The Brain Tumor Segmentation (BraTS) Challenge 2023: Focus on Pediatrics (CBTN-CONNECT-DIPGR-ASNR-MICCAI BraTS-PEDs)

Pediatric tumors of the central nervous system are the most common cause of cancer-related death in children. The five-year survival rate for high-grade gliomas in children is less than 20\%. Due to their rarity, the diagnosis of these entities is often delayed, their treatment is mainly based on historic treatment concepts, and clinical trials require multi-institutional collaborations. The MICCAI Brain Tumor Segmentation (BraTS) Challenge is a landmark community benchmark event with a successful history of 12 years of resource creation for the segmentation and analysis of adult glioma. Here we present the CBTN-CONNECT-DIPGR-ASNR-MICCAI BraTS-PEDs 2023 challenge, which represents the first BraTS challenge focused on pediatric brain tumors with data acquired across multiple international consortia dedicated to pediatric neuro-oncology and clinical trials. The BraTS-PEDs 2023 challenge focuses on benchmarking the development of volumentric segmentation algorithms for pediatric brain glioma through standardized quantitative performance evaluation metrics utilized across the BraTS 2023 cluster of challenges. Models gaining knowledge from the BraTS-PEDs multi-parametric structural MRI (mpMRI) training data will be evaluated on separate validation and unseen test mpMRI dataof high-grade pediatric glioma. The CBTN-CONNECT-DIPGR-ASNR-MICCAI BraTS-PEDs 2023 challenge brings together clinicians and AI/imaging scientists to lead to faster development of automated segmentation techniques that could benefit clinical trials, and ultimately the care of children with brain tumors.

The University of California San Francisco Brain Metastases Stereotactic Radiosurgery (UCSF-BMSR) MRI Dataset

Supplemental material is available for this article.

3D U-Net for automated detection of multiple sclerosis lesions: utility of transfer learning from other pathologies

Background and purpose

Deep learning algorithms for segmentation of multiple sclerosis (MS) plaques generally require training on large datasets. This manuscript evaluates the effect of transfer learning from segmentation of another pathology to facilitate use of smaller MS-specific training datasets. That is, a model trained for detection of one type of pathology was re-trained to identify MS lesions and active demyelination.

Materials and methods

In this retrospective study using MRI exams from 149 patients spanning 4/18/2014 to 7/8/2021, 3D convolutional neural networks were trained with a variable number of manually-segmented MS studies. Models were trained for FLAIR lesion segmentation at a single timepoint, new FLAIR lesion segmentation comparing two timepoints, and enhancing (actively demyelinating) lesion segmentation on T1 post-contrast imaging. Models were trained either de-novo or fine-tuned with transfer learning applied to a pre-existing model initially trained on non-MS data. Performance was evaluated with lesionwise sensitivity and positive predictive value (PPV).

Results

For single timepoint FLAIR lesion segmentation with 10 training studies, a fine-tuned model demonstrated improved performance [lesionwise sensitivity 0.55 ± 0.02 (mean ± standard error), PPV 0.66 ± 0.02] compared to a de-novo model (sensitivity 0.49 ± 0.02, p = 0.001; PPV 0.32 ± 0.02, p < 0.001). For new lesion segmentation with 30 training studies and their prior comparisons, a fine-tuned model demonstrated similar sensitivity (0.49 ± 0.05) and significantly improved PPV (0.60 ± 0.05) compared to a de-novo model (sensitivity 0.51 ± 0.04, p = 0.437; PPV 0.43 ± 0.04, p = 0.002). For enhancement segmentation with 20 training studies, a fine-tuned model demonstrated significantly improved overall performance (sensitivity 0.74 ± 0.06, PPV 0.69 ± 0.05) compared to a de-novo model (sensitivity 0.44 ± 0.09, p = 0.001; PPV 0.37 ± 0.05, p = 0.001).

Conclusion

By fine-tuning models trained for other disease pathologies with MS-specific data, competitive models identifying existing MS plaques, new MS plaques, and active demyelination can be built with substantially smaller datasets than would otherwise be required to train new models.

The Brain Tumor Segmentation (BraTS) Challenge 2023: Glioma Segmentation in Sub-Saharan Africa Patient Population (BraTS-Africa)

Gliomas are the most common type of primary brain tumors. Although gliomas are relatively rare, they are among the deadliest types of cancer, with a survival rate of less than 2 years after diagnosis. Gliomas are challenging to diagnose, hard to treat and inherently resistant to conventional therapy. Years of extensive research to improve diagnosis and treatment of gliomas have decreased mortality rates across the Global North, while chances of survival among individuals in low- and middle-income countries (LMICs) remain unchanged and are significantly worse in Sub-Saharan Africa (SSA) populations. Long-term survival with glioma is associated with the identification of appropriate pathological features on brain MRI and confirmation by histopathology. Since 2012, the Brain Tumor Segmentation (BraTS) Challenge have evaluated state-of-the-art machine learning methods to detect, characterize, and classify gliomas. However, it is unclear if the state-of-the-art methods can be widely implemented in SSA given the extensive use of lower-quality MRI technology, which produces poor image contrast and resolution and more importantly, the propensity for late presentation of disease at advanced stages as well as the unique characteristics of gliomas in SSA (i.e., suspected higher rates of gliomatosis cerebri). Thus, the BraTS-Africa Challenge provides a unique opportunity to include brain MRI glioma cases from SSA in global efforts through the BraTS Challenge to develop and evaluate computer-aided-diagnostic (CAD) methods for the detection and characterization of glioma in resource-limited settings, where the potential for CAD tools to transform healthcare are more likely.

Neuroimaging Patterns of Intracranial Infections: Meningitis, Cerebritis, and Their Complications

Neuroimaging provides rapid, noninvasive visualization of central nervous system infections for optimal diagnosis and management. Generalizable and characteristic imaging patterns help radiologists distinguish different types of intracranial infections including meningitis and cerebritis from a variety of bacterial, viral, fungal, and/or parasitic causes. Here, we describe key radiologic patterns of meningeal enhancement and diffusion restriction through profiles of meningitis, cerebritis, abscess, and ventriculitis. We discuss various imaging modalities and recent diagnostic advances such as deep learning through a survey of intracranial pathogens and their radiographic findings. Moreover, we explore critical complications and differential diagnoses of intracranial infections.

Development of Gestational Age-Based Fetal Brain and Intracranial Volume Reference Norms Using Deep Learning

BACKGROUND AND PURPOSE

Fetal brain MR imaging interpretations are subjective and require subspecialty expertise. We aimed to develop a deep learning algorithm for automatically measuring intracranial and brain volumes of fetal brain MRIs across gestational ages.

MATERIALS AND METHODS

This retrospective study included 246 patients with singleton pregnancies at 19-38 weeks gestation. A 3D U-Net was trained to segment the intracranial contents of 2D fetal brain MRIs in the axial, coronal, and sagittal planes. An additional 3D U-Net was trained to segment the brain from the output of the first model. Models were tested on MRIs of 10 patients (28 planes) via Dice coefficients and volume comparison with manual reference segmentations. Trained U-Nets were applied to 200 additional MRIs to develop normative reference intracranial and brain volumes across gestational ages and then to 9 pathologic fetal brains.

RESULTS

Fetal intracranial and brain compartments were automatically segmented in a mean of 6.8 (SD, 1.2) seconds with median Dices score of 0.95 and 0.90, respectively (interquartile ranges, 0.91-0.96/0.89-0.91) on the test set. Correlation with manual volume measurements was high (Pearson r = 0.996, P < .001). Normative samples of intracranial and brain volumes across gestational ages were developed. Eight of 9 pathologic fetal intracranial volumes were automatically predicted to be >2 SDs from this age-specific reference mean. There were no effects of fetal sex, maternal diabetes, or maternal age on intracranial or brain volumes across gestational ages.

CONCLUSIONS

Deep learning techniques can quickly and accurately quantify intracranial and brain volumes on clinical fetal brain MRIs and identify abnormal volumes on the basis of a normative reference standard.

The University of California San Francisco Preoperative Diffuse Glioma MRI Dataset

Supplemental material is available for this article. Keywords: Informatics, MR Diffusion Tensor Imaging, MR Perfusion, MR Imaging, Neuro-Oncology, CNS, Brain/Brain Stem, Oncology, Radiogenomics, Radiology-Pathology Integration © RSNA, 2022.

Longitudinal Assessment of Posttreatment Diffuse Glioma Tissue Volumes with Three-dimensional Convolutional Neural Networks

Neural networks were trained for segmentation and longitudinal assessment of posttreatment diffuse glioma. A retrospective cohort (from January 2018 to December 2019) of 298 patients with diffuse glioma (mean age, 52 years ± 14 [SD]; 177 men; 152 patients with glioblastoma, 72 patients with astrocytoma, and 74 patients with oligodendroglioma) who underwent two consecutive multimodal MRI examinations were randomly selected into training (n = 198) and testing (n = 100) samples. A posttreatment tumor segmentation three-dimensional nnU-Net convolutional neural network with multichannel inputs (T1, T2, and T1 postcontrast and fluid-attenuated inversion recovery [FLAIR]) was trained to segment three multiclass tissue types (peritumoral edematous, infiltrated, or treatment-changed tissue [ED]; active tumor or enhancing tissue [AT]; and necrotic core). Separate longitudinal change nnU-Nets were trained on registered and subtracted FLAIR and T1 postlongitudinal images to localize and better quantify and classify changes in ED and AT. Segmentation Dice scores, volume similarities, and 95th percentile Hausdorff distances ranged from 0.72 to 0.89, 0.90 to 0.96, and 2.5 to 3.6 mm, respectively. Accuracy rates of the posttreatment tumor segmentation and longitudinal change networks being able to classify longitudinal changes in ED and AT as increased, decreased, or unchanged were 76%-79% and 90%-91%, respectively. The accuracy levels of the longitudinal change networks were not significantly different from those of three neuroradiologists (accuracy, 90%-92%; κ, 0.58-0.63; P > .05). The results of this study support the potential clinical value of artificial intelligence-based automated longitudinal assessment of posttreatment diffuse glioma. Keywords: MR Imaging, Neuro-Oncology, Neural Networks, CNS, Brain/Brain Stem, Segmentation, Quantification, Convolutional Neural Network (CNN) Supplemental material is available for this article. © RSNA, 2022.

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.

Combining radiomics and deep convolutional neural network features from preoperative MRI for predicting clinically relevant genetic biomarkers in glioblastoma

Background

Glioblastoma is the most common primary brain malignancy, yet treatment options are limited, and prognosis remains guarded. Individualized tumor genetic assessment has become important for accurate prognosis and for guiding emerging targeted therapies. However, challenges remain for widespread tumor genetic testing due to costs and the need for tissue sampling. The aim of this study is to evaluate a novel artificial intelligence method for predicting clinically relevant genetic biomarkers from preoperative brain MRI in patients with glioblastoma.

Methods

We retrospectively analyzed preoperative MRI data from 400 patients with glioblastoma and grade 4 astrocytoma, IDH mutant who underwent resection and genetic testing. Nine genetic biomarkers were assessed: hotspot mutations of IDH1 or TERT promoter, pathogenic mutations of TP53, PTEN, ATRX, or CDKN2A/B, MGMT promoter methylation, EGFR amplification, and combined aneuploidy of chromosomes 7 & 10. Models were developed to predict biomarker status from MRI data using radiomics features, convolutional neural network (CNN) features, and a combination of both.

Results

Combined model performance was good for IDH1 and TERT promoter hotspot mutations, pathogenic mutations of ATRX and CDKN2A/B, and combined aneuploidy of chromosomes 7 & 10, with receiver operating characteristic area under the curve (ROC AUC) &gt; 0.85 and was fair for all other tested biomarkers with ROC AUC &gt; 0.7. Combined model performance was statistically superior to individual radiomics and CNN feature models for prediction chromosome 7 & 10 aneuploidy, MGMT promoter methylation, and PTEN loss.

Conclusions

Combining radiomics and CNN features from preoperative MRI yields improved non-invasive genetic biomarker prediction performance in patients with grade 4 diffuse gliomas.

Economic impact of selective use of contrast for routine follow-up MRI of patients with multiple sclerosis

BACKGROUND AND PURPOSE

Imaging and autopsy studies show intracranial gadolinium deposition in patients who have undergone serial contrast-enhanced MRIs. This observation has raised concerns when using contrast administration in patients who receive frequent MRIs. To address this, we implemented a contrast-conditional protocol wherein gadolinium is administered only for multiple sclerosis (MS) patients with imaging evidence of new disease activity on precontrast imaging. In this study, we explore the economic impact of our new MRI protocol.

METHODS

We compared scanner time and Medicare reimbursement using our contrast-conditional methodology versus that of prior protocols where all patients received gadolinium.

RESULTS

For 422 patients over 4 months, the contrast-conditional protocol amounted to 60% decrease in contrast injection and savings of approximately 20% of MRI scanner time. If the extra scanner time was used for performing MS follow-up MRIs in additional patients, the contrast-conditional protocol would amount to net revenue loss of $21,707 (∼3.7%).

CONCLUSIONS

Implementation of a new protocol to limit contrast in MS follow-up MRIs led to a minimal decrease in revenue when controlled for scanner time utilized and is outweighed by other benefits, including substantial decreased gadolinium administration, increased patient comfort, and increased availability of scanner time, which depending on type of studies performed could result in additional financial benefit.

Clinical Assessment of Deep Learning-based Super-Resolution for 3D Volumetric Brain MRI

Artificial intelligence (AI)-based image enhancement has the potential to reduce scan times while improving signal-to-noise ratio (SNR) and maintaining spatial resolution. This study prospectively evaluated AI-based image enhancement in 32 consecutive patients undergoing clinical brain MRI. Standard-of-care (SOC) three-dimensional (3D) T1 precontrast, 3D T2 fluid-attenuated inversion recovery, and 3D T1 postcontrast sequences were performed along with 45% faster versions of these sequences using half the number of phase-encoding steps. Images from the faster sequences were processed by a Food and Drug Administration-cleared AI-based image enhancement software for resolution enhancement. Four board-certified neuroradiologists scored the SOC and AI-enhanced image series independently on a five-point Likert scale for image SNR, anatomic conspicuity, overall image quality, imaging artifacts, and diagnostic confidence. While interrater κ was low to fair, the AI-enhanced scans were noninferior for all metrics and actually demonstrated a qualitative SNR improvement. Quantitative analyses showed that the AI software restored the high spatial resolution of small structures, such as the septum pellucidum. In conclusion, AI-based software can achieve noninferior image quality for 3D brain MRI sequences with a 45% scan time reduction, potentially improving the patient experience and scanner efficiency without sacrificing diagnostic quality. Keywords: MR Imaging, CNS, Brain/Brain Stem, Reconstruction Algorithms © RSNA, 2022.

Interinstitutional Portability of a Deep Learning Brain MRI Lesion Segmentation Algorithm

PURPOSE

To assess how well a brain MRI lesion segmentation algorithm trained at one institution performed at another institution, and to assess the effect of multi-institutional training datasets for mitigating performance loss.

MATERIALS AND METHODS

In this retrospective study, a three-dimensional U-Net for brain MRI abnormality segmentation was trained on data from 293 patients from one institution (IN1) (median age, 54 years; 165 women; patients treated between 2008 and 2018) and tested on data from 51 patients from a second institution (IN2) (median age, 46 years; 27 women; patients treated between 2003 and 2019). The model was then trained on additional data from various sources: (a) 285 multi-institution brain tumor segmentations, (b) 198 IN2 brain tumor segmentations, and (c) 34 IN2 lesion segmentations from various brain pathologic conditions. All trained models were tested on IN1 and external IN2 test datasets, assessing segmentation performance using Dice coefficients.

RESULTS

The U-Net accurately segmented brain MRI lesions across various pathologic conditions. Performance was lower when tested at an external institution (median Dice score, 0.70 [IN2] vs 0.76 [IN1]). Addition of 483 training cases of a single pathologic condition, including from IN2, did not raise performance (median Dice score, 0.72; P = .10). Addition of IN2 training data with heterogeneous pathologic features, representing only 10% (34 of 329) of total training data, increased performance to baseline (Dice score, 0.77; P < .001). This final model produced total lesion volumes with a high correlation to the reference standard (Spearman r = 0.98).

CONCLUSION

For brain MRI lesion segmentation, adding a modest amount of relevant training data from an external institution to a previously trained model supported successful application of the model to this external institution.Keywords: Neural Networks, Brain/Brain Stem, Segmentation Supplemental material is available for this article. © RSNA, 2021.

NIMG-28. PROSPECTIVE HISTOPATHOLOGY-VALIDATED MACHINE LEARNING FOR DISTINGUISHING TRUE PROGRESSION FROM TREATMENT-RELATED CHANGES IN GLIOBLASTOMA PATIENTS

PURPOSE

Decision making about the best course of treatment for glioblastoma patients becomes challenging when a new enhancing lesion appears in the vicinity of the surgical bed on follow-up MRI (after maximal safe tumor resection and chemoradiation), raising concerns for tumor progression (TP). Literature indicates 30-50% of these new lesions describe primarily treatment-related changes (TRC). We hypothesize that quantitative analysis of specific and sensitive features extracted from multi-parametric MRI (mpMRI) via machine learning (ML) techniques may yield non-invasive imaging signatures that distinguish TP from TRC and facilitate better treatment personalization.

METHODS

We have generated an ML model on a retrospective cohort of 58 subjects, and prospectively evaluated on an independent cohort of 58 previously unseen patients who underwent second resection for suspicious recurrence and had availability of advanced mpMRI (T1, T1-Gd, T2, T2-FLAIR, DTI, DSC). The features selected by our retrospective model, representing principal components analysis of intensity distributions, morphological, statistical, and texture descriptors, were extracted from the mpMRI of the prospective cohort. Integration of these features revealed signatures distinguishing between TP, mixed response, and TRC. Independently, a board-certified neuropathologist evaluated the resected tissue by blindly classifying it in the above three categories, based on mitotic figures, pseudopalisading necrosis, geographic necrosis, dystrophic calcification, vascular changes, and Ki67.

RESULTS

Tissues classified as TRC by the neuropathologist were associated with imaging phenotypes of lower angiogenesis (DSC-derived features), lower cellularity (DTI-derived features), and higher water concentration (T2, T2-FLAIR features). Our ML model characterized TP with 78% accuracy (sensitivity:86%, specificity:70%, AUC:0.80 (95%CI, 0.68-0.92)) and TRC with 81% accuracy (sensitivity:80%, specificity:81%, AUC:0.87 (95%CI, 0.72-1.00)).

CONCLUSION

Our proposed ML model reveals distinct non-invasive markers of TP and TRC, directly associated with histopathological changes in prospective glioblastoma patients. Reliable stratification of TP and TRC entities may help to noninvasively determine whether the course of treatment should change.

Automated multiclass tissue segmentation of clinical brain MRIs with lesions

•

A U-Net incorporating spatial prior information can successfully segment 6 brain tissue types.

•

The U-Net was able to segment gray and white matter in the presence of lesions.

•

The U-Net surpassed the performance of its source algorithm in an external dataset.

•

Segmentations were produced in a hundredth of the time of its predecessor algorithm.

Delineation and quantification of normal and abnormal brain tissues on Magnetic Resonance Images is fundamental to the diagnosis and longitudinal assessment of neurological diseases. Here we sought to develop a convolutional neural network for automated multiclass tissue segmentation of brain MRIs that was robust at typical clinical resolutions and in the presence of a variety of lesions. We trained a 3D U-Net for full brain multiclass tissue segmentation from a prior atlas-based segmentation method on an internal dataset that consisted of 558 clinical T1-weighted brain MRIs (453/52/53; training/validation/test) of patients with one of 50 different diagnostic entities (n = 362) or with a normal brain MRI (n = 196). We then used transfer learning to refine our model on an external dataset that consisted of 7 patients with hand-labeled tissue types. We evaluated the tissue-wise and intra-lesion performance with different loss functions and spatial prior information in the validation set and applied the best performing model to the internal and external test sets. The network achieved an average overall Dice score of 0.87 and volume similarity of 0.97 in the internal test set. Further, the network achieved a median intra-lesion tissue segmentation accuracy of 0.85 inside lesions within white matter and 0.61 inside lesions within gray matter. After transfer learning, the network achieved an average overall Dice score of 0.77 and volume similarity of 0.96 in the external dataset compared to human raters. The network had equivalent or better performance than the original atlas-based method on which it was trained across all metrics and produced segmentations in a hundredth of the time. We anticipate that this pipeline will be a useful tool for clinical decision support and quantitative analysis of clinical brain MRIs in the presence of lesions.

Brain MRI Deep Learning and Bayesian Inference System Augments Radiology Resident Performance

Automated quantitative and probabilistic medical image analysis has the potential to improve the accuracy and efficiency of the radiology workflow. We sought to determine whether AI systems for brain MRI diagnosis could be used as a clinical decision support tool to augment radiologist performance. We utilized previously developed AI systems that combine convolutional neural networks and expert-derived Bayesian networks to distinguish among 50 diagnostic entities on multimodal brain MRIs. We tested whether these systems could augment radiologist performance through an interactive clinical decision support tool known as Adaptive Radiology Interpretation and Education System (ARIES) in 194 test cases. Four radiology residents and three academic neuroradiologists viewed half of the cases unassisted and half with the results of the AI system displayed on ARIES. Diagnostic accuracy of radiologists for top diagnosis (TDx) and top three differential diagnosis (T3DDx) was compared with and without ARIES. Radiology resident performance was significantly better with ARIES for both TDx (55% vs 30%; P < .001) and T3DDx (79% vs 52%; P = 0.002), with the largest improvement for rare diseases (39% increase for T3DDx; P < 0.001). There was no significant difference between attending performance with and without ARIES for TDx (72% vs 69%; P = 0.48) or T3DDx (86% vs 89%; P = 0.39). These findings suggest that a hybrid deep learning and Bayesian inference clinical decision support system has the potential to augment diagnostic accuracy of non-specialists to approach the level of subspecialists for a large array of diseases on brain MRI.

Feasibility of Simulated Postcontrast MRI of Glioblastomas and Lower-Grade Gliomas by Using Three-dimensional Fully Convolutional Neural Networks

PURPOSE

To evaluate the feasibility and accuracy of simulated postcontrast T1-weighted brain MR images generated by using precontrast MR images in patients with brain glioma.

MATERIALS AND METHODS

In this retrospective study, a three-dimensional deep convolutional neural network was developed to simulate T1-weighted postcontrast images from eight precontrast sequences in 400 patients (mean age, 57 years; 239 men; from 2015 to 2020), including 332 with glioblastoma and 68 with lower-grade gliomas. Performance was evaluated by using quantitative image similarity and error metrics and enhancing tumor overlap analysis. Performance was also assessed on a multicenter external dataset (n = 286 from the 2019 Multimodal Brain Tumor Segmentation Challenge; mean age, 60 years; ratio of men to women unknown) by using transfer learning. A subset of cases was reviewed by neuroradiologist readers to assess whether simulated images affected the ability to determine the tumor grade.

RESULTS

Simulated whole-brain postcontrast images were both qualitatively and quantitatively similar to the real postcontrast images in terms of quantitative image similarity (structural similarity index of 0.84 ± 0.05), pixelwise error (symmetric mean absolute percent error of 3.65%), and enhancing tumor compartment overlap (Dice coefficient, 0.65 ± 0.25). Similar results were achieved with the external dataset (Dice coefficient, 0.62 ± 0.27). There was no difference in the ability of the neuroradiologist readers to determine the tumor grade in real versus simulated images (accuracy, 87.7% vs 90.6%; P = .87).

CONCLUSION

The developed model was capable of producing simulated postcontrast T1-weighted MR images that were similar to real acquired images as determined by both quantitative analysis and radiologist assessment.Keywords: MR-Contrast Agent, MR-Imaging, CNS, Brain/Brain Stem, Contrast Agents-Intravenous, Neoplasms-Primary, Experimental Investigations, Technology Assessment, Supervised Learning, Transfer Learning, Convolutional Neural Network, Deep Learning Algorithms, Machine Learning Algorithms Supplemental material is available for this article. © RSNA, 2021.

Three-dimensional U-Net Convolutional Neural Network for Detection and Segmentation of Intracranial Metastases

PURPOSE

To develop and validate a neural network for automated detection and segmentation of intracranial metastases on brain MRI studies obtained for stereotactic radiosurgery treatment planning.

MATERIALS AND METHODS

In this retrospective study, 413 patients (average age, 61 years ± 12 [standard deviation]; 238 women) with a total of 5202 intracranial metastases (median volume, 0.05 cm³; interquartile range, 0.02-0.18 cm³) undergoing stereotactic radiosurgery at one institution were included (January 2017 to February 2020). A total of 563 MRI examinations were performed among the patients, and studies were split into training (n = 413), validation (n = 50), and test (n = 100) datasets. A three-dimensional (3D) U-Net convolutional network was trained and validated on 413 T1 postcontrast or subtraction scans, and several loss functions were evaluated. After model validation, 100 discrete test patients, who underwent imaging after the training and validation patients, were used for final model evaluation. Performance for detection and segmentation of metastases was evaluated using Dice scores, false discovery rates, and false-negative rates, and a comparison with neuroradiologist interrater reliability was performed.

RESULTS

The median Dice score for segmenting enhancing metastases in the test set was 0.75 (interquartile range, 0.63-0.84). There were strong correlations between manually segmented and predicted metastasis volumes (r = 0.98, P < .001) and between the number of manually segmented and predicted metastases (R = 0.95, P < .001). Higher Dice scores were strongly correlated with larger metastasis volumes on a logarithmically transformed scale (r = 0.71). Sensitivity across the whole test sample was 70.0% overall and 96.4% for metastases larger than 6 mm. There was an average of 0.46 false-positive results per scan, with the positive predictive value being 91.5%. In comparison, the median Dice score between two neuroradiologists was 0.85 (interquartile range, 0.80-0.89), with sensitivity across the test sample being 87.9% overall and 98.4% for metastases larger than 6 mm.

CONCLUSION

A 3D U-Net-based convolutional neural network was able to segment brain metastases with high accuracy and perform detection at the level of human interrater reliability for metastases larger than 6 mm.Keywords: Adults, Brain/Brain Stem, CNS, Feature detection, MR-Imaging, Neural Networks, Neuro-Oncology, Quantification, Segmentation© RSNA, 2021.

Cancer Imaging Phenomics via CaPTk: Multi-Institutional Prediction of Progression-Free Survival and Pattern of Recurrence in Glioblastoma

PURPOSE

To construct a multi-institutional radiomic model that supports upfront prediction of progression-free survival (PFS) and recurrence pattern (RP) in patients diagnosed with glioblastoma multiforme (GBM) at the time of initial diagnosis.

PATIENTS AND METHODS

We retrospectively identified data for patients with newly diagnosed GBM from two institutions (institution 1, n = 65; institution 2, n = 15) who underwent gross total resection followed by standard adjuvant chemoradiation therapy, with pathologically confirmed recurrence, sufficient follow-up magnetic resonance imaging (MRI) scans to reliably determine PFS, and available presurgical multiparametric MRI (MP-MRI). The advanced software suite Cancer Imaging Phenomics Toolkit (CaPTk) was leveraged to analyze standard clinical brain MP-MRI scans. A rich set of imaging features was extracted from the MP-MRI scans acquired before the initial resection and was integrated into two distinct imaging signatures for predicting mean shorter or longer PFS and near or distant RP. The predictive signatures for PFS and RP were evaluated on the basis of different classification schemes: single-institutional analysis, multi-institutional analysis with random partitioning of the data into discovery and replication cohorts, and multi-institutional assessment with data from institution 1 as the discovery cohort and data from institution 2 as the replication cohort.

RESULTS

These predictors achieved cross-validated classification performance (ie, area under the receiver operating characteristic curve) of 0.88 (single-institution analysis) and 0.82 to 0.83 (multi-institution analysis) for prediction of PFS and 0.88 (single-institution analysis) and 0.56 to 0.71 (multi-institution analysis) for prediction of RP.

CONCLUSION

Imaging signatures of presurgical MP-MRI scans reveal relatively high predictability of time and location of GBM recurrence, subject to the patients receiving standard first-line chemoradiation therapy. Through its graphical user interface, CaPTk offers easy accessibility to advanced computational algorithms for deriving imaging signatures predictive of clinical outcome and could similarly be used for a variety of radiomic and radiogenomic analyses.

Subspecialty-Level Deep Gray Matter Differential Diagnoses with Deep Learning and Bayesian Networks on Clinical Brain MRI: A Pilot Study

Purpose

To develop and validate a system that could perform automated diagnosis of common and rare neurologic diseases involving deep gray matter on clinical brain MRI studies.

Materials and Methods

In this retrospective study, multimodal brain MRI scans from 212 patients (mean age, 55 years ± 17 [standard deviation]; 113 women) with 35 neurologic diseases and normal brain MRI scans obtained between January 2008 and January 2018 were included (110 patients in the training set, 102 patients in the test set). MRI scans from 178 patients (mean age, 48 years ± 17; 106 women) were used to supplement training of the neural networks. Three-dimensional convolutional neural networks and atlas-based image processing were used for extraction of 11 imaging features. Expert-derived Bayesian networks incorporating domain knowledge were used for differential diagnosis generation. The performance of the artificial intelligence (AI) system was assessed by comparing diagnostic accuracy with that of radiologists of varying levels of specialization by using the generalized estimating equation with robust variance estimator for the top three differential diagnoses (T3DDx) and the correct top diagnosis (TDx), as well as with receiver operating characteristic analyses.

Results

In the held-out test set, the imaging pipeline detected 11 key features on brain MRI scans with 89% accuracy (sensitivity, 81%; specificity, 95%) relative to academic neuroradiologists. The Bayesian network, integrating imaging features with clinical information, had an accuracy of 85% for T3DDx and 64% for TDx, which was better than that of radiology residents (n = 4; 56% for T3DDx, 36% for TDx; P < .001 for both) and general radiologists (n = 2; 53% for T3DDx, 31% for TDx; P < .001 for both). The accuracy of the Bayesian network was better than that of neuroradiology fellows (n = 2) for T3DDx (72%; P = .003) but not for TDx (59%; P = .19) and was not different from that of academic neuroradiologists (n = 2; 84% T3DDx, 65% TDx; P > .09 for both).

Conclusion

A hybrid AI system was developed that simultaneously provides a quantitative assessment of disease burden, explainable intermediate imaging features, and a probabilistic differential diagnosis that performed at the level of academic neuroradiologists. This type of approach has the potential to improve clinical decision making for common and rare diseases.Supplemental material is available for this article.© RSNA, 2020.

Artificial Intelligence System Approaching Neuroradiologist-level Differential Diagnosis Accuracy at Brain MRI

Background Although artificial intelligence (AI) shows promise across many aspects of radiology, the use of AI to create differential diagnoses for rare and common diseases at brain MRI has not been demonstrated. Purpose To evaluate an AI system for generation of differential diagnoses at brain MRI compared with radiologists. Materials and Methods This retrospective study tested performance of an AI system for probabilistic diagnosis in patients with 19 common and rare diagnoses at brain MRI acquired between January 2008 and January 2018. The AI system combines data-driven and domain-expertise methodologies, including deep learning and Bayesian networks. First, lesions were detected by using deep learning. Then, 18 quantitative imaging features were extracted by using atlas-based coregistration and segmentation. Third, these image features were combined with five clinical features by using Bayesian inference to develop probability-ranked differential diagnoses. Quantitative feature extraction algorithms and conditional probabilities were fine-tuned on a training set of 86 patients (mean age, 49 years ± 16 [standard deviation]; 53 women). Accuracy was compared with radiology residents, general radiologists, neuroradiology fellows, and academic neuroradiologists by using accuracy of top one, top two, and top three differential diagnoses in 92 independent test set patients (mean age, 47 years ± 18; 52 women). Results For accuracy of top three differential diagnoses, the AI system (91% correct) performed similarly to academic neuroradiologists (86% correct; P = .20), and better than radiology residents (56%; P < .001), general radiologists (57%; P < .001), and neuroradiology fellows (77%; P = .003). The performance of the AI system was not affected by disease prevalence (93% accuracy for common vs 85% for rare diseases; P = .26). Radiologists were more accurate at diagnosing common versus rare diagnoses (78% vs 47% across all radiologists; P < .001). Conclusion An artificial intelligence system for brain MRI approached overall top one, top two, and top three differential diagnoses accuracy of neuroradiologists and exceeded that of less-specialized radiologists. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Zaharchuk in this issue.

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.

Multi-Disease Segmentation of Gliomas and White Matter Hyperintensities in the BraTS Data Using a 3D Convolutional Neural Network

An important challenge in segmenting real-world biomedical imaging data is the presence of multiple disease processes within individual subjects. Most adults above age 60 exhibit a variable degree of small vessel ischemic disease, as well as chronic infarcts, which will manifest as white matter hyperintensities (WMH) on brain MRIs. Subjects diagnosed with gliomas will also typically exhibit some degree of abnormal T2 signal due to WMH, rather than just due to tumor. We sought to develop a fully automated algorithm to distinguish and quantify these distinct disease processes within individual subjects’ brain MRIs. To address this multi-disease problem, we trained a 3D U-Net to distinguish between abnormal signal arising from tumors vs. WMH in the 3D multi-parametric MRI (mpMRI, i.e., native T1-weighted, T1-post-contrast, T2, T2-FLAIR) scans of the International Brain Tumor Segmentation (BraTS) 2018 dataset (n_training = 285, n_validation = 66). Our trained neuroradiologist manually annotated WMH on the BraTS training subjects, finding that 69% of subjects had WMH. Our 3D U-Net model had a 4-channel 3D input patch (80 × 80 × 80) from mpMRI, four encoding and decoding layers, and an output of either four [background, active tumor (AT), necrotic core (NCR), peritumoral edematous/infiltrated tissue (ED)] or five classes (adding WMH as the fifth class). For both the four- and five-class output models, the median Dice for whole tumor (WT) extent (i.e., union of AT, ED, NCR) was 0.92 in both training and validation sets. Notably, the five-class model achieved significantly (p = 0.002) lower/better Hausdorff distances for WT extent in the training subjects. There was strong positive correlation between manually segmented and predicted volumes for WT (r = 0.96) and WMH (r = 0.89). Larger lesion volumes were positively correlated with higher/better Dice scores for WT (r = 0.33), WMH (r = 0.34), and across all lesions (r = 0.89) on a log(10) transformed scale. While the median Dice for WMH was 0.42 across training subjects with WMH, the median Dice was 0.62 for those with at least 5 cm³ of WMH. We anticipate the development of computational algorithms that are able to model multiple diseases within a single subject will be a critical step toward translating and integrating artificial intelligence systems into the heterogeneous real-world clinical workflow.

Multivariate Analysis of Preoperative Magnetic Resonance Imaging Reveals Transcriptomic Classification of de novo Glioblastoma Patients

Glioblastoma, the most frequent primary malignant brain neoplasm, is genetically diverse and classified into four transcriptomic subtypes, i. e., classical, mesenchymal, proneural, and neural. Currently, detection of transcriptomic subtype is based on ex vivo analysis of tissue that does not capture the spatial tumor heterogeneity. In view of accumulative evidence of in vivo imaging signatures summarizing molecular features of cancer, this study seeks robust non-invasive radiographic markers of transcriptomic classification of glioblastoma, based solely on routine clinically-acquired imaging sequences. A pre-operative retrospective cohort of 112 pathology-proven de novo glioblastoma patients, having multi-parametric MRI (T1, T1-Gd, T2, T2-FLAIR), collected from the Hospital of the University of Pennsylvania were included. Following tumor segmentation into distinct radiographic sub-regions, diverse imaging features were extracted and support vector machines were employed to multivariately integrate these features and derive an imaging signature of transcriptomic subtype. Extracted features included intensity distributions, volume, morphology, statistics, tumors' anatomical location, and texture descriptors for each tumor sub-region. The derived signature was evaluated against the transcriptomic subtype of surgically-resected tissue specimens, using a 5-fold cross-validation method and a receiver-operating-characteristics analysis. The proposed model was 71% accurate in distinguishing among the four transcriptomic subtypes. The accuracy (sensitivity/specificity) for distinguishing each subtype (classical, mesenchymal, proneural, neural) from the rest was equal to 88.4% (71.4/92.3), 75.9% (83.9/72.8), 82.1% (73.1/84.9), and 75.9% (79.4/74.4), respectively. The findings were also replicated in The Cancer Genomic Atlas glioblastoma dataset. The obtained imaging signature for the classical subtype was dominated by associations with features related to edge sharpness, whereas for the mesenchymal subtype had more pronounced presence of higher T2 and T2-FLAIR signal in edema, and higher volume of enhancing tumor and edema. The proneural and neural subtypes were characterized by the lower T1-Gd signal in enhancing tumor and higher T2-FLAIR signal in edema, respectively. Our results indicate that quantitative multivariate analysis of features extracted from clinically-acquired MRI may provide a radiographic biomarker of the transcriptomic profile of glioblastoma. Importantly our findings can be influential in surgical decision-making, treatment planning, and assessment of inoperable tumors.

Convolutional Neural Network for Automated FLAIR Lesion Segmentation on Clinical Brain MR Imaging

BACKGROUND AND PURPOSE

Most brain lesions are characterized by hyperintense signal on FLAIR. We sought to develop an automated deep learning-based method for segmentation of abnormalities on FLAIR and volumetric quantification on clinical brain MRIs across many pathologic entities and scanning parameters. We evaluated the performance of the algorithm compared with manual segmentation and existing automated methods.

MATERIALS AND METHODS

We adapted a U-Net convolutional neural network architecture for brain MRIs using 3D volumes. This network was retrospectively trained on 295 brain MRIs to perform automated FLAIR lesion segmentation. Performance was evaluated on 92 validation cases using Dice scores and voxelwise sensitivity and specificity, compared with radiologists' manual segmentations. The algorithm was also evaluated on measuring total lesion volume.

RESULTS

Our model demonstrated accurate FLAIR lesion segmentation performance (median Dice score, 0.79) on the validation dataset across a large range of lesion characteristics. Across 19 neurologic diseases, performance was significantly higher than existing methods (Dice, 0.56 and 0.41) and approached human performance (Dice, 0.81). There was a strong correlation between the predictions of lesion volume of the algorithm compared with true lesion volume (ρ = 0.99). Lesion segmentations were accurate across a large range of image-acquisition parameters on >30 different MR imaging scanners.

CONCLUSIONS

A 3D convolutional neural network adapted from a U-Net architecture can achieve high automated FLAIR segmentation performance on clinical brain MR imaging across a variety of underlying pathologies and image acquisition parameters. The method provides accurate volumetric lesion data that can be incorporated into assessments of disease burden or into radiologic reports.

Artificial intelligence for precision education in radiology

In the era of personalized medicine, the emphasis of health care is shifting from populations to individuals. Artificial intelligence (AI) is capable of learning without explicit instruction and has emerging applications in medicine, particularly radiology. Whereas much attention has focused on teaching radiology trainees about AI, here our goal is to instead focus on how AI might be developed to better teach radiology trainees. While the idea of using AI to improve education is not new, the application of AI to medical and radiological education remains very limited. Based on the current educational foundation, we highlight an AI-integrated framework to augment radiology education and provide use case examples informed by our own institution's practice. The coming age of "AI-augmented radiology" may enable not only "precision medicine" but also what we describe as "precision medical education," where instruction is tailored to individual trainees based on their learning styles and needs.

A Roadmap for Foundational Research on Artificial Intelligence in Medical Imaging: From the 2018 NIH/RSNA/ACR/The Academy Workshop

Imaging research laboratories are rapidly creating machine learning systems that achieve expert human performance using open-source methods and tools. These artificial intelligence systems are being developed to improve medical image reconstruction, noise reduction, quality assurance, triage, segmentation, computer-aided detection, computer-aided classification, and radiogenomics. In August 2018, a meeting was held in Bethesda, Maryland, at the National Institutes of Health to discuss the current state of the art and knowledge gaps and to develop a roadmap for future research initiatives. Key research priorities include: 1, new image reconstruction methods that efficiently produce images suitable for human interpretation from source data; 2, automated image labeling and annotation methods, including information extraction from the imaging report, electronic phenotyping, and prospective structured image reporting; 3, new machine learning methods for clinical imaging data, such as tailored, pretrained model architectures, and federated machine learning methods; 4, machine learning methods that can explain the advice they provide to human users (so-called explainable artificial intelligence); and 5, validated methods for image de-identification and data sharing to facilitate wide availability of clinical imaging data sets. This research roadmap is intended to identify and prioritize these needs for academic research laboratories, funding agencies, professional societies, and industry.

An Initiative to Reduce Unnecessary Gadolinium-Based Contrast in Multiple Sclerosis Patients

OBJECTIVE

Patients with multiple sclerosis (MS) routinely undergo serial contrast-enhanced MRIs. Given concerns regarding tissue deposition of gadolinium-based contrast agents (GBCAs) and evidence that enhancement of lesions is only seen in patients with new disease activity on noncontrast imaging, we set out to implement a prospective quality improvement project whereby intravenous contrast would be reserved only for patients with evidence of new disease activity on noncontrast images.

METHODS

To prospectively implement such a protocol, we leveraged our in-house computer-assisted detection (CAD) software and 3-D laboratory radiology technologists to perform real-time preliminary assessments of the CAD-processed T2 fluid attenuated inversion recovery (FLAIR) noncontrast images as a basis for deciding whether to inject contrast. Before implementation, we held multidisciplinary meetings with neurology, neuroradiology, and MR technologists and distributed surveys to objectively assess opinions and obstacles to clinical implementation. We evaluated reduction in GBCA utilization and technologist performance relative to final neuroradiologist interpretations.

RESULTS

During a 2-month trial period, 153 patients were imaged under the new protocol. Technologists using the CAD software were able to identify patients with new or enlarging lesions on FLAIR images with 95% accuracy and 97% negative predictive value relative to final neuroradiologist interpretations, which allowed us to avoid the use of contrast and additional imaging sequences in 87% of patients.

DISCUSSION

A multidisciplinary effort to implement a quality improvement project to limit contrast in MS patients receiving follow-up MRIs allowed for improved safety and cost by targeting patients that would benefit from the use of intravenous contrast in real-time.

Emerging Applications of Artificial Intelligence in Neuro-Oncology

Due to the exponential growth of computational algorithms, artificial intelligence (AI) methods are poised to improve the precision of diagnostic and therapeutic methods in medicine. The field of radiomics in neuro-oncology has been and will likely continue to be at the forefront of this revolution. A variety of AI methods applied to conventional and advanced neuro-oncology MRI data can already delineate infiltrating margins of diffuse gliomas, differentiate pseudoprogression from true progression, and predict recurrence and survival better than methods used in daily clinical practice. Radiogenomics will also advance our understanding of cancer biology, allowing noninvasive sampling of the molecular environment with high spatial resolution and providing a systems-level understanding of underlying heterogeneous cellular and molecular processes. By providing in vivo markers of spatial and molecular heterogeneity, these AI-based radiomic and radiogenomic tools have the potential to stratify patients into more precise initial diagnostic and therapeutic pathways and enable better dynamic treatment monitoring in this era of personalized medicine. Although substantial challenges remain, radiologic practice is set to change considerably as AI technology is further developed and validated for clinical use.

Neuroimaging of Dilated Perivascular Spaces: From Benign and Pathologic Causes to Mimics

Perivascular spaces (PVSs), also known as Virchow-Robin spaces, are pial-lined, fluid-filled structures found in characteristic locations throughout the brain. They can become abnormally enlarged or dilated and in rare cases can cause hydrocephalus. Dilated PVSs can pose a diagnostic dilemma for radiologists because of their varied appearance, sometimes mimicking more serious entities such as cystic neoplasms, including dysembryoplastic neuroepithelial tumor and multinodular and vacuolating neuronal tumor, or cystic infections including toxoplasmosis and neurocysticercosis. In addition, various pathologic processes, including cryptococcosis and chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids, can spread into the brain via PVSs, resulting in characteristic magnetic resonance imaging appearances. This review aims to describe the key imaging characteristics of normal and dilated PVSs, as well as cystic mimics and pathologic processes that directly involve PVSs.

Machine learning classification of mesial temporal sclerosis in epilepsy patients

BACKGROUND AND PURPOSE

Novel approaches applying machine-learning methods to neuroimaging data seek to develop individualized measures that will aid in the diagnosis and treatment of brain-based disorders such as temporal lobe epilepsy (TLE). Using a large cohort of epilepsy patients with and without mesial temporal sclerosis (MTS), we sought to automatically classify MTS using measures of cortical morphology, and to further relate classification probabilities to measures of disease burden.

MATERIALS AND METHODS

Our sample consisted of high-resolution T1 structural scans of 169 adults with epilepsy collected across five different 1.5T and four different 3T scanners at UCLA. We applied a multiple support vector machine recursive feature elimination algorithm to morphological measures generated from FreeSurfer's automated segmentation and parcellation in order to classify Epilepsy patients with MTS (n=85) from those without MTS (N=84).

RESULTS

In addition to hippocampal volume, we found that alterations in cortical thickness, surface area, volume and curvature in inferior frontal and anterior and inferior temporal regions contributed to a classification accuracy of up to 81% (p=1.3×10(-17)) in identifying MTS. We also found that MTS classification probabilities were associated with a longer duration of disease for epilepsy patients both with and without MTS.

CONCLUSIONS

In addition to implicating extra-hippocampal involvement of MTS, these findings shed further light on the pathogenesis of TLE and may ultimately assist in the development of automated tools that incorporate multiple neuroimaging measures to assist clinicians in detecting more subtle cases of TLE and MTS.

Overreactive brain responses to sensory stimuli in youth with autism spectrum disorders

OBJECTIVES

Sensory over-responsivity (SOR), defined as a negative response to or avoidance of sensory stimuli, is both highly prevalent and extremely impairing in youth with autism spectrum disorders (ASD), yet little is known about the neurological bases of SOR. This study aimed to examine the functional neural correlates of SOR by comparing brain responses to sensory stimuli in youth with and without ASD.

METHOD

A total of 25 high-functioning youth with ASD and 25 age- and IQ-equivalent typically developing (TD) youth were presented with mildly aversive auditory and visual stimuli during a functional magnetic resonance imaging (fMRI) scan. Parents provided ratings of children's SOR and anxiety symptom severity.

RESULTS

Compared to TD participants, ASD participants displayed greater activation in primary sensory cortical areas as well as amygdala, hippocampus, and orbital-frontal cortex. In both groups, the level of activity in these areas was positively correlated with level of SOR severity as rated by parents, over and above behavioral ratings of anxiety.

CONCLUSIONS

This study demonstrates that youth with ASD show neural hyper-responsivity to sensory stimuli, and that behavioral symptoms of SOR may be related to both heightened responsivity in primary sensory regions as well as areas related to emotion processing and regulation.

An fMRI investigation of responses to peer rejection in adolescents with autism spectrum disorders

Peer rejection is particularly pervasive among adolescents with autism spectrum disorders (ASD). However, how adolescents with ASD differ from typically developing adolescents in their responses to peer rejection is poorly understood. The goal of the current investigation was to examine neural responses to peer exclusion among adolescents with ASD compared to typically developing adolescents. Nineteen adolescents with ASD and 17 typically developing controls underwent fMRI as they were ostensibly excluded by peers during an online game called Cyberball. Afterwards, participants reported their distress about the exclusion. Compared to typically developing adolescents, those with ASD displayed less activity in regions previously linked with the distressing aspect of peer exclusion, including the subgenual anterior cingulate and anterior insula, as well as less activity in regions previously linked with the regulation of distress responses during peer exclusion, including the ventrolateral prefrontal cortex and ventral striatum. Interestingly, however, both groups self-reported equivalent levels of distress. This suggests that adolescents with ASD may engage in differential processing of social experiences at the neural level, but be equally aware of, and concerned about, peer rejection. Overall, these findings contribute new insights about how this population may differentially experience negative social events in their daily lives.

The UCLA multimodal connectivity database: a web-based platform for brain connectivity matrix sharing and analysis

Brain connectomics research has rapidly expanded using functional MRI (fMRI) and diffusion-weighted MRI (dwMRI). A common product of these varied analyses is a connectivity matrix (CM). A CM stores the connection strength between any two regions (“nodes”) in a brain network. This format is useful for several reasons: (1) it is highly distilled, with minimal data size and complexity, (2) graph theory can be applied to characterize the network's topology, and (3) it retains sufficient information to capture individual differences such as age, gender, intelligence quotient (IQ), or disease state. Here we introduce the UCLA Multimodal Connectivity Database (http://umcd.humanconnectomeproject.org), an openly available website for brain network analysis and data sharing. The site is a repository for researchers to publicly share CMs derived from their data. The site also allows users to select any CM shared by another user, compute graph theoretical metrics on the site, visualize a report of results, or download the raw CM. To date, users have contributed over 2000 individual CMs, spanning different imaging modalities (fMRI, dwMRI) and disorders (Alzheimer's, autism, Attention Deficit Hyperactive Disorder). To demonstrate the site's functionality, whole brain functional and structural connectivity matrices are derived from 60 subjects' (ages 26–45) resting state fMRI (rs-fMRI) and dwMRI data and uploaded to the site. The site is utilized to derive graph theory global and regional measures for the rs-fMRI and dwMRI networks. Global and nodal graph theoretical measures between functional and structural networks exhibit low correspondence. This example demonstrates how this tool can enhance the comparability of brain networks from different imaging modalities and studies. The existence of this connectivity-based repository should foster broader data sharing and enable larger-scale meta-analyses comparing networks across imaging modality, age group, and disease state.

Autism-associated promoter variant in MET impacts functional and structural brain networks

As genes that confer increased risk for autism spectrum disorder (ASD) are identified, a crucial next step is to determine how these risk factors impact brain structure and function and contribute to disorder heterogeneity. With three converging lines of evidence, we show that a common, functional ASD risk variant in the Met Receptor Tyrosine Kinase (MET) gene is a potent modulator of key social brain circuitry in children and adolescents with and without ASD. MET risk genotype predicted atypical fMRI activation and deactivation patterns to social stimuli (i.e., emotional faces), as well as reduced functional and structural connectivity in temporo-parietal regions known to have high MET expression, particularly within the default mode network. Notably, these effects were more pronounced in individuals with ASD. These findings highlight how genetic stratification may reduce heterogeneity and help elucidate the biological basis of complex neuropsychiatric disorders such as ASD.

Insights into multimodal imaging classification of ADHD

Attention deficit hyperactivity disorder (ADHD) currently is diagnosed in children by clinicians via subjective ADHD-specific behavioral instruments and by reports from the parents and teachers. Considering its high prevalence and large economic and societal costs, a quantitative tool that aids in diagnosis by characterizing underlying neurobiology would be extremely valuable. This provided motivation for the ADHD-200 machine learning (ML) competition, a multisite collaborative effort to investigate imaging classifiers for ADHD. Here we present our ML approach, which used structural and functional magnetic resonance imaging data, combined with demographic information, to predict diagnostic status of individuals with ADHD from typically developing (TD) children across eight different research sites. Structural features included quantitative metrics from 113 cortical and non-cortical regions. Functional features included Pearson correlation functional connectivity matrices, nodal and global graph theoretical measures, nodal power spectra, voxelwise global connectivity, and voxelwise regional homogeneity. We performed feature ranking for each site and modality using the multiple support vector machine recursive feature elimination (SVM-RFE) algorithm, and feature subset selection by optimizing the expected generalization performance of a radial basis function kernel SVM (RBF-SVM) trained across a range of the top features. Site-specific RBF-SVMs using these optimal feature sets from each imaging modality were used to predict the class labels of an independent hold-out test set. A voting approach was used to combine these multiple predictions and assign final class labels. With this methodology we were able to predict diagnosis of ADHD with 55% accuracy (versus a 39% chance level in this sample), 33% sensitivity, and 80% specificity. This approach also allowed us to evaluate predictive structural and functional features giving insight into abnormal brain circuitry in ADHD.

Regional fMRI hypoactivation and altered functional connectivity during emotion processing in nonmedicated depressed patients with bipolar II disorder

OBJECTIVE

Although the amygdala and ventrolateral prefrontal cortex have been implicated in the pathophysiology of bipolar I disorder, the neural mechanisms underlying bipolar II disorder remain unknown. The authors examined neural activity in response to negative emotional faces during an emotion perception task that reliably activates emotion regulatory regions.

METHOD

Twenty-one nonmedicated depressed bipolar II patients and 21 healthy comparison subjects underwent functional MRI (fMRI) while performing an emotional face-matching task. Within- and between-group whole-brain fMRI activation and seed-based connectivity analyses were conducted.

RESULTS

In depressed bipolar II patients, random-effects between-group fMRI analyses revealed a significant reduction in activation in several regions, including the left and right ventrolateral prefrontal cortices (Brodmann's area [BA] 47) and the right amygdala, a priori regions of interest. Additionally, bipolar patients exhibited significantly reduced negative functional connectivity between the right amygdala and the right orbitofrontal cortex (BA 10) as well as the right dorsolateral prefrontal cortex (BA 46) relative to healthy comparison subjects.

CONCLUSIONS

These findings suggest that bipolar II depression is characterized by reduced regional orbitofrontal and limbic activation and altered connectivity in a fronto-temporal circuit implicated in working memory and emotional learning. While the amygdala hypoactivation observed in bipolar II depression is opposite to the direction seen in bipolar I mania and may therefore be state dependent, the observed orbitofrontal cortex hypoactivation is consistent with findings in bipolar I depression, mania, and euthymia, suggesting a physiologic trait marker of the disorder.

Frontostriatal connectivity in children during working memory and the effects of prenatal methamphetamine, alcohol, and polydrug exposure

Various abnormalities in frontal and striatal regions have been reported in children with prenatal alcohol and/or methamphetamine exposure. In a recent fMRI study, we observed a correlation between accuracy on a working-memory task and functional activation in the putamen in children with prenatal methamphetamine and polydrug exposure. Because the putamen is part of the corticostriatal motor loop whereas the caudate is involved in the executive loop, we hypothesized that a loss of segregation between distinct corticostriatal networks may occur in these participants. The current study was designed to test this hypothesis using functional connectivity MRI. We examined 50 children ranging in age from 7 to 15, including 19 with prenatal methamphetamine exposure (15 of whom had concomitant prenatal alcohol exposure), 13 with prenatal exposure to alcohol but not methamphetamine, and 18 unexposed controls. We measured the coupling between blood oxygenation level dependent (BOLD) fluctuations during a working-memory task in four striatal seed regions and those in the rest of the brain. We found that the putamen seeds showed increased connectivity with frontal brain regions involved in executive functions while the caudate seeds showed decreased connectivity with some of these regions in both groups of exposed subjects compared to controls. These findings suggest that localized brain abnormalities resulting from prenatal exposure to alcohol and/or methamphetamine lead to a partial rewiring of corticostriatal networks. These results represent important progress in the field, and could have substantial clinical significance in helping devise more targeted treatments and remediation strategies designed to better serve the needs of this population.

Atypical Neural Processing of Ironic and Sincere Remarks in Children and Adolescents with Autism Spectrum Disorders

Individuals with ASD show consistent impairment in processing pragmatic language when attention to multiple social cues (e.g., facial expression, tone of voice) is often needed to navigate social interactions. Building upon prior fMRI work examining how facial affect and prosodic cues are used to infer a speaker's communicative intent, the authors examined whether children and adolescents with ASD differ from typically developing (TD) controls in their processing of sincere versus ironic remarks. At the behavioral level, children and adolescents with ASD and matched TD controls were able to determine whether a speaker's remark was sincere or ironic equally well, with both groups showing longer response times for ironic remarks. At the neural level, for both sincere and ironic scenarios, an extended cortical network-including canonical language areas in the left hemisphere and their right hemisphere counterparts-was activated in both groups, albeit to a lesser degree in the ASD sample. Despite overall similar patterns of activity observed for the two conditions in both groups, significant modulation of activity was detected when directly comparing sincere and ironic scenarios within and between groups. While both TD and ASD groups showed significantly greater activity in several nodes of this extended network when processing ironic versus sincere remarks, increased activity was largely confined to left language areas in TD controls, whereas the ASD sample showed a more bilateral activation profile which included both language and "theory of mind" areas (i.e., ventromedial prefrontal cortex). These findings suggest that, for high-functioning individuals with ASD, increased activity in right hemisphere homologues of language areas in the left hemisphere, as well as regions involved in social cognition, may reflect compensatory mechanisms supporting normative behavioral task performance.

Reduced functional integration and segregation of distributed neural systems underlying social and emotional information processing in autism spectrum disorders

A growing body of evidence suggests that autism spectrum disorders (ASDs) are related to altered communication between brain regions. Here, we present findings showing that ASD is characterized by a pattern of reduced functional integration as well as reduced segregation of large-scale brain networks. Twenty-three children with ASD and 25 typically developing matched controls underwent functional magnetic resonance imaging while passively viewing emotional face expressions. We examined whole-brain functional connectivity of two brain structures previously implicated in emotional face processing in autism: the amygdala bilaterally and the right pars opercularis of the inferior frontal gyrus (rIFGpo). In the ASD group, we observed reduced functional integration (i.e., less long-range connectivity) between amygdala and secondary visual areas, as well as reduced segregation between amygdala and dorsolateral prefrontal cortex. For the rIFGpo seed, we observed reduced functional integration with parietal cortex and increased integration with right frontal cortex as well as right nucleus accumbens. Finally, we observed reduced segregation between rIFGpo and the ventromedial prefrontal cortex. We propose that a systems-level approach-whereby the integration and segregation of large-scale brain networks in ASD is examined in relation to typical development-may provide a more detailed characterization of the neural basis of ASD.

Altered functional connectivity in frontal lobe circuits is associated with variation in the autism risk gene CNTNAP2

Genetic studies are rapidly identifying variants that shape risk for disorders of human cognition, but the question of how such variants predispose to neuropsychiatric disease remains. Noninvasive human brain imaging allows assessment of the brain in vivo, and the combination of genetics and imaging phenotypes remains one of the only ways to explore functional genotype-phenotype associations in human brain. Common variants in contactin-associated protein-like 2 (CNTNAP2), a neurexin superfamily member, have been associated with several allied neurodevelopmental disorders, including autism and specific language impairment, and CNTNAP2 is highly expressed in frontal lobe circuits in the developing human brain. Using functional neuroimaging, we have demonstrated a relationship between frontal lobar connectivity and common genetic variants in CNTNAP2. These data provide a mechanistic link between specific genetic risk for neurodevelopmental disorders and empirical data implicating dysfunction of long-range connections within the frontal lobe in autism. The convergence between genetic findings and cognitive-behavioral models of autism provides evidence that genetic variation at CNTNAP2 predisposes to diseases such as autism in part through modulation of frontal lobe connectivity.

Altered structural brain connectivity in healthy carriers of the autism risk gene, CNTNAP2

Recently, carriers of a common variant in the autism risk gene, CNTNAP2, were found to have altered functional brain connectivity using functional MRI. Here, we scanned 328 young adults with high-field (4-Tesla) diffusion imaging, to test the hypothesis that carriers of this gene variant would have altered structural brain connectivity. All participants (209 women, 119 men, age: 23.4±2.17 SD years) were scanned with 105-gradient high-angular-resolution diffusion imaging (HARDI) at 4 Tesla. After performing a whole-brain fiber tractography using the full angular resolution of the diffusion scans, 70 cortical surface-based regions of interest were created from each individual's co-registered anatomical data to compute graph metrics for all pairs of cortical regions. In graph theory analyses, subjects homozygous for the risk allele (CC) had lower characteristic path length, greater small-worldness and global efficiency in whole-brain analyses, and lower [corrected] eccentricity (maximum path length) in 60 of the 70 nodes in regional analyses. These results were not reducible to differences in more commonly studied traits such as fiber density or fractional anisotropy. This is the first study that links graph theory metrics of brain structural connectivity to a common genetic variant linked with autism and will help us understand the neurobiology of the circuits implicated in the risk for autism.