High-amplitude cofluctuations in cortical activity drive functional connectivity.

Resting-state functional connectivity is used throughout neuroscience to study brain organization and to generate biomarkers of development, disease, and cognition. The processes that give rise to correlated activity are, however, poorly understood. Here we decompose resting-state functional connectivity using a temporal unwrapping procedure to assess the contributions of moment-to-moment activity cofluctuations to the overall connectivity pattern. This approach temporally resolves functional connectivity at a timescale of single frames, which enables us to make direct comparisons of cofluctuations of network organization with fluctuations in the blood oxygen level-dependent (BOLD) time series. We show that surprisingly, only a small fraction of frames exhibiting the strongest cofluctuation amplitude are required to explain a significant fraction of variance in the overall pattern of connection weights as well as the network's modular structure. These frames coincide with frames of high BOLD activity amplitude, corresponding to activity patterns that are remarkably consistent across individuals and identify fluctuations in default mode and control network activity as the primary driver of resting-state functional connectivity. Finally, we demonstrate that cofluctuation amplitude synchronizes across subjects during movie watching and that high-amplitude frames carry detailed information about individual subjects (whereas low-amplitude frames carry little). Our approach reveals fine-scale temporal structure of resting-state functional connectivity and discloses that frame-wise contributions vary across time. These observations illuminate the relation of brain activity to functional connectivity and open a number of directions for future research.

Cortico-subcortical interactions in overlapping communities of edge functional connectivity.

Both cortical and subcortical regions can be functionally organized into networks. Regions of the basal ganglia are extensively interconnected with the cortex via reciprocal connections that relay and modulate cortical function. Here we employ an edge-centric approach, which computes co-fluctuations among region pairs in a network to investigate the role and interaction of subcortical regions with cortical systems. By clustering edges into communities, we show that cortical systems and subcortical regions couple via multiple edge communities, with hippocampus and amygdala having a distinct pattern from striatum and thalamus. We show that the edge community structure of cortical networks is highly similar to one obtained from cortical nodes when the subcortex is present in the network. Additionally, we show that the edge community profile of both cortical and subcortical nodes can be estimates solely from cortico-subcortical interactions. Finally, we used a motif analysis focusing on edge community triads where a subcortical region coupled to two cortical regions and found that two community triads where one community couples the subcortex to the cortex were overrepresented. In summary, our results show organized coupling of the subcortex to the cortex that may play a role in cortical organization of primary sensorimotor/attention and heteromodal systems and puts forth the motif analysis of edge community triads as a promising method for investigation of communication patterns in networks.

Pure Prosodic Type of Primary Progressive Apraxia of Speech Mimicking Nonfluent Aphasia and Later Progressing to Corticobasal Syndrome.

Primary progressive apraxia of speech (PPAOS), a rare neurodedegenerative disorder, can be subdivided into predominant phonetic or prosodic type. Pure prosodic type of PPAOS as an isolated disorder has been hardly found. We present 2 cases of patients with pure prosodic PPAOS who initially were misdiagnosed as nonfluent variant of primary progressive aphasia and later turned out to be corticobasal syndrome. A 65-year-old woman and a 72-year-old man were referred to our speech-language clinic under the clinical impression of nonfluent variant of primary progressive aphasia. The neurological examinations revealed no definite abnormalities except for slow and effortful speech with the production of simple sentences. However, their receptive and expressive language abilities were normal. Their brain magnetic resonance imaging was unremarkable. We initially entertained the diagnosis of pure prosodic type of PPAOS. During several years of follow up, they gradually developed extrapyramidal symptoms which are compatible with corticobasal syndrome. The characteristics of the patients and the results of neuroimaging studies are discussed.

Subject identification using edge-centric functional connectivity.

Group-level studies do not capture individual differences in network organization, an important prerequisite for understanding neural substrates shaping behavior and for developing interventions in clinical conditions. Recent studies have employed 'fingerprinting' analyses on functional connectivity to identify subjects' idiosyncratic features. Here, we develop a complementary approach based on an edge-centric model of functional connectivity, which focuses on the co-fluctuations of edges. We first show whole-brain edge functional connectivity (eFC) to be a robust substrate that improves identifiability over nodal FC (nFC) across different datasets and parcellations. Next, we characterize subjects' identifiability at different spatial scales, from single nodes to the level of functional systems and clusters using k-means clustering. Across spatial scales, we find that heteromodal brain regions exhibit consistently greater identifiability than unimodal, sensorimotor, and limbic regions. Lastly, we show that identifiability can be further improved by reconstructing eFC using specific subsets of its principal components. In summary, our results highlight the utility of the edge-centric network model for capturing meaningful subject-specific features and sets the stage for future investigations into individual differences using edge-centric models.

Modularity maximization as a flexible and generic framework for brain network exploratory analysis.

The modular structure of brain networks supports specialized information processing, complex dynamics, and cost-efficient spatial embedding. Inter-individual variation in modular structure has been linked to differences in performance, disease, and development. There exist many data-driven methods for detecting and comparing modular structure, the most popular of which is modularity maximization. Although modularity maximization is a general framework that can be modified and reparamaterized to address domain-specific research questions, its application to neuroscientific datasets has, thus far, been narrow. Here, we highlight several strategies in which the "out-of-the-box" version of modularity maximization can be extended to address questions specific to neuroscience. First, we present approaches for detecting "space-independent" modules and for applying modularity maximization to signed matrices. Next, we show that the modularity maximization frame is well-suited for detecting task- and condition-specific modules. Finally, we highlight the role of multi-layer models in detecting and tracking modules across time, tasks, subjects, and modalities. In summary, modularity maximization is a flexible and general framework that can be adapted to detect modular structure resulting from a wide range of hypotheses. This article highlights multiple frontiers for future research and applications.

Prediction of Core Signaling Pathway by Using Diffusion- and Perfusion-based MRI Radiomics and Next-generation Sequencing in Isocitrate Dehydrogenase Wild-type Glioblastoma.

Background Next-generation sequencing (NGS) enables highly sensitive cancer genomics analysis, but its clinical implications for therapeutic options from imaging-based prediction have been limited. Purpose To predict core signaling pathways in isocitrate dehydrogenase (IDH) wild-type glioblastoma by using diffusion and perfusion MRI radiomics and NGS. Materials and Methods The radiogenomics model was developed by using retrospective patients with glioma who underwent NGS and anatomic, diffusion-, and perfusion-weighted imaging between March 2017 and February 2019. For testing model performance in predicting core signaling pathway, patients with IDH wild-type glioblastoma from a retrospective analysis from a registry (ClinicalTrials.gov NCT02619890) were evaluated. Radiogenomic feature selection was performed by using t tests, least absolute shrinkage and selection operator penalization, and random forest. Combining radiogenomic features, age, and location, the performance of predicting receptor tyrosine kinase (RTK), tumor protein p53 (P53), and retinoblastoma 1 pathways was evaluated by using the area under the receiver operating characteristic curve (AUC). Results There were 120 patients (52 years ± 13 [standard deviation]; 61 women) who were evaluated. Eighty-five patients (51 years ± 13; 43 men) were in the training set and 35 patients with IDH wild-type glioblastoma (56 years ± 12; 19 women) were in the validation set. Radiogenomics model identified 71 features in the RTK, 17 features in P53, and 35 features in the retinoblastoma pathway. The combined model showed better performance than anatomic imaging-based prediction in the RTK (P = .03) and retinoblastoma (P = .03) and perfusion imaging-based prediction in the P53 pathway (P = .04) in the training set. AUC values of the combined model for the prediction of core signaling pathways were 0.88 (95% confidence interval [CI]: 0.74, 1) for RTK, 0.76 (95% CI: 0.59, 0.92) for P53, and 0.81 (95% CI: 0.64, 0.97) for retinoblastoma in the validation set. Conclusion A diffusion- and perfusion-weighted MRI radiomics model can help characterize core signaling pathways and potentially guide targeted therapy for isocitrate dehydrogenase wild-type glioblastoma. © RSNA, 2019 Online supplemental material is available for this article.

Advanced imaging parameters improve the prediction of diffuse lower-grade gliomas subtype, IDH mutant with no 1p19q codeletion: added value to the T2/FLAIR mismatch sign.

OBJECTIVES: A combination of T2/FLAIR mismatch sign and advanced imaging parameters may improve the determination of molecular subtypes of diffuse lower-grade glioma. We assessed the diagnostic value of adding the apparent diffusion coefficient (ADC) and cerebral blood volume (CBV) to the T2/FLAIR mismatch sign for differentiation of the IDH mutation or 1p/19q codeletion. MATERIALS AND METHODS: Preoperative conventional, diffusion-weighted, and dynamic susceptibility contrast imaging were performed on 110 patients with diffuse lower-grade gliomas. The study population was classified into three groups using molecular subtype, namely IDH mutation and 1p/19q codeletion (IDHmut-Codel), IDH wild type (IDHwt) and IDH mutation and no 1p/19q codeletion (IDHmut-Noncodel). T2/FLAIR mismatch sign and the histogram parameters of apparent diffusion coefficient (ADC) and normalised cerebral blood volume (nCBV) values were assessed. A multivariate logistic regression model was constructed to distinguish IDHmut-Noncodel from IDHmut-Codel and IDHwt and from IDHwt, and the performance was compared with that of single parameters using the area under the receiver operating characteristics curve (AUC). RESULTS: Positive visual T2/FLAIR mismatch sign and higher nCBV skewness were significant variables to distinguish IDHmut-Noncodel from the other two groups (AUC, 0.88; 95% CI, 0.81-0.96). A lower ADC10 was a significant variable for distinguishing IDHmut-Noncodel from the IDHwt group (AUC, 0.75; 95% CI, 0.62-0.89). Adding ADC or CBV histogram parameters to T2/FLAIR mismatch sign improved performance in distinguishing IDHmut-Noncodel from the other two groups (AUC 0.882 vs. AUC 0.810) or from IDHwt (AUC 0.923 vs. AUC 0.868). CONCLUSIONS: The combination of the T2/FLAIR mismatch sign with ADC or CBV histogram parameters can improve the identification of IDHmut-Noncodel diffuse lower-grade gliomas, which can be easily applied in clinical practice. KEY POINTS: • The combination of the T2/FLAIR mismatch sign with the ADC or CBV histogram parameters can improve the identification of IDHmut-Noncodel diffuse lower-grade gliomas. • The multivariable model showed a significantly better performance for distinguishing the IDHmut-Noncodel group from other diffuse lower-grade gliomas than the T2/FLAIR mismatch sign alone or any single parameter. • The IDHmut-Noncodel type was associated with intermediate treatment outcomes; therefore, the identification of IDHmut-Noncodel diffuse lower-grade gliomas could be helpful for determining the clinical approach.

Repeatability of amide proton transfer-weighted signals in the brain according to clinical condition and anatomical location.

OBJECTIVES: To investigate whether clinical condition, imaging session, and locations affect repeatability of amide proton transfer-weighted (APTw) magnetic resonance imaging (MRI) in the brain. MATERIALS AND METHODS: Three APTw MRI data sets were acquired, involving two intrasession scans and one intersession scan for 19 healthy, 15 glioma, and 12 acute stroke adult participants (mean age 53.8, 54.6, and 68.5, respectively) on a 3T MR scanner. The mean APTw signals from five locations in healthy brain (supratentorial and infratentorial locations) and from entire tumor and stroke lesions (supratentorial location) were calculated. The within-subject coefficient of variation (wCV) and intraclass correlation coefficient (ICC) were calculated for each clinical conditions, image sessions, and anatomic locations. Differences in APTw signals between sessions were analyzed using repeated-measures analysis of variance. RESULTS: The ICC and wCV were 0.96 (95% confidence interval [CI], 0.91-0.99) and 16.1 (12.6-21.3) in glioma, 0.93 (0.82-0.98) and 15.0 (11.4-20.6) in stroke, and 0.84 (0.72-0.91) and 34.0 (28.7-41.0) in healthy brain. There were no significant differences in APTw signal between three sessions, irrespective of disease condition and location. The ICC and wCV were 0.85 (0.68-0.94) and 27.4 (21.8-35.6) in supratentorial, and 0.44 (- 0.18 to 0.76) and 32.7 (25.9 to 42.9) in infratentorial locations. There were significant differences in APTw signal between supra- (mean, 0.49%; 95% CI, 0.38-0.61) and infratentorial locations (1.09%, 0.98-1.20; p < 0.001). CONCLUSION: The repeatability of APTw signal was excellent in supratentorial locations, while it was poor in infratentorial locations due to severe B0 inhomogeneity and susceptibility which affects MTR asymmetry. KEY POINTS: • In supratentorial locations, APTw MRI showed excellent intrasession and intersession repeatability in brains of healthy controls and patients with glioma, as well as in stroke-affected regions. • APTw MRI showed excellent repeatability in supratentorial locations, but poor repeatability in infratentorial locations. • Considering poor repeatability in the infratentorial locations, the use of APTw MRI in longitudinal assessment in infratentorial locations is not indicated.

Modular subgraphs in large-scale connectomes underpin spontaneous co-fluctuation events in mouse and human brains.

Previous studies have adopted an edge-centric framework to study fine-scale network dynamics in human fMRI. To date, however, no studies have applied this framework to data collected from model organisms. Here, we analyze structural and functional imaging data from lightly anesthetized mice through an edge-centric lens. We find evidence of "bursty" dynamics and events - brief periods of high-amplitude network connectivity. Further, we show that on a per-frame basis events best explain static FC and can be divided into a series of hierarchically-related clusters. The co-fluctuation patterns associated with each cluster centroid link distinct anatomical areas and largely adhere to the boundaries of algorithmically detected functional brain systems. We then investigate the anatomical connectivity undergirding high-amplitude co-fluctuation patterns. We find that events induce modular bipartitions of the anatomical network of inter-areal axonal projections. Finally, we replicate these same findings in a human imaging dataset. In summary, this report recapitulates in a model organism many of the same phenomena observed in previously edge-centric analyses of human imaging data. However, unlike human subjects, the murine nervous system is amenable to invasive experimental perturbations. Thus, this study sets the stage for future investigation into the causal origins of fine-scale brain dynamics and high-amplitude co-fluctuations. Moreover, the cross-species consistency of the reported findings enhances the likelihood of future translation.

Superiority of magnetic resonance imaging in small renal mass diagnosis where image reports mismatches between computed tomography and magnetic resonance imaging.

PURPOSE: To analyze malignancy of computed tomography (CT) and magnetic resonance imaging (MRI) results in the same renal mass. MATERIALS AND METHODS: We retrospectively reviewed 1,216 patients who underwent partial nephrectomy from January 2017 to December 2021 in our institute. Patients who had both CT and MRI reports prior to surgery were included. We compared the diagnostic accuracy between the CT and the MRI. The patients were divided into two groups according to the consistency of reports: the 'Consistent group' and the 'Inconsistent group'. The Inconsistent group was further divided into two subgroups. Group 1 is the case that showed benign findings on CT but malignancy on MRI. Group 2 is the cases of malignancy on CT but benign on MRI. RESULTS: 410 patients were identified. Benign lesion was identified in 68 cases (16.6%). The sensitivity, specificity and diagnostic accuracy of MRI was 91.2%, 36.8%, and 82.2% respectively, whereas that of CT was 84.8%, 41.2%, and 77.6% respectively. Consistent group were 335 cases (81.7%) and inconsistent group were 75 cases (18.3%). The mean mass size was significantly smaller in the inconsistent group compared to the consistent group (consistent group vs. inconsistent group: 2.31±0.84 cm vs.1.84±0.75 cm, p<0.001). Also, the Group 1 had higher odds of malignancy compared to Group 2 in the renal mass size 2-4 cm (odds ratio, 5.62 [1.02-30.90]). CONCLUSIONS: Smaller mass size affects the discrepancy of CT and MRI reports. In addition, MRI showed better diagnostic performance in mismatch cases in the small renal masses.

Kaempferol Alleviates Mitochondrial Damage by Reducing Mitochondrial Reactive Oxygen Species Production in Lipopolysaccharide-Induced Prostate Organoids.

Common prostate diseases such as prostatitis and benign prostatic hyperplasia (BPH) have a high incidence at any age. Cellular stresses, such as reactive oxygen species (ROS) and chronic inflammation, are implicated in prostate enlargement and cancer progression and development. Kaempferol is a flavonoid found in abundance in various plants, including broccoli and spinach, and has been reported to exhibit positive biological activities, such as antioxidant and anti-inflammatory properties. In the present study, we introduced prostate organoids to investigate the protective effects of kaempferol against various cellular stresses. The levels of COX-2, iNOS, p-IκB, a pro-inflammatory cytokine, and ROS were increased by LPS treatment but reversed by kaempferol treatment. Kaempferol activated the nuclear factor erythroid 2-related factor 2(Nrf2)-related pathway and enhanced the mitochondrial quality control proteins PGC-1α, PINK1, Parkin, and Beclin. The increase in mitochondrial ROS and oxygen consumption induced by LPS was stabilized by kaempferol treatment. First, our study used prostate organoids as a novel evaluation platform. Secondly, it was demonstrated that kaempferol could alleviate the mitochondrial damage in LPS-induced induced prostate organoids by reducing the production of mitochondrial ROS.

The diversity and multiplexity of edge communities within and between brain systems.

The human brain is composed of functionally specialized systems that support cognition. Recently, we proposed an edge-centric model for detecting overlapping communities. It remains unclear how these communities and brain systems are related. Here, we address this question using data from the Midnight Scan Club and show that all brain systems are linked via at least two edge communities. We then examine the diversity of edge communities within each system, finding that heteromodal systems are more diverse than sensory systems. Next, we cluster the entire cortex to reveal it according to the regions' edge-community profiles. We find that regions in heteromodal systems are more likely to form their own clusters. Finally, we show that edge communities are personalized. Our work reveals the pervasive overlap of edge communities across the cortex and their relationship with brain systems. Our work provides pathways for future research using edge-centric brain networks.

Predicting Insignificant Prostate Cancer: Analysis of the Pathological Outcomes of Candidates for Active Surveillance according to the Pre-International Society of Urological Pathology (Pre-ISUP) 2014 Era Versus the Post-ISUP2014 Era.

PURPOSE: To analyze the difference in the prediction accuracy with an active surveillance (AS) protocol between two eras (pre-International Society of Urological Pathology [pre-ISUP]-2014 vs. post-ISUP2014). MATERIALS AND METHODS: We retrospectively analyzed 118 candidates for AS who underwent radical prostatectomy between 2009 and 2017. We divided our patients into two groups (group 1 [n=57], operation date 2009-2015; group 2 [n=61], operation date 2016-2017). Pathologic slides in group 1 were reviewed to distinguish men with cribriform pattern (CP) because the determination of Gleason scores in old era had been based on pre-ISUP2014 classification. Postoperative outcomes in the two eras were analyzed twice: first, all men in group 1 vs. group 2; second, the remaining men after excluding those with CPs in group 1 vs. group 2. RESULTS: The proportion of men with insignificant prostate cancer (iPCa) was significantly lower in group 1 than in group 2 (36.8% vs. 57.4%, p=0.040). After excluding 11 men with CPs from group 1, those remaining (46 men) were compared again with group 2. In this analysis, the proportion of men with iPCa was similar between the two groups (old vs. contemporary era: 41.3% vs. 57.4%, p=0.146). Nine of 11 men with CP had violated the criteria for iPCa in the earlier comparison. CONCLUSIONS: The accuracy of the AS protocol has been affected by the coexistence of CPs and pure Gleason 6 tumors in the pre-ISUP2014 era. We suggest to use only contemporary (post-ISUP2014) data to analyze the accuracy with AS protocols in future studies.

Edge-centric functional network representations of human cerebral cortex reveal overlapping system-level architecture.

Network neuroscience has relied on a node-centric network model in which cells, populations and regions are linked to one another via anatomical or functional connections. This model cannot account for interactions of edges with one another. In this study, we developed an edge-centric network model that generates constructs 'edge time series' and 'edge functional connectivity' (eFC). Using network analysis, we show that, at rest, eFC is consistent across datasets and reproducible within the same individual over multiple scan sessions. We demonstrate that clustering eFC yields communities of edges that naturally divide the brain into overlapping clusters, with regions in sensorimotor and attentional networks exhibiting the greatest levels of overlap. We show that eFC is systematically modulated by variation in sensory input. In future work, the edge-centric approach could be useful for identifying novel biomarkers of disease, characterizing individual variation and mapping the architecture of highly resolved neural circuits.

Delayed Functional Networks Development and Altered Fast Oscillation Dynamics in a Rat Model of Cortical Malformation.

Malformations of cortical development (MCD) is associated with a wide range of developmental delay and drug resistant epilepsy in children. By using resting-state functional magnetic resonance imaging (RS-fMRI) and event-related spectral perturbation (ERSP) of cortical electroencephalography (EEG) data, we tried to investigate the neural changes of spatiotemporal functional connectivity (FC) and fast oscillation (FO) dynamics in a rat model of methylazoxymethanol (MAM)-induced MCD. A total of 28 infant rats with prenatal exposure to MAM and those of age matched 28 controls with prenatal saline exposure were used. RS-fMRI were acquired at postnatal day 15 (P15) and 29 (P29), and correlation coefficient analysis of eleven region of interests (ROI) was done to find the differences of functional networks between four groups. Two hour-cortical EEGs were also recorded at P15 and P29 and the ERSP of gamma (30-80 Hz) and ripples (80-200 Hz) were analyzed. The rats with MCD showed significantly delayed development of superior colliculus-brainstem network compared to control rats at P15. In contrast to marked maturation of default mode network (DMN) in controls from P15 to P29, there was no clear development in MCD rats. The MCD rats showed significantly higher cortical gamma and ripples-ERSP at P15 and lower cortical ripples-ERSP at P29 than those of control rats. This study demonstrated delayed development of FC and altered cortical FO dynamics in rats with malformed brain. The results should be further investigated in terms of the epileptogenesis and cognitive dysfunction in patients with MCD.

Radiomics prognostication model in glioblastoma using diffusion- and perfusion-weighted MRI.

We aimed to develop and validate a multiparametric MR radiomics model using conventional, diffusion-, and perfusion-weighted MR imaging for better prognostication in patients with newly diagnosed glioblastoma. A total of 216 patients with newly diagnosed glioblastoma were enrolled from two tertiary medical centers and divided into training (n = 158) and external validation sets (n = 58). Radiomic features were extracted from contrast-enhanced T1-weighted imaging, fluid-attenuated inversion recovery, diffusion-weighted imaging, and dynamic susceptibility contrast imaging. After radiomic feature selection using LASSO regression, an individualized radiomic score was calculated. A multiparametric MR prognostic model was built using the radiomic score and clinical predictors. The results showed that the multiparametric MR prognostic model (radiomics score + clinical predictors) exhibited good discrimination (C-index, 0.74) and performed better than a conventional MR radiomics model (C-index, 0.65, P < 0.0001) or clinical predictors (C-index, 0.66; P < 0.0001). The multiparametric MR prognostic model also showed robustness in external validation (C-index, 0.70). Our results indicate that the incorporation of diffusion- and perfusion-weighted MR imaging into an MR radiomics model to improve prognostication in glioblastoma patients improved its performance over that achievable using clinical predictors alone.

Treatment of refractory IgA vasculitis with dapsone: a systematic review.

IgA vasculitis, formerly known as Henoch-Schönlein purpura, is a systemic IgA-mediated vasculitis of the small vessels commonly seen in children. The natural history of IgA vasculitis is generally self-limiting; however, one-third of patients experience symptom recurrence and a refractory course. This systematic review examined the use of dapsone in refractory IgA vasculitis cases. A literature search of PubMed databases retrieved 13 articles published until June 14, 2018. The most common clinical feature was a palpable rash (100% of patients), followed by joint pain (69.2%). Treatment response within 1-2 days was observed in 6 of 26 patients (23.1%) versus within 3-7 days in 17 patients (65.4%). Relapse after treatment discontinuation was reported in 17 patients (65.4%) but not in 3 patients (11.5 %). Four of the 26 patients (15.4%) reported adverse effects of dapsone including arthralgia (7.7%), rash (7.7%), and dapsone hypersensitivity syndrome (3.8%). Our findings suggest that dapsone may affect refractory IgA vasculitis. Multicenter randomized placebo-controlled trials are necessary to determine the standard dosage of dapsone at initial or tapering of treatment in IgA vasculitis patients and evaluate whether dapsone has a significant benefit versus steroids or other medications.

Incorporating diffusion- and perfusion-weighted MRI into a radiomics model improves diagnostic performance for pseudoprogression in glioblastoma patients.

BACKGROUND: Pseudoprogression is a diagnostic challenge in early posttreatment glioblastoma. We therefore developed and validated a radiomics model using multiparametric MRI to differentiate pseudoprogression from early tumor progression in patients with glioblastoma. METHODS: The model was developed from the enlarging contrast-enhancing portions of 61 glioblastomas within 3 months after standard treatment with 6472 radiomic features being obtained from contrast-enhanced T1-weighted imaging, fluid-attenuated inversion recovery imaging, and apparent diffusion coefficient (ADC) and cerebral blood volume (CBV) maps. Imaging features were selected using a LASSO (least absolute shrinkage and selection operator) logistic regression model with 10-fold cross-validation. Diagnostic performance for pseudoprogression was compared with that for single parameters (mean and minimum ADC and mean and maximum CBV) and single imaging radiomics models using the area under the receiver operating characteristics curve (AUC). The model was validated with an external cohort (n = 34) imaged on a different scanner and internal prospective registry data (n = 23). RESULTS: Twelve significant radiomic features (3 from conventional, 2 from diffusion, and 7 from perfusion MRI) were selected for model construction. The multiparametric radiomics model (AUC, 0.90) showed significantly better performance than any single ADC or CBV parameter (AUC, 0.57-0.79, P < 0.05), and better than a single radiomics model using conventional MRI (AUC, 0.76, P = 0.012), ADC (AUC, 0.78, P = 0.014), or CBV (AUC, 0.80, P = 0.43). The multiparametric radiomics showed higher performance in the external validation (AUC, 0.85) and internal validation (AUC, 0.96) than any single approach, thus demonstrating robustness. CONCLUSIONS: Incorporating diffusion- and perfusion-weighted MRI into a radiomics model improved diagnostic performance for identifying pseudoprogression and showed robustness in a multicenter setting.