Search

Federated ML (FL) provides an alternative to train accurate and generalizable ML models, by only sharing numerical model updates. Here, the authors present the largest FL study to-date to generate an automatic tumor boundary detector for glioblastoma.

This pilot study investigates structural alterations and their relationships with cognitive function in survivors of diffuse gliomas. Twenty-four survivors of diffuse gliomas (mean age 44.5 ± 11.5), from whom high-resolution T1-weighted images, neuropsychological tests, and self-report questionnaires were obtained, were analyzed. Patients were grouped by degree of cognitive impairment, and interregional correlations of cortical thickness were computed to generate morphometric correlation networks (MCNs). The results show that the cortical thickness of the right insula (R2 = 0.3025, p = 0.0054) was negatively associated with time since the last treatment, and the cortical thickness of the left superior temporal gyrus (R2 = 0.2839, p = 0.0107) was positively associated with cognitive performance. Multiple cortical regions in the default mode, salience, and language networks were identified as predominant nodes in the MCNs of survivors of diffuse gliomas. Compared to cognitively impaired patients, cognitively non-impaired patients tended to have higher network stability in network nodes removal analysis, especially when the fraction of removed nodes (among 66 nodes in total) exceeded 55%. These findings suggest that structural networks are altered in survivors of diffuse gliomas and that their cortical structures may also be adapting to support cognitive function during survivorship.

Abstract This study aimed to differentiate isocitrate dehydrogenase (IDH) mutation status with the voxel-wise clustering method of multiparametric magnetic resonance imaging (MRI) and to discover biological underpinnings of the clusters. A total of 69 patients with treatment-naïve diffuse glioma were scanned with pH-sensitive amine chemical exchange saturation transfer MRI, diffusion-weighted imaging, fluid-attenuated inversion recovery, and contrast-enhanced T1-weighted imaging at 3 T. An unsupervised two-level clustering approach was used for feature extraction from acquired images. The logarithmic ratio of the labels in each class within tumor regions was applied to a support vector machine to differentiate IDH status. The highest performance to predict IDH mutation status was found for 10-class clustering, with a mean area under the curve, accuracy, sensitivity, and specificity of 0.94, 0.91, 0.90, and 0.91, respectively. Targeted biopsies revealed that the tissues with labels 7–10 showed high expression levels of hypoxia-inducible factor 1-alpha, glucose transporter 3, and hexokinase 2, which are typical of IDH wild-type glioma, whereas those with labels 1 showed low expression of these proteins. In conclusion, A machine learning model successfully predicted the IDH mutation status of gliomas, and the resulting clusters properly reflected the metabolic status of the tumors.

Abstract Glioblastoma (GBM) has high metabolic demands, which can lead to acidification of the tumor microenvironment. We hypothesize that a machine learning model built on temporal principal component analysis (PCA) of dynamic susceptibility contrast-enhanced (DSC) perfusion MRI can be used to estimate tumor acidity in GBM, as estimated by pH-sensitive amine chemical exchange saturation transfer echo-planar imaging (CEST-EPI). We analyzed 78 MRI scans in 32 treatment naïve and post-treatment GBM patients. All patients were imaged with DSC-MRI, and pH-weighting that was quantified from CEST-EPI estimation of the magnetization transfer ratio asymmetry (MTRasym) at 3 ppm. Enhancing tumor (ET), non-enhancing core (NC), and peritumoral T2 hyperintensity (namely, edema, ED) were used to extract principal components (PCs) and to build support vector machines regression (SVR) models to predict MTRasym values using PCs. Our predicted map correlated with MTRasym values with Spearman’s r equal to 0.66, 0.47, 0.67, 0.71, in NC, ET, ED, and overall, respectively (p

Aim: Understanding and supporting quality of life (QoL) and daily functioning in glioma patients is a clinical imperative. In this study, we examined the relationship between cognition, psychological factors, measures of health-related QoL and functioning in glioma survivors. Materials & methods: We examined neuropsychological, self-reported cognition, mood and QoL correlates of work and non-work-related daily functioning in 23 glioma survivors, and carried out linear models of the best predictors. Results & conclusion: A total of 13/23 participants were working at the time of enrollment. The best model for worse work-related functioning (R2 = .83) included worse self-reported cognitive function, depression, loneliness and brain tumor symptoms. The best model for worse non-work-related functioning (R2 = .61) included worse self-reported cognitive functioning, anxiety, sleep disturbance and physical functioning. Neuropsychological variables were not among the most highly correlated with function. Worse cognitive, particularly self-reported and psychosocial outcomes may compromise optimal functioning in glioma survivors.

Abstract Background The purpose of this study was to develop a voxel-wise clustering method of multiparametric magnetic resonance imaging (MRI) and 3,4-dihydroxy-6-[18F]-fluoro-L-phenylalanine (FDOPA) positron emission tomography (PET) images using an unsupervised, two-level clustering approach followed by support vector machine in order to classify the isocitrate dehydrogenase (IDH) status of gliomas. Methods Sixty-two treatment-naïve glioma patients who underwent FDOPA PET and MRI were retrospectively included. Contrast enhanced T1-weighted images, T2-weighted images, fluid-attenuated inversion recovery images, apparent diffusion coefficient maps, and relative cerebral blood volume maps, and FDOPA PET images were used for voxel-wise feature extraction. An unsupervised two-level clustering approach, including a self-organizing map followed by the K-means algorithm was used, and each class label was applied to the original images. The logarithmic ratio of labels in each class within tumor regions was applied to a support vector machine to differentiate IDH mutation status. The area under the curve (AUC) of receiver operating characteristic curves, accuracy, and F1-socore were calculated and used as metrics for performance. Results The associations of multiparametric imaging values in each cluster were successfully visualized. Multiparametric images with 16-class clustering revealed the highest classification performance to differentiate IDH status with the AUC, accuracy, and F1-score of 0.81, 0.76, and 0.76, respectively. Conclusions Machine learning using an unsupervised two-level clustering approach followed by a support vector machine classified the IDH mutation status of gliomas, and visualized voxel-wise features from multiparametric MRI and FDOPA PET images. Unsupervised clustered features may improve the understanding of prioritizing multiparametric imaging for classifying IDH status.

Background and PurposeWhile relative cerebral blood volume (rCBV) may be diagnostic and prognostic for survival in glioblastoma (GBM), changes in rCBV during chemoradiation in the subset of newly diagnosed GBM with subtotal resection and the impact of MGMT promoter methylation status on survival have not been explored. This study aimed to investigate the association between rCBV response, MGMT methylation status, and progression-free (PFS) and overall survival (OS) in newly diagnosed GBM with measurable enhancing lesions.Methods1,153 newly diagnosed IDH wild-type GBM patients were screened and 53 patients (4.6%) had measurable post-surgical tumor (>1mL). rCBV was measured before and after patients underwent chemoradiation. Patients with a decrease in rCBV >10% were considered rCBV Responders, while patients with an increase or a decrease in rCBV

Abstract Purpose To identify clinically relevant magnetic resonance imaging (MRI) features of different types of metastatic brain lesions, including standard anatomical, diffusion weighted imaging (DWI) and dynamic susceptibility contrast (DSC) perfusion MRI. Methods MRI imaging was retrospectively assessed on one hundred and fourteen (N = 114) brain metastases including breast (n = 27), non-small cell lung cancer (NSCLC, n = 43) and ‘other’ primary tumors (n = 44). Based on 114 patient’s MRI scans, a total of 346 individual contrast enhancing tumors were manually segmented. In addition to tumor volume, apparent diffusion coefficients (ADC) and relative cerebral blood volume (rCBV) measurements, an independent component analysis (ICA) was performed with raw DSC data in order to assess arterio-venous components and the volume of overlap (AVOL) relative to tumor volume, as well as time to peak (TTP) of T2* signal from each component. Results Results suggests non-breast or non-NSCLC (‘other’) tumors had higher volume compare to breast and NSCLC patients (p = 0.0056 and p = 0.0003, respectively). No differences in median ADC or rCBV were observed across tumor types; however, breast and NSCLC tumors had a significantly higher “arterial” proportion of the tumor volume as indicated by ICA (p = 0.0062 and p = 0.0018, respectively), while a higher “venous” proportion were prominent in breast tumors compared with NSCLC (p = 0.0027) and ‘other’ lesions (p = 0.0011). The AVOL component was positively related to rCBV in all groups, but no correlation was found for arterial and venous components with respect to rCBV values. Median time to peak of arterial and venous components were 8.4 s and 12.6 s, respectively (p

Purpose: To quantify abnormal metabolism of diffuse gliomas using “aerobic glycolytic imaging” and investigate its biological correlation. Methods: All subjects underwent a pH-weighted amine chemical exchange saturation transfer spin-and-gradient-echo echoplanar imaging (CEST-SAGE-EPI) and dynamic susceptibility contrast perfusion MRI. Relative oxygen extraction fraction (rOEF) was estimated as the ratio of reversible transverse relaxation rate R2′ to normalized relative cerebral blood volume. An aerobic glycolytic index (AGI) was derived by the ratio of pH-weighted image contrast (MTRasym at 3.0 ppm) to rOEF. AGI was compared between different tumor types (N = 51, 30 IDH mutant and 21 IDH wild type). Metabolic MR parameters were correlated with 18F-FDG uptake (N = 8, IDH wild-type glioblastoma), expression of key glycolytic proteins using immunohistochemistry (N = 38 samples, 21 from IDH mutant and 17 from IDH wild type), and bioenergetics analysis on purified tumor cells (N = 7, IDH wild-type high grade). Results: AGI was significantly lower in IDH mutant than wild-type gliomas (0.48 ± 0.48 vs. 0.70 ± 0.48; P = 0.03). AGI was strongly correlated with 18F-FDG uptake both in non-enhancing tumor (Spearman, ρ = 0.81; P = 0.01) and enhancing tumor (ρ = 0.81; P = 0.01). AGI was significantly correlated with glucose transporter 3 (ρ = 0.71; P = 0.004) and hexokinase 2 (ρ = 0.73; P = 0.003) in IDH wild-type glioma, and monocarboxylate transporter 1 (ρ = 0.59; P = 0.009) in IDH mutant glioma. Additionally, a significant correlation was found between AGI derived from bioenergetics analysis and that estimated from MRI (ρ = 0.79; P = 0.04). Conclusion: AGI derived from molecular MRI was correlated with glucose uptake (18F-FDG and glucose transporter 3/hexokinase 2) and cellular AGI in IDH wild-type gliomas, whereas AGI in IDH mutant gliomas appeared associated with monocarboxylate transporter density.

Cerebral acidosis is a consequence of secondary injury mechanisms following traumatic brain injury (TBI), including excitotoxicity and ischemia, with potentially significant clinical implications. However, there remains an unmet clinical need for technology for non-invasive, high resolution pH imaging of human TBI for studying metabolic changes following injury. The current study examined 17 patients with TBI and 20 healthy controls using amine chemical exchange saturation transfer echoplanar imaging (CEST EPI), a novel pH-weighted molecular MR imaging technique, on a clinical 3T MR scanner. Results showed significantly elevated pH-weighted image contrast (MTRasym at 3 ppm) in areas of T2 hyperintensity or edema (P