritumoral edema that confounds traditional diffusion-based magnetic resonance metrics. OBJECTIVE: To augment fiber-tracking through peritumoral regions by performing novel edema correction on clinically feasible DTI acquisitions and assess the accuracy of the fiber-tracks using intraoperative stimulation mapping (ISM), task-based functional magnetic resonance imaging (fMRI) activation maps, and postoperative follow-up as reference standards. METHODS: Edema correction, using our bi-compartment free water modeling algorithm (FERNET), was performed on clinically acquired DTI data from a cohort of 10 patients presenting with suspected high-grade glioma and peritumoral edema in proximity to and/or infiltrating language or motor pathways. Deterministic fiber-tracking was then performed on the corrected and uncorrected DTI to identify tracts pertaining to the eloquent region involved (language or motor). Tracking results were compared visually and quantitatively using mean fiber count, voxel count, and mean fiber length. The tracts through the edematous region were verified based on overlay with the corresponding motor or language task-based fMRI activation maps and intraoperative ISM points, as well as at time points after surgery when peritumoral edema had subsided. RESULTS: Volume and number of fibers increased with application of edema correction; concordantly, mean fractional anisotropy decreased. Overlay with functional activation maps and ISM-verified eloquence of the increased fibers. Comparison with postsurgical follow-up scans with lower edema further confirmed the accuracy of the tracts. CONCLUSION: This method of edema correction can be applied to standard clinical DTI to improve visualization of motor and language tracts in patients with glioma-associated peritumoral edema.

Emerging Techniques in Imaging of Glioma Microenvironment.

Magnetic resonance imaging (MRI) has been the cornerstone of imaging of brain tumors in the past 4 decades. Conventional MRI remains the workhorse for neuro-oncologic imaging, not only for basic information such as location, extent, and navigation but also able to provide information regarding proliferation and infiltration, angiogenesis, hemorrhage, and more. More sophisticated MRI sequences have extended the ability to assess and quantify these features; for example, permeability and perfusion acquisitions can assess blood-brain barrier disruption and angiogenesis, diffusion techniques can assess cellularity and infiltration, and spectroscopy can address metabolism. Techniques such as fMRI and diffusion fiber tracking can be helpful in diagnostic planning for resection and radiation therapy, and more sophisticated iterations of these techniques can extend our understanding of neurocognitive effects of these tumors and associated treatment responses and effects. More recently, MRI has been used to go beyond such morphological, physiological, and functional characteristics to assess the tumor microenvironment. The current review highlights multiple recent and emerging approaches in MRI to characterize the tumor microenvironment.

Histopathology-validated machine learning radiographic biomarker for noninvasive discrimination between true progression and pseudo-progression in glioblastoma.

BACKGROUND: Imaging of glioblastoma patients after maximal safe resection and chemoradiation commonly demonstrates new enhancements that raise concerns about tumor progression. However, in 30% to 50% of patients, these enhancements primarily represent the effects of treatment, or pseudo-progression (PsP). We hypothesize that quantitative machine learning analysis of clinically acquired multiparametric magnetic resonance imaging (mpMRI) can identify subvisual imaging characteristics to provide robust, noninvasive imaging signatures that can distinguish true progression (TP) from PsP. METHODS: We evaluated independent discovery (n = 40) and replication (n = 23) cohorts of glioblastoma patients who underwent second resection due to progressive radiographic changes suspicious for recurrence. Deep learning and conventional feature extraction methods were used to extract quantitative characteristics from the mpMRI scans. Multivariate analysis of these features revealed radiophenotypic signatures distinguishing among TP, PsP, and mixed response that compared with similar categories blindly defined by board-certified neuropathologists. Additionally, interinstitutional validation was performed on 20 new patients. RESULTS: Patients who demonstrate TP on neuropathology are significantly different (P < .0001) from those with PsP, showing imaging features reflecting higher angiogenesis, higher cellularity, and lower water concentration. The accuracy of the proposed signature in leave-one-out cross-validation was 87% for predicting PsP (area under the curve [AUC], 0.92) and 84% for predicting TP (AUC, 0.83), whereas in the discovery/replication cohort, the accuracy was 87% for predicting PsP (AUC, 0.84) and 78% for TP (AUC, 0.80). The accuracy in the interinstitutional cohort was 75% (AUC, 0.80). CONCLUSION: Quantitative mpMRI analysis via machine learning reveals distinctive noninvasive signatures of TP versus PsP after treatment of glioblastoma. Integration of the proposed method into clinical studies can be performed using the freely available Cancer Imaging Phenomics Toolkit.

Arterial Spin Labeling and Dynamic Susceptibility Contrast-enhanced MR Imaging for evaluation of arteriovenous shunting and tumor hypoxia in glioblastoma.

Glioblastoma (GBM) is the most common primary malignant brain tumor in adults and carries a dismal prognosis. Significant challenges in the care of patients with GBM include marked vascular heterogeneity and arteriovenous (AV) shunting, which results in tumor hypoxia and inadequate delivery of systemic treatments to reach tumor cells. In this study, we investigated the utility of different MR perfusion techniques to detect and quantify arteriovenous (AV) shunting and tumor hypoxia in patients with GBM. Macrovascular shunting was present in 33% of subjects, with the degree of shunting ranging from (37-60%) using arterial spin labeling perfusion. Among the dynamic susceptibility contrast-enhanced perfusion curve features, there were a strong negative correlation between hypoxia score, DSC perfusion curve recovery slope (r = -0.72, P = 0.018) and angle (r = -0.73, P = 0.015). The results of this study support the possibility of using arterial spin labeling and pattern analysis of dynamic susceptibility contrast-enhanced MR Imaging for evaluation of arteriovenous shunting and tumor hypoxia in glioblastoma.

Focused Ultrasound Thalamotomy for Essential Tremor in the Setting of a Ventricular Shunt: Technical Report.

BACKGROUND: A recent randomized controlled trial of magnetic resonance imaging (MRI)-guided focused ultrasound (FUS) for essential tremor (ET) demonstrated safety and efficacy. Patients with ventricular shunts may be good candidates for FUS to minimize hardware-associated infections. OBJECTIVE: To demonstrate feasibility of FUS in this subset of patients. METHODS: A 74-yr-old male with medically refractory ET, and a right-sided ventricular shunt for normal pressure hydrocephalus, underwent FUS to the right ventro-intermedius (VIM) nucleus. The VIM nucleus was directly targeted using deterministic tractography. Clinical outcomes were measured using the Clinical Rating Scale for Tremor. RESULTS: Shunt components required 6% of the total ultrasound transducer elements to be shut off. Eight therapeutic sonications were delivered (maximum temperature, 64°), leading to a 90% improvement in hand tremor and a 100% improvement in functional disability at the 3-mo follow-up. No complications were noted. CONCLUSION: This is the first case of FUS thalamotomy in a patient with a shunt. Direct VIM targeting and achievement of therapeutic temperatures with acoustic energy is feasible in this subset of patients.

Differentiation of brain infection from necrotic glioblastoma using combined analysis of diffusion and perfusion MRI.

BACKGROUND: Accurate differentiation of brain infections from necrotic glioblastomas (GBMs) may not always be possible on morphologic MRI or on diffusion tensor imaging (DTI) and dynamic susceptibility contrast perfusion-weighted imaging (DSC-PWI) if these techniques are used independently. PURPOSE: To investigate the combined analysis of DTI and DSC-PWI in distinguishing brain injections from necrotic GBMs. STUDY TYPE: Retrospective. POPULATION: Fourteen patients with brain infections and 21 patients with necrotic GBMs. FIELD STRENGTH/SEQUENCE: 3T MRI, DTI, and DSC-PWI. ASSESSMENT: Parametric maps of mean diffusivity (MD), fractional anisotropy (FA), coefficient of linear (CL), and planar anisotropy (CP) and leakage corrected cerebral blood volume (CBV) were computed and coregistered with postcontrast T1 -weighted and FLAIR images. All lesions were segmented into the central core and enhancing region. For each region, median values of MD, FA, CL, CP, relative CBV (rCBV), and top 90th percentile of rCBV (rCBVmax ) were measured. STATISTICAL TESTS: All parameters from both regions were compared between brain infections and necrotic GBMs using Mann-Whitney tests. Logistic regression analyses were performed to obtain the best model in distinguishing these two conditions. RESULTS: From the central core, significantly lower MD (0.90 × 10-3 ± 0.44 × 10-3 mm2 /s vs. 1.66 × 10-3 ± 0.62 × 10-3 mm2 /s, P = 0.001), significantly higher FA (0.15 ± 0.06 vs. 0.09 ± 0.03, P < 0.001), and CP (0.07 ± 0.03 vs. 0.04 ± 0.02, P = 0.009) were observed in brain infections compared to those in necrotic GBMs. Additionally, from the contrast-enhancing region, significantly lower rCBV (1.91 ± 0.95 vs. 2.76 ± 1.24, P = 0.031) and rCBVmax (3.46 ± 1.41 vs. 5.89 ± 2.06, P = 0.001) were observed from infective lesions compared to necrotic GBMs. FA from the central core and rCBVmax from enhancing region provided the best classification model in distinguishing brain infections from necrotic GBMs, with a sensitivity of 91% and a specificity of 93%. DATA CONCLUSION: Combined analysis of DTI and DSC-PWI may provide better performance in differentiating brain infections from necrotic GBMs. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;49:184-194.

Multiparametric magnetic resonance imaging in the assessment of anti-EGFRvIII chimeric antigen receptor T cell therapy in patients with recurrent glioblastoma.

EGFRvIII targeted chimeric antigen receptor T (CAR-T) cell therapy has recently been reported for treating glioblastomas (GBMs); however, physiology-based MRI parameters have not been evaluated in this setting. Ten patients underwent multiparametric MRI at baseline, 1, 2 and 3 months after CAR-T therapy. Logistic regression model derived progression probabilities (PP) using imaging parameters were used to assess treatment response. Four lesions from "early surgery" group demonstrated high PP at baseline suggestive of progression, which was confirmed histologically. Out of eight lesions from remaining six patients, three lesions with low PP at baseline remained stable. Two lesions with high PP at baseline were associated with large decreases in PP reflecting treatment response, whereas other two lesions with high PP at baseline continued to demonstrate progression. One patient didn't have baseline data but demonstrated progression on follow-up. Our findings indicate that multiparametric MRI may be helpful in monitoring CAR-T related early therapeutic changes in GBM patients.

Three-dimensional echo planar spectroscopic imaging for differentiation of true progression from pseudoprogression in patients with glioblastoma.

Accurate differentiation of true progression (TP) from pseudoprogression (PsP) in patients with glioblastomas (GBMs) is essential for planning adequate treatment and for estimating clinical outcome measures and future prognosis. The purpose of this study was to investigate the utility of three-dimensional echo planar spectroscopic imaging (3D-EPSI) in distinguishing TP from PsP in GBM patients. For this institutional review board approved and HIPAA compliant retrospective study, 27 patients with GBM demonstrating enhancing lesions within six months of completion of concurrent chemo-radiation therapy were included. Of these, 18 were subsequently classified as TP and 9 as PsP based on histological features or follow-up MRI studies. Parametric maps of choline/creatine (Cho/Cr) and choline/N-acetylaspartate (Cho/NAA) were computed and co-registered with post-contrast T1 -weighted and FLAIR images. All lesions were segmented into contrast enhancing (CER), immediate peritumoral (IPR), and distal peritumoral (DPR) regions. For each region, Cho/Cr and Cho/NAA ratios were normalized to corresponding metabolite ratios from contralateral normal parenchyma and compared between TP and PsP groups. Logistic regression analyses were performed to obtain the best model to distinguish TP from PsP. Significantly higher Cho/NAA was observed from CER (2.69 ± 1.00 versus 1.56 ± 0.51, p = 0.003), IPR (2.31 ± 0.92 versus 1.53 ± 0.56, p = 0.030), and DPR (1.80 ± 0.68 versus 1.19 ± 0.28, p = 0.035) regions in TP patients compared with those with PsP. Additionally, significantly elevated Cho/Cr (1.74 ± 0.44 versus 1.34 ± 0.26, p = 0.023) from CER was observed in TP compared with PsP. When these parameters were incorporated in multivariate regression analyses, a discriminatory model with a sensitivity of 94% and a specificity of 87% was observed in distinguishing TP from PsP. These results indicate the utility of 3D-EPSI in differentiating TP from PsP with high sensitivity and specificity.

Prognostic value of contrast enhancement and FLAIR for survival in newly diagnosed glioblastoma treated with and without bevacizumab: results from ACRIN 6686.

Background: ACRIN 6686/RTOG 0825 was a phase III trial of conventional chemoradiation plus adjuvant temozolomide with bevacizumab or without (placebo) in newly diagnosed glioblastoma. This study investigated whether changes in contrast-enhancing and fluid attenuated inversion recovery (FLAIR)-hyperintense tumor assessed by central reading prognosticate overall survival (OS). Methods: Two hundred eighty-four patients (171 men; median age 57 y, range 19-79; 159 on bevacizumab) had MRI at post-op (baseline) and pre-cycle 4 of adjuvant temozolomide (22 wk post chemoradiation initiation). Four central readers measured bidimensional lesion enhancement (2D-T1) and FLAIR hyperintensity at both time points. Changes from baseline to pre-cycle 4 for both markers were dichotomized (increasing vs non-increasing). Cox proportional hazards model and Kaplan-Meier survival estimates were used for inference. Results: Adjusting for treatment, increasing 2D-T1 (n = 262, hazard ratio [HR] = 2.07, 95% CI: 1.48-2.91, P < 0.0001) and FLAIR (n = 273, HR = 1.75, 95% CI: 1.26-2.41, P = 0.0008) significantly predicted worse OS. Median OS (days) was significantly shorter for patients with increasing versus non-increasing 2D-T1 for both bevacizumab (443 vs 535, P = 0.004) and placebo (526 vs 887, P = 0.001). Median OS was significantly shorter for patients with increasing versus non-increasing FLAIR for placebo (595 vs 872, P = 0.001), and trended similarly for bevacizumab (499 vs 535, P = 0.0935). Adjusting for 2D-T1 and treatment, increasing FLAIR represented significantly higher risk for death (HR = 1.59 [1.11-2.26], P = 0.01). Conclusion: Increased 2D-T1 significantly predicts worse OS in both treatment groups, implying absence of a substantial proportion of pseudoprogression 22 weeks after initiation of standard therapy. FLAIR adds value beyond 2D-T1 in predicting OS, potentially addressing the pseudoresponse effect by substratifying bevacizumab-treated patients with non-increasing 2D-T1.

Population-based MRI atlases of spatial distribution are specific to patient and tumor characteristics in glioblastoma.

BACKGROUND AND PURPOSE: In treating glioblastoma (GB), surgical and chemotherapeutic treatment guidelines are, for the most part, independent of tumor location. In this work, we compiled imaging data from a large cohort of GB patients to create statistical atlases illustrating the disease spatial frequency as a function of patient demographics as well as tumor characteristics. MATERIALS AND METHODS: Two-hundred-six patients with pathology-proven glioblastoma were included. Of those, 65 had pathology-proven recurrence and 113 had molecular subtype and genetic information. We used validated software to segment the tumors in all patients and map them from patient space into a common template. We then created statistical maps that described the spatial location of tumors with respect to demographics and tumor characteristics. We applied a chi-square test to determine whether pattern differences were statistically significant. RESULTS: The most frequent location for glioblastoma in our patient population is the right temporal lobe. There are statistically significant differences when comparing patterns using demographic data such as gender (p = 0.0006) and age (p = 0.006). Small and large tumors tend to occur in separate locations (p = 0.0007). The tumors tend to occur in different locations according to their molecular subtypes (p < 10(- 6)). The classical subtype tends to spare the frontal lobes, the neural subtype tend to involve the inferior right frontal lobe. Although the sample size is limited, there was a difference in location according to EGFR VIII genotype (p < 10(- 4)), with a right temporal dominance for EFGR VIII negative tumors, and frontal lobe dominance in EGFR VIII positive tumors. CONCLUSIONS: Spatial location of GB is an important factor that correlates with demographic factors and tumor characteristics, which should therefore be considered when evaluating a patient with GB and might assist in personalized treatment.

Imaging Surrogates of Infiltration Obtained Via Multiparametric Imaging Pattern Analysis Predict Subsequent Location of Recurrence of Glioblastoma.

BACKGROUND: Glioblastoma is an aggressive and highly infiltrative brain cancer. Standard surgical resection is guided by enhancement on postcontrast T1-weighted (T1) magnetic resonance imaging, which is insufficient for delineating surrounding infiltrating tumor. OBJECTIVE: To develop imaging biomarkers that delineate areas of tumor infiltration and predict early recurrence in peritumoral tissue. Such markers would enable intensive, yet targeted, surgery and radiotherapy, thereby potentially delaying recurrence and prolonging survival. METHODS: Preoperative multiparametric magnetic resonance images (T1, T1-gadolinium, T2-weighted, T2-weighted fluid-attenuated inversion recovery, diffusion tensor imaging, and dynamic susceptibility contrast-enhanced magnetic resonance images) from 31 patients were combined using machine learning methods, thereby creating predictive spatial maps of infiltrated peritumoral tissue. Cross-validation was used in the retrospective cohort to achieve generalizable biomarkers. Subsequently, the imaging signatures learned from the retrospective study were used in a replication cohort of 34 new patients. Spatial maps representing the likelihood of tumor infiltration and future early recurrence were compared with regions of recurrence on postresection follow-up studies with pathology confirmation. RESULTS: This technique produced predictions of early recurrence with a mean area under the curve of 0.84, sensitivity of 91%, specificity of 93%, and odds ratio estimates of 9.29 (99% confidence interval: 8.95-9.65) for tissue predicted to be heavily infiltrated in the replication study. Regions of tumor recurrence were found to have subtle, yet fairly distinctive multiparametric imaging signatures when analyzed quantitatively by pattern analysis and machine learning. CONCLUSION: Visually imperceptible imaging patterns discovered via multiparametric pattern analysis methods were found to estimate the extent of infiltration and location of future tumor recurrence, paving the way for improved targeted treatment.

Individualized Map of White Matter Pathways: Connectivity-Based Paradigm for Neurosurgical Planning.

BACKGROUND: Advances in white matter tractography enhance neurosurgical planning and glioma resection, but white matter tractography is limited by biological variables such as edema, mass effect, and tract infiltration or selection biases related to regions of interest or fractional anisotropy values. OBJECTIVE: To provide an automated tract identification paradigm that corrects for artifacts created by tumor edema and infiltration and provides a consistent, accurate method of fiber bundle identification. METHODS: An automated tract identification paradigm was developed and evaluated for glioma surgery. A fiber bundle atlas was generated from 6 healthy participants. Fibers of a test set (including 3 healthy participants and 10 patients with brain tumors) were clustered adaptively with this atlas. Reliability of the identified tracts in both groups was assessed by comparison with 2 experts with the Cohen κ used to quantify concurrence. We evaluated 6 major fiber bundles: cingulum bundle, fornix, uncinate fasciculus, arcuate fasciculus, inferior fronto-occipital fasciculus, and inferior longitudinal fasciculus, the last 3 tracts mediating language function. RESULTS: The automated paradigm demonstrated a reliable and practical method to identify white mater tracts, despite mass effect, edema, and tract infiltration. When the tumor demonstrated significant mass effect or shift, the automated approach was useful for providing an initialization to guide the expert with identification of the specific tract of interest. CONCLUSION: We report a reliable paradigm for the automated identification of white matter pathways in patients with gliomas. This approach should enhance the neurosurgical objective of maximal safe resections. ABBREVIATIONS: AF, arcuate fasciculusDTI, diffusion tensor imagingIFOF, inferior fronto-occipital fasciculusILF, inferior longitudinal fasciculusROI, region of interestWM, white matter.

Imaging patterns predict patient survival and molecular subtype in glioblastoma via machine learning techniques.

BACKGROUND: MRI characteristics of brain gliomas have been used to predict clinical outcome and molecular tumor characteristics. However, previously reported imaging biomarkers have not been sufficiently accurate or reproducible to enter routine clinical practice and often rely on relatively simple MRI measures. The current study leverages advanced image analysis and machine learning algorithms to identify complex and reproducible imaging patterns predictive of overall survival and molecular subtype in glioblastoma (GB). METHODS: One hundred five patients with GB were first used to extract approximately 60 diverse features from preoperative multiparametric MRIs. These imaging features were used by a machine learning algorithm to derive imaging predictors of patient survival and molecular subtype. Cross-validation ensured generalizability of these predictors to new patients. Subsequently, the predictors were evaluated in a prospective cohort of 29 new patients. RESULTS: Survival curves yielded a hazard ratio of 10.64 for predicted long versus short survivors. The overall, 3-way (long/medium/short survival) accuracy in the prospective cohort approached 80%. Classification of patients into the 4 molecular subtypes of GB achieved 76% accuracy. CONCLUSIONS: By employing machine learning techniques, we were able to demonstrate that imaging patterns are highly predictive of patient survival. Additionally, we found that GB subtypes have distinctive imaging phenotypes. These results reveal that when imaging markers related to infiltration, cell density, microvascularity, and blood-brain barrier compromise are integrated via advanced pattern analysis methods, they form very accurate predictive biomarkers. These predictive markers used solely preoperative images, hence they can significantly augment diagnosis and treatment of GB patients.

Microvesicles as a Biomarker for Tumor Progression versus Treatment Effect in Radiation/Temozolomide-Treated Glioblastoma Patients.

UNLABELLED: The standard of care for glioblastoma (GB) is surgery followed by concurrent radiation therapy (RT) and temozolomide (TMZ) and then adjuvant TMZ. This regime is associated with increased survival but also increased occurrence of equivocal imaging findings, e.g., tumor progression (TP) versus treatment effect (TE), which is also referred to as pseudoprogression (PsP). Equivocal findings make decisions regarding further treatment difficult and often delayed. Because none of the current imaging assays have proven sensitive and specific for differentiation of TP versus TE/PsP, we investigated whether blood-derived microvesicles (MVs) would be a relevant assay. METHODS: 2.8 ml of citrated blood was collected from patients with GB at the time of their RT simulation, at the end of chemoradiation therapy (CRT), and multiple times following treatment. MVs were collected following multiple centrifugations (300g, 2500g, and 15,000g). The pellet from the final spin was analyzed using flow cytometry. A diameter of approximately 300 nm or greater and Pacific Blue-labeled Annexin V positivity were used to identify the MVs reported herein. RESULTS: We analyzed 19 blood samples from 11 patients with GB. MV counts in the patients with stable disease or TE/PsP were significantly lower than patients who developed TP (P = .014). CONCLUSION: These preliminary data suggest that blood analysis for MVs from GB patients receiving CRT may be useful to distinguish TE/PsP from TP. MVs may add clarity to standard imaging for decision making in patients with equivocal imaging findings.

Pattern analysis of dynamic susceptibility contrast-enhanced MR imaging demonstrates peritumoral tissue heterogeneity.

PURPOSE: To augment the analysis of dynamic susceptibility contrast material-enhanced magnetic resonance (MR) images to uncover unique tissue characteristics that could potentially facilitate treatment planning through a better understanding of the peritumoral region in patients with glioblastoma. MATERIALS AND METHODS: Institutional review board approval was obtained for this study, with waiver of informed consent for retrospective review of medical records. Dynamic susceptibility contrast-enhanced MR imaging data were obtained for 79 patients, and principal component analysis was applied to the perfusion signal intensity. The first six principal components were sufficient to characterize more than 99% of variance in the temporal dynamics of blood perfusion in all regions of interest. The principal components were subsequently used in conjunction with a support vector machine classifier to create a map of heterogeneity within the peritumoral region, and the variance of this map served as the heterogeneity score. RESULTS: The calculated principal components allowed near-perfect separability of tissue that was likely highly infiltrated with tumor and tissue that was unlikely infiltrated with tumor. The heterogeneity map created by using the principal components showed a clear relationship between voxels judged by the support vector machine to be highly infiltrated and subsequent recurrence. The results demonstrated a significant correlation (r = 0.46, P < .0001) between the heterogeneity score and patient survival. The hazard ratio was 2.23 (95% confidence interval: 1.4, 3.6; P < .01) between patients with high and low heterogeneity scores on the basis of the median heterogeneity score. CONCLUSION: Analysis of dynamic susceptibility contrast-enhanced MR imaging data by using principal component analysis can help identify imaging variables that can be subsequently used to evaluate the peritumoral region in glioblastoma. These variables are potentially indicative of tumor infiltration and may become useful tools in guiding therapy, as well as individualized prognostication.