VASARI-auto: Equitable, efficient, and economical featurisation of glioma MRI.

The VASARI MRI feature set is a quantitative system designed to standardise glioma imaging descriptions. Though effective, deriving VASARI is time-consuming and seldom used clinically. We sought to resolve this problem with software automation and machine learning. Using glioma data from 1172 patients, we developed VASARI-auto, an automated labelling software applied to open-source lesion masks and an openly available tumour segmentation model. Consultant neuroradiologists independently quantified VASARI features in 100 held-out glioblastoma cases. We quantified 1) agreement across neuroradiologists and VASARI-auto, 2) software equity, 3) an economic workforce analysis, and 4) fidelity in predicting survival. Tumour segmentation was compatible with the current state of the art and equally performant regardless of age or sex. A modest inter-rater variability between in-house neuroradiologists was comparable to between neuroradiologists and VASARI-auto, with far higher agreement between VASARI-auto methods. The time for neuroradiologists to derive VASARI was substantially higher than VASARI-auto (mean time per case 317 vs. 3 s). A UK hospital workforce analysis forecast that three years of VASARI featurisation would demand 29,777 consultant neuroradiologist workforce hours and >£1.5 ($1.9) million, reducible to 332 hours of computing time (and £146 of power) with VASARI-auto. The best-performing survival model utilised VASARI-auto features instead of those derived by neuroradiologists. VASARI-auto is a highly efficient and equitable automated labelling system, a favourable economic profile if used as a decision support tool, and non-inferior survival prediction. Future work should iterate upon and integrate such tools to enhance patient care.

AB066. From imaging to molecular diagnosis: VASARI magnetic resonance imaging features to predict isocitrate dehydrogenase mutation status in glioma.

BACKGROUND: Glioma, the most common brain tumor, poses significant challenges in patient care and economic burden. Clinicians often struggle with management strategies, especially under the 2021 World Health Organization (WHO) central nervous system (CNS) classification emphasizing molecular diagnosis. Isocitrate dehydrogenase (IDH) mutation status is crucial in glioma management. However, many facilities lack the capability for comprehensive molecular tests, and not all patients are candidates for invasive biopsies. MRI offers a non-invasive method to evaluate glioma characteristics. The Visually Accessible Rembrandt Images (VASARI) MRI feature set provides a systematic approach to analyzing brain glioma. This study examines the association of VASARI features with IDH mutation status and their predictive capability. METHODS: This study included 105 glioma patients treated between 2017 and 2022 who had not undergone surgery, chemotherapy, or radiotherapy. Brain MRIs were assessed using VASARI MRI features by two blinded radiologists. Pathological and molecular examinations were conducted per the 2021 WHO CNS tumor classification. IDH mutations were assessed using polymerase chain reaction (PCR) followed by DNA sequencing. Chi-squared analysis identified VASARI features significantly associated with IDH mutation status. A random forest model predicted IDH mutation status using these features. RESULTS: Brain MRI assessments using VASARI terminology showed good inter-observer agreement (kappa =0.714-0.831) and excellent intra-observer agreement (kappa =0.910). Thirteen VASARI features were significantly associated with IDH mutation status. The prediction model based on VASARI MRI features achieved an area under the curve (AUC) of 0.97, with 93.75% sensitivity, 75% specificity, and 84.38% accuracy on test data. CONCLUSIONS: The VASARI MRI feature set is a reliable method for evaluating glioma patients and is feasible for routine radiological practice. Several VASARI features significantly associate with IDH mutation status, aiding glioma patient management. The IDH mutation prediction model based on VASARI features performs excellently and warrants further validation before routine implementation.

Human performance in predicting enhancement quality of gliomas using gadolinium-free MRI sequences.

BACKGROUND AND PURPOSE: To develop and test a decision tree for predicting contrast enhancement quality and shape using precontrast magnetic resonance imaging (MRI) sequences in a large adult-type diffuse glioma cohort. METHODS: Preoperative MRI scans (development/optimization/test sets: n = 31/38/303, male = 17/22/189, mean age = 52/59/56.7 years, high-grade glioma = 22/33/249) were retrospectively evaluated, including pre- and postcontrast T1-weighted, T2-weighted, fluid-attenuated inversion recovery, and diffusion-weighted imaging sequences. Enhancement prediction decision tree (EPDT) was developed using development and optimization sets, incorporating four imaging features: necrosis, diffusion restriction, T2 inhomogeneity, and nonenhancing tumor margins. EPDT accuracy was assessed on a test set by three raters of variable experience. True enhancement features (gold standard) were evaluated using pre- and postcontrast T1-weighted images. Statistical analysis used confusion matrices, Cohen's/Fleiss' kappa, and Kendall's W. Significance threshold was p < .05. RESULTS: Raters 1, 2, and 3 achieved overall accuracies of .86 (95% confidence interval [CI]: .81-.90), .89 (95% CI: .85-.92), and .92 (95% CI: .89-.95), respectively, in predicting enhancement quality (marked, mild, or no enhancement). Regarding shape, defined as the thickness of enhancing margin (solid, rim, or no enhancement), accuracies were .84 (95% CI: .79-.88), .88 (95% CI: .84-.92), and .89 (95% CI: .85-.92). Intrarater intergroup agreement comparing predicted and true enhancement features consistently reached substantial levels (≥.68 [95% CI: .61-.75]). Interrater comparison showed at least moderate agreement (group: ≥.42 [95% CI: .36-.48], pairwise: ≥.61 [95% CI: .50-.72]). Among the imaging features in the EPDT, necrosis assessment displayed the highest intra- and interrater consistency (≥.80 [95% CI: .73-.88]). CONCLUSION: The proposed EPDT has high accuracy in predicting enhancement patterns of gliomas irrespective of rater experience.

Glioma Grade and Molecular Markers: Comparing Machine-Learning Approaches Using VASARI (Visually AcceSAble Rembrandt Images) Radiological Assessment.

OBJECTIVES: This study aimed to leverage Visually AcceSAble Rembrandt Images (VASARI) radiological features, extracted from magnetic resonance imaging (MRI) scans, and machine-learning techniques to predict glioma grade, isocitrate dehydrogenase (IDH) mutation status, and O6-methylguanine-DNA methyltransferase (MGMT) methylation. METHODOLOGY: A retrospective evaluation was undertaken, analyzing MRI and molecular data from 107 glioma patients treated at a tertiary hospital. Patients underwent MRI scans using established protocols and were evaluated based on VASARI criteria. Tissue samples were assessed for glioma grade and underwent molecular testing for IDH mutations and MGMT methylation. Four machine learning models, namely, Random Forest, Elastic-Net, multivariate adaptive regression spline (MARS), and eXtreme Gradient Boosting (XGBoost), were trained on 27 VASARI features using fivefold internal cross-validation. The models' predictive performances were assessed using the area under the curve (AUC), sensitivity, and specificity. RESULTS: For glioma grade prediction, XGBoost exhibited the highest AUC (0.978), sensitivity (0.879), and specificity (0.964), with f6 (proportion of non-enhancing) and f12 (definition of enhancing margin) as the most important predictors. In predicting IDH mutation status, XGBoost achieved an AUC of 0.806, sensitivity of 0.364, and specificity of 0.880, with f1 (tumor location), f12, and f30 (perpendicular diameter to f29) as primary predictors. For MGMT methylation, XGBoost displayed an AUC of 0.580, sensitivity of 0.372, and specificity of 0.759, highlighting f29 (longest diameter) as the key predictor. CONCLUSIONS: This study underscores the robust potential of combining VASARI radiological features with machine learning models in predicting glioma grade, IDH mutation status, and MGMT methylation. The best and most balanced performance was achieved using the XGBoost model. While the prediction of glioma grade showed promising results, the sensitivity in discerning IDH mutations and MGMT methylation still leaves room for improvement. Follow-up studies with larger datasets and more advanced artificial intelligence techniques can further refine our understanding and management of gliomas.

Recursive partitioning analysis for survival stratification and early imaging prediction of molecular biomarker in glioma patients.

BACKGROUND: Glioma is the most common primary brain tumor with high mortality and disability rates. Recent studies have highlighted the significant prognostic consequences of subtyping molecular pathological markers using tumor samples, such as IDH, 1p/19q, and TERT. However, the relative importance of individual markers or marker combinations in affecting patient survival remains unclear. Moreover, the high cost and reliance on postoperative tumor samples hinder the widespread use of these molecular markers in clinical practice, particularly during the preoperative period. We aim to identify the most prominent molecular biomarker combination that affects patient survival and develop a preoperative MRI-based predictive model and clinical scoring system for this combination. METHODS: A cohort dataset of 2,879 patients was compiled for survival risk stratification. In a subset of 238 patients, recursive partitioning analysis (RPA) was applied to create a survival subgroup framework based on molecular markers. We then collected MRI data and applied Visually Accessible Rembrandt Images (VASARI) features to construct predictive models and clinical scoring systems. RESULTS: The RPA delineated four survival groups primarily defined by the status of IDH and TERT mutations. Predictive models incorporating VASARI features and clinical data achieved AUC values of 0.85 for IDH and 0.82 for TERT mutations. Nomogram-based scoring systems were also formulated to facilitate clinical application. CONCLUSIONS: The combination of IDH-TERT mutation status alone can identify the most distinct survival differences in glioma patients. The predictive model based on preoperative MRI features, supported by clinical assessments, offers a reliable method for early molecular mutation prediction and constitutes a valuable scoring tool for clinicians in guiding treatment strategies.

Clinically Available and Reproducible Prediction Models for IDH and CDKN2A/B Gene Status in Adult-type Diffuse Gliomas.

PURPOSE: Isocitrate dehydrogenase (IDH) and cyclin-dependent kinase inhibitor (CDKN) 2A/B status holds important prognostic value in diffuse gliomas. We aimed to construct prediction models using clinically available and reproducible characteristics for predicting IDH-mutant and CDKN2A/B homozygous deletion in adult-type diffuse glioma patients. MATERIALS AND METHODS: This retrospective, two-center study analysed 272 patients with adult-type diffuse glioma (230 for primary cohort and 42 for external validation cohort). Two radiologists independently assessed the patients' images according to the Visually AcceSAble Rembrandt Images (VASARI) feature set. Least absolute shrinkage and selection operator (LASSO) regression analysis was used to optimise variable selection. Multivariable logistic regression analysis was used to develop the prediction models. Calibration plots, receiver operating characteristic (ROC) curves, and decision curve analysis (DCA) were used to validate the models. Nomograms were developed visually based on the prediction models. RESULTS: The interobserver agreement between the two radiologists for VASARI features was excellent (κ range, 0.813-1). For the IDH-mutant prediction model, the area under the curves (AUCs) was 0.88-0.96 in the internal and external validation sets, For the CDKN2A/B homozygous deletion model, the AUCs were 0.80-0.86 in the internal and external validation sets. The decision curves show that both prediction models had good net benefits. CONCLUSION: The prediction models which basing on VASARI and clinical features provided a reliable and clinically meaningful preoperative prediction for IDH and CDKN2A/B status in diffuse glioma patients. These findings provide a foundation for precise preoperative non-invasive diagnosis and personalised treatment approaches for adult-type diffuse glioma patients.

Nomogram incorporating preoperative MRI-VASARI features for differentiating intracranial extraventricular ependymoma from glioblastoma.

Background: Intracranial extraventricular ependymoma (IEE) and glioblastoma (GBM) may have similar imaging findings but different prognosis. This study aimed to develop and validate a nomogram based on magnetic resonance imaging (MRI) Visually AcceSAble Rembrandt Images (VASARI) features for preoperatively differentiating IEE from GBM. Methods: The clinical data and the MRI-VASARI features of patients with confirmed IEE (n=114) and confirmed GBM (n=258) in a multicenter cohort were retrospectively analyzed. Predictive models for differentiating IEE from GBM were built using a multivariate logistic regression method. A nomogram was generated and the performance of the nomogram was assessed with respect to its calibration, discrimination, and clinical usefulness. Results: The predictors identified in this study consisted of six VASARI features and four clinical features. Compared with the individual models, the combined model incorporating clinical and VASARI features had the highest area under the curve (AUC) value [training set: 0.99, 95% confidence interval (CI): 0.98-1.00; validation set: 0.97, 95% CI: 0.94-1.00] in comparison to the clinical model. The nomogram was well calibrated with significant clinical benefit according to the calibration curve and decision curve analyses. Conclusions: The nomogram combining clinical and MRI-VASARI characteristics was robust for differentiating IEE from GBM preoperatively and may potentially assist in diagnosis and treatment of brain tumors.

Imaging predictors of 4q12 amplified and RB1 mutated glioblastoma IDH-wildtype.

INTRODUCTION: Recent studies have identified that glioblastoma IDH-wildtype consists of different molecular subgroups with distinct prognoses. In order to accurately describe and classify gliomas, the Visually AcceSAble Rembrandt Images (VASARI) system was developed. The goal of this study was to evaluate the VASARI characteristics in molecular subgroups of IDH-wildtype glioblastoma. METHODS: A retrospective analysis of glioblastoma IDH- wildtype with comprehensive next-generation sequencing and pre-operative and post-operative MRI was performed. VASARI characteristics and 205 genes were evaluated. Multiple comparison adjustment by the Bejamin-Hochberg false discovery rate (BH-FDR) was performed. A 1:3 propensity score match (PSM) with a Caliper of 0.2 was done. RESULTS: 178 patients with GBM IDH-WT met the inclusion criteria. 4q12 amplified patients (n = 20) were associated with cyst presence (30% vs. 12%, p = 0.042), decreased hemorrhage (35% vs. 62%, p = 0.028), and non-restricting/mixed (35%/60%) rather than restricting diffusion pattern (5%), meanwhile, 4q12 non-amplified patients had mostly restricting (47.4%) rather than a non-restricting/mixed diffusion pattern (28.4%/23.4%). This remained statistically significant after BH-FDR adjustment (p = 0.002). PSM by 4q12 amplification showed that diffusion characteristics continued to be significantly different. Among RB1-mutant patients, 96% had well-defined enhancing margins vs. 70.6% of RB1-WT (p = 0.018), however, this was not significant after BH-FDR or PSM. CONCLUSIONS: Patients with glioblastoma IDH-wildtype harboring 4q12 amplification rarely have restricting DWI patterns compared to their wildtype counterparts, in which this DWI pattern is present in ~ 50% of patients. This suggests that some phenotypic imaging characteristics can be identified among molecular subtypes of IDH-wildtype glioblastoma.

Beyond invasive biopsies: using VASARI MRI features to predict grade and molecular parameters in gliomas.

BACKGROUND: Gliomas present a significant economic burden and patient management challenge. The 2021 WHO classification incorporates molecular parameters, which guide treatment decisions. However, acquiring these molecular data involves invasive biopsies, prompting a need for non-invasive diagnostic methods. This study aims to assess the potential of Visually AcceSAble Rembrandt Images (VASARI) MRI features to predict glioma characteristics such as grade, IDH mutation, and MGMT methylation status. METHODS: This study enrolled 107 glioma patients treated between 2017 and 2022, meeting specific criteria including the absence of prior chemotherapy/radiation therapy, and the presence of molecular and MRI data. Images were assessed using the 27 VASARI MRI features by two blinded radiologists. Pathological and molecular assessments were conducted according to WHO 2021 CNS Tumor classification. Cross-validation Least Absolute Shrinkage and Selection Operator (CV-LASSO) logistic regression was applied for statistical analysis to identify significant VASARI features in determining glioma grade, IDH mutation, and MGMT methylation status. RESULTS: The study demonstrated substantial observer agreement in VASARI feature evaluation (inter- and intra-observer κ = 0.714 - 0.831 and 0.910, respectively). Patient imaging characteristics varied significantly with glioma grade, IDH mutation, and MGMT methylation. A predictive model was established using VASARI features for glioma grade prediction, exhibiting an AUC of 0.995 (95% CI = 0.986 - 0.998), 100% sensitivity, and 92.86% specificity. IDH mutation status was predicted with AUC 0.930 (95% CI = 0.882 - 0.977), and improved slightly to 0.933 with 'age-at-diagnosis' added. A model predicting MGMT methylation had a satisfactory performance (AUC 0.757, 95% CI = 0.645 - 0.868), improving to 0.791 when 'age-at-diagnosis' was added. CONCLUSIONS: The T1/FLAIR ratio, enhancement quality, hemorrhage, and proportion enhancing predict glioma grade with excellent accuracy. The proportion enhancing, thickness of enhancing margin, and T1/FLAIR ratio are significant predictors for IDH mutation status. Lastly, MGMT methylation is related to the longest diameter of the lesion, edema crossing the midline, and the proportion of the non-enhancing lesion. VASARI MRI features offer non-invasive and accurate predictive models for glioma grade, IDH mutation, and MGMT methylation status, enhancing glioma patient management.

Magnetic resonance imaging findings of intracranial extraventricular ependymoma: A retrospective multi-center cohort study of 114 cases.

BACKGROUND: Intracranial extraventricular ependymoma (IEE) is an ependymoma located in the brain parenchyma outside the ventricles. IEE has overlapping clinical and imaging characteristics with glioblastoma multiforme (GBM) but different treatment strategy and prognosis. Therefore, an accurate preoperative diagnosis is necessary for optimizing therapy for IEE. METHODS: A retrospective multicenter cohort of IEE and GBM was identified. MR imaging characteristics assessed with the Visually Accessible Rembrandt Images (VASARI) feature set and clinicopathological findings were recorded. Independent predictors for IEE were identified using multivariate logistic regression, which was used to construct a diagnostic score for differentiating IEE from GBM. RESULTS: Compared to GBM, IEE tended to occur in younger patients. Multivariate logistic regression analysis identified seven independent predictors for IEE. Among them, 3 predictors including tumor necrosis rate (F7), age, and tumor-enhancing margin thickness (F11), demonstrated higher diagnostic performance with an Area Under Curve (AUC) of more than 70% in distinguishing IEE from GBM. The AUC was 0.85, 0.78, and 0.70, with sensitivity of 92.98%, 72.81%, and 96.49%, and specificity of 65.50%, 73.64%, and 43.41%, for F7, age, and F11, respectively. CONCLUSION: We identified specific MR imaging features such as tumor necrosis and thickness of enhancing tumor margins that could help to differentiate IEE from GBM. Our study results should be helpful to assist in diagnosis and clinical management of this rare brain tumor.

Magnetic Resonance Imaging Features of Zinc Finger Translocation Associated-RELA Fusion Ependymoma Compared to Its Wild-Type Counterpart.

OBJECTIVE: To explore the predictive value of quantitative features extracted from conventional magnetic resonance imaging (MRI) in distinguishing Zinc Finger Translocation Associated (ZFTA)-RELA fusion-positive and wild-type ependymomas. METHODS: Twenty-seven patients with pathologically confirmed ependymomas (17 patients with ZFTA-RELA fusions and 10 ZFTA-RELA fusion-negative patients) who underwent conventional MRI were enrolled in this retrospective study. Two experienced neuroradiologists who were blinded to the histopathological subtypes independently extracted imaging features using Visually Accessible Rembrandt Images annotations. The consistency between the readers was evaluated with the Kappa test. The imaging features with significant differences between the 2 groups were obtained using the least absolute shrinkage and selection operator regression model. Logistic regression analysis and receiver operating characteristic analysis were performed to analyze the diagnostic performance of the imaging features in predicting the ZFTA-RELA fusion status in ependymoma. RESULTS: There was a good interevaluator agreement on the imaging features (kappa value range 0.601-1.000). Enhancement quality, thickness of the enhancing margin, and edema crossing the midline have high predictive performance in identifying ZFTA-RELA fusion-positive and ZFTA-RELA fusion-negative ependymomas (C-index = 0.862 and area under the curve= 0.8618). CONCLUSIONS: Quantitative features extracted from preoperative conventional MRI by Visually Accessible Rembrandt Images provide high discriminatory accuracy in predicting the ZFTA-RELA fusion status of ependymoma.

The combination of radiomics features and VASARI standard to predict glioma grade.

Background and Purpose: Radiomics features and The Visually AcceSAble Rembrandt Images (VASARI) standard appear to be quantitative and qualitative evaluations utilized to determine glioma grade. This study developed a preoperative model to predict glioma grade and improve the efficacy of clinical strategies by combining these two assessment methods. Materials and Methods: Patients diagnosed with glioma between March 2017 and September 2018 who underwent surgery and histopathology were enrolled in this study. A total of 3840 radiomic features were calculated; however, using the least absolute shrinkage and selection operator (LASSO) method, only 16 features were chosen to generate a radiomic signature. Three predictive models were developed using radiomic features and VASARI standard. The performance and validity of models were evaluated using decision curve analysis and 10-fold nested cross-validation. Results: Our study included 102 patients: 35 with low-grade glioma (LGG) and 67 with high-grade glioma (HGG). Model 1 utilized both radiomics and the VASARI standard, which included radiomic signatures, proportion of edema, and deep white matter invasion. Models 2 and 3 were constructed with radiomics or VASARI, respectively, with an area under the receiver operating characteristic curve (AUC) of 0.937 and 0.831, respectively, which was less than that of Model 1, with an AUC of 0.966. Conclusion: The combination of radiomics features and the VASARI standard is a robust model for predicting glioma grades.

Vasari Scoring System in Discerning between Different Degrees of Glioma and IDH Status Prediction: A Possible Machine Learning Application?

(1) The aim of our study is to evaluate the capacity of the Visually AcceSAble Rembrandt Images (VASARI) scoring system in discerning between the different degrees of glioma and Isocitrate Dehydrogenase (IDH) status predictions, with a possible application in machine learning. (2) A retrospective study was conducted on 126 patients with gliomas (M/F = 75/51; mean age: 55.30), from which we obtained their histological grade and molecular status. Each patient was analyzed with all 25 features of VASARI, blinded by two residents and three neuroradiologists. The interobserver agreement was assessed. A statistical analysis was conducted to evaluate the distribution of the observations using a box plot and a bar plot. We then performed univariate and multivariate logistic regressions and a Wald test. We also calculated the odds ratios and confidence intervals for each variable and the evaluation matrices with receiver operating characteristic (ROC) curves in order to identify cut-off values that are predictive of a diagnosis. Finally, we did the Pearson correlation test to see if the variables grade and IDH were correlated. (3) An excellent ICC estimate was obtained. For the grade and IDH status prediction, there were statistically significant results by evaluation of the degree of post-contrast impregnation (F4) and the percentage of impregnated area (F5), not impregnated area (F6), and necrotic (F7) tissue. These models showed good performances according to the area under the curve (AUC) values (>70%). (4) Specific MRI features can be used to predict the grade and IDH status of gliomas, with important prognostic implications. The standardization and improvement of these data (aim: AUC > 80%) can be used for programming machine learning software.

Identifying overall survival in 98 glioblastomas using VASARI features at 3T.

PURPOSE: This study aims to evaluate qualitative and quantitative imaging metrics along with clinical features affecting overall survival in glioblastomas and to classify them into high survival and low survival groups based on 12, 19, and 24 months thresholds using machine learning. METHODS: The cohort consisted of 98 adult glioblastomas. A standard brain tumor magnetic resonance (MR) imaging protocol, was performed on a 3T MR scanner. Visually Accessible REMBRANDT Images (VASARI) features were assessed. A Kaplan-Meier survival analysis followed by a log-rank test and multivariate Cox regression analysis were used to investigate the effects of VASARI features along with the age, gender, the extent of resection, pre- and post-KPS, ki67 and P53 mutation status on overall survival. Supervised machine learning algorithms were employed to predict the survival of glioblastoma patients based on 12, 19, and 24 months thresholds. RESULTS: Tumor location (p<0.001), the proportion of non-enhancing component (p=0.0482), and proportion of necrosis (p=0.02) were significantly associated with overall survival based on Kaplan-Meier analysis. Multivariate Cox regression analysis revealed that increases in proportion of non-enhancing component (p=0.040) and proportion of necrosis (p=0.039) were significantly associated with overall survival. Machine-learning models were successful in differentiating patients living longer than 12 months with 96.40% accuracy (sensitivity=97.22%, specificity=95.55%). The classification accuracies based on 19 and 24 months survival thresholds were 70.87% (sensitivity=83.02%, specificity=60.11%) and 74.66% (sensitivity=67.58%, specificity=82.08%), respectively. CONCLUSION: Employing clinical and VASARI features together resulted in a successful classification of glioblastomas that would have a longer overall survival.

Preoperative Magnetic Resonance Imaging Radiomics for Predicting Early Recurrence of Glioblastoma.

PURPOSE: Early recurrence of glioblastoma after standard treatment makes patient care challenging. This study aimed to assess preoperative magnetic resonance imaging (MRI) radiomics for predicting early recurrence of glioblastoma. PATIENTS AND METHODS: A total of 122 patients (training cohort: n = 86; validation cohort: n = 36) with pathologically confirmed glioblastoma were included in this retrospective study. Preoperative brain MRI images were analyzed for both radiomics and the Visually Accessible Rembrandt Image (VASARI) features of glioblastoma. Models incorporating MRI radiomics, the VASARI parameters, and clinical variables were developed and presented in a nomogram. Performance was assessed based on calibration, discrimination, and clinical usefulness. RESULTS: The nomogram consisting of the radiomic signatures, the VASARI parameters, and blood urea nitrogen (BUN) values showed good discrimination between the patients with early recurrence and those with later recurrence, with an area under the curve of 0.85 (95% CI, 0.77-0.94) in the training cohort and 0.84 [95% CI, 0.71-0.97] in the validation cohort. Decision curve analysis demonstrated favorable clinical application of the nomogram. CONCLUSION: This study showed the potential usefulness of preoperative brain MRI radiomics in predicting the early recurrence of glioblastoma, which should be helpful in personalized management of glioblastoma.

Prediction of H3K27M mutation status of diffuse midline gliomas using MRI features.

BACKGROUND AND PURPOSE: Presurgical prediction of H3K27M mutation in diffuse midline gliomas (DMGs) on MRI is desirable. The purpose of this study is to elaborate conventional MRI (cMRI) features of H3K27M-mutant DMGs and identify features that could discriminate them from wild-type (WT) DMGs. METHODS: CMRI features of 123 patients with DMG were evaluated conforming to the institutional research protocols. Multimodality MRI was performed on 1.5 or 3.0 Tesla MR Scanners with imaging protocol, including T1-weighted (w), T2w, fluid-attenuated inversion recovery, diffusion-weighted, susceptibility-weighted, and postcontrast T1w sequences. Pertinent cMRI features were annotated along the lines of Visually AcceSAble Rembrandt Images features, and Intra Tumoral Susceptibility Signal score (ITSS) was evaluated. R software was used for statistical analysis. RESULTS: Sixty-one DMGs were H3K27M-mutant (mutant DMGs). The patients in the H3K27M-mutant DMG group were younger compared to the WT-DMG group (mean age 24.13 ± 13.13 years vs. 35.79±18.74 years) (p = 0.016). The two groups differed on five cMRI features--(1) enhancement quality (p = 0.032), (2) thickness of enhancing margin (p = 0.05), (3) proportion of edema (p = 0.002), (4) definition of noncontrast-enhancing tumor (NCET) margin (p = 0.001), and (5) cortical invasion (p = 0.037). The mutant DMGs showed greater enhancement and greater thickness of enhancing margin, while the WT DMGs exhibited significantly larger edema proportion with poorly defined NCET margins and cortical invasion. ITSS was not significantly different among the groups. CONCLUSION: CMRI features like enhancement quality, the thickness of the enhancing margin, proportion of edema, definition of NCET margin, and cortical invasion can discriminate between the H3K27M-mutant and WT DMGs.

Radiogenomics of Gliomas.

The 2016 World Health Organization brain tumor classification is based on genomic and molecular profile of tumor tissue. These characteristics have improved understanding of the brain tumor and played an important role in treatment planning and prognostication. There is an ongoing effort to develop noninvasive imaging techniques that provide insight into tissue characteristics at the cellular and molecular levels. This article focuses on the molecular characteristics of gliomas, transcriptomic subtypes, and radiogenomic studies using semantic and radiomic features. The limitations and future directions of radiogenomics as a standalone diagnostic tool also are discussed.

Machine learning-based radiomic, clinical and semantic feature analysis for predicting overall survival and MGMT promoter methylation status in patients with glioblastoma.

INTRODUCTION: Survival varies in patients with glioblastoma due to intratumoral heterogeneity and radiomics/imaging biomarkers have potential to demonstrate heterogeneity. The objective was to combine radiomic, semantic and clinical features to improve prediction of overall survival (OS) and O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status from pre-operative MRI in patients with glioblastoma. METHODS: A retrospective study of 181 MRI studies (mean age 58 ± 13 years, mean OS 497 ± 354 days) performed in patients with histopathology-proven glioblastoma. Tumour mass, contrast-enhancement and necrosis were segmented from volumetric contrast-enhanced T1-weighted imaging (CE-T1WI). 333 radiomic features were extracted and 16 Visually Accessible Rembrandt Images (VASARI) features were evaluated by two experienced neuroradiologists. Top radiomic, VASARI and clinical features were used to build machine learning models to predict MGMT status, and all features including MGMT status were used to build Cox proportional hazards regression (Cox) and random survival forest (RSF) models for OS prediction. RESULTS: The optimal cut-off value for MGMT promoter methylation index was 12.75%; 42 radiomic features exhibited significant differences between high and low-methylation groups. However, model performance accuracy combining radiomic, VASARI and clinical features for MGMT status prediction varied between 45 and 67%. For OS predication, the RSF model based on clinical, VASARI and CE radiomic features achieved the best performance with an average iAUC of 96.2 ± 1.7 and C-index of 90.0 ± 0.3. CONCLUSIONS: VASARI features in combination with clinical and radiomic features from the enhancing tumour show promise for predicting OS with a high accuracy in patients with glioblastoma from pre-operative volumetric CE-T1WI.

Clinical, molecular, and radiomic profile of gliomas with FGFR3-TACC3 fusions.

BACKGROUND: Actionable fibroblast growth factor receptor 3 (FGFR3)-transforming acidic coiled-coil protein 3 fusions (F3T3) are found in approximately 3% of gliomas, but their characteristics and prognostic significance are still poorly defined. Our goal was to characterize the clinical, radiological, and molecular profile of F3T3 positive diffuse gliomas. METHODS: We screened F3T3 fusion by real-time (RT)-PCR and FGFR3 immunohistochemistry in a large series of gliomas, characterized for main genetic alterations, histology, and clinical evolution. We performed a radiological and radiomic case control study, using an exploratory and a validation cohort. RESULTS: We screened 1162 diffuse gliomas (951 unselected cases and 211 preselected for FGFR3 protein immunopositivity), identifying 80 F3T3 positive gliomas. F3T3 was mutually exclusive with IDH mutation (P < 0.001) and EGFR amplification (P = 0.01), defining a distinct molecular cluster associated with CDK4 (P = 0.04) and MDM2 amplification (P = 0.03). F3T3 fusion was associated with longer survival for the whole series and for glioblastomas (median overall survival was 31.1 vs 19.9 mo, P = 0.02) and was an independent predictor of better outcome on multivariate analysis.F3T3 positive gliomas had specific MRI features, affecting preferentially insula and temporal lobe, and with poorly defined tumor margins. F3T3 fusion was correctly predicted by radiomics analysis on both the exploratory (area under the curve [AUC] = 0.87) and the validation MRI (AUC = 0.75) cohort. Using Cox proportional hazards models, radiomics predicted survival with a high C-index (0.75, SD 0.04), while the model combining clinical, genetic, and radiomic data showed the highest C-index (0.81, SD 0.04). CONCLUSION: F3T3 positive gliomas have distinct molecular and radiological features, and better outcome.

Modelling MR and clinical features in grade II/III astrocytomas to predict IDH mutation status.

BACKGROUND AND PURPOSE: There is increasing evidence that many IDH wildtype (IDHwt) astrocytomas have a poor prognosis and although MR features have been identified, there remains diagnostic uncertainty in the clinic. We have therefore conducted a comprehensive analysis of conventional MR features of IDHwt astrocytomas and performed a Bayesian logistic regression model to identify critical radiological and basic clinical features that can predict IDH mutation status. MATERIALS AND METHODS: 146 patients comprising 52 IDHwt astrocytomas (19 WHO Grade II diffuse astrocytomas (A II) and 33 WHO Grade III anaplastic astrocytomas (A III)), 68 IDHmut astrocytomas (53 A II and 15 A III) and 26 GBM were studied. Age, sex, presenting symptoms and Overall Survival were recorded. Two neuroradiologists assessed 23 VASARI imaging descriptors of MRI features and the relation between IDH mutation status and MR and basic clinical features was modelled by Bayesian logistic regression, and survival by Kaplan-Meier plots. RESULTS: The features of greatest predictive power for IDH mutation status were, age at presentation (OR = 0.94 +/-0.03), tumour location within the thalamus (OR = 0.15 +/-0.25), involvement of speech receptive areas (OR = 0.21 +/-0.26), deep white matter invasion of the brainstem (OR = 0.10 +/-0.32), and T1/FLAIR signal ratio (OR = 1.63 +/-0.64). A logistic regression model based on these five features demonstrated excellent out-of-sample predictive performance (AUC = 0.92 +/-0.07; balanced accuracy 0.81 +/-0.09). Stepwise addition of further VASARI variables did not improve performance. CONCLUSION: Five demographic and VASARI features enable excellent individual prediction ofIDH mutation status, opening the way to identifying patients with IDHwt astrocytomas for earlier tissue diagnosis and more aggressive management.