Several studies underscore the potential of deep learning in identifying complex patterns, leading to diagnostic and prognostic biomarkers. Identifying sufficiently large and diverse datasets, required for training, is a significant challenge in medicine and can rarely be found in individual institutions. Multi-institutional collaborations based on centrally-shared patient data face privacy and ownership challenges. federated learning is a novel paradigm for data-private multi-institutional collaborations, where model-learning leverages all available data without sharing data between institutions, by distributing the model-training to the data-owners and aggregating their results. We show that federated learning among 10 institutions results in models reaching 99% of the model quality achieved with centralized data, and evaluate generalizability on data from institutions outside the federation. We further investigate the effects of data distribution across collaborating institutions on model quality and learning patterns, indicating that increased access to data through data private multi-institutional collaborations can benefit model quality more than the errors introduced by the collaborative method. finally, we compare with other collaborative-learning approaches demonstrating the superiority of federated learning, and discuss practical implementation considerations. clinical adoption of federated learning is expected to lead to models trained on datasets of unprecedented size, hence have a catalytic impact towards precision/personalized medicine. Abbreviations CDS Collaborative data sharing FL Federated learning IIL Institutional incremental learning CIIL Cyclic institutional incremental learning IID Independent and identically distributed BraTS Brain tumor segmentation

† People involved in the organization of the challenge. ‡ People contributing data from their institutions.§ Equal senior authors.

The identifying or sensitive anatomical features in MR and CT images used in research raise patient privacy concerns when such data are shared. In order to protect human subject privacy, we developed a method of anatomical surface modification and investigated the effects of such modification on image statistics and common neuroimaging processing tools. Common approaches to obscuring facial features typically remove large portions of the voxels. The approach described here focuses on blurring the anatomical surface instead, to avoid impinging on areas of interest and hard edges that can confuse processing tools. The algorithm proceeds by extracting a thin boundary layer containing surface anatomy from a region of interest. This layer is then “stretched” and “flattened” to fit into a thin “box” volume. After smoothing along a plane roughly parallel to anatomy surface, this volume is transformed back onto the boundary layer of the original data. The above method, named normalized anterior filtering, was coded in MATLAB and applied on a number of high resolution MR and CT scans. To test its effect on automated tools, we compared the output of selected common skull stripping and MR gain field correction methods used on unmodified and obscured data. With this paper, we hope to improve the understanding of the effect of surface deformation approaches on the quality of de-identified data and to provide a useful de-identification tool for MR and CT acquisitions.

Brain extraction, or skull-stripping, is an essential pre-processing step in neuro-imaging that has a direct impact on the quality of all subsequent processing and analyses steps. It is also a key requirement in multi-institutional collaborations to comply with privacy-preserving regulations. Existing automated methods, including Deep Learning (DL) based methods that have obtained state-of-the-art results in recent years, have primarily targeted brain extraction without considering pathologically-affected brains. Accordingly, they perform sub-optimally when applied on magnetic resonance imaging (MRI) brain scans with apparent pathologies such as brain tumors. Furthermore, existing methods focus on using only T1-weighted MRI scans, even though multi-parametric MRI (mpMRI) scans are routinely acquired for patients with suspected brain tumors. In this study, we present a comprehensive performance evaluation of recent deep learning architectures for brain extraction, training models on mpMRI scans of pathologically-affected brains, with a particular focus on seeking a practically-applicable, low computational footprint approach, generalizable across multiple institutions, further facilitating collaborations. We identified a large retrospective multi-institutional dataset of n = 3340 mpMRI brain tumor scans, with manually-inspected and approved gold-standard segmentations, acquired during standard clinical practice under varying acquisition protocols, both from private institutional data and public (TCIA) collections. To facilitate optimal utilization of rich mpMRI data, we further introduce and evaluate a novel “modality-agnostic training” technique that can be applied using any available modality, without need for model retraining. Our results indicate that the modality-agnostic approach 1 obtains accurate results, providing a generic and practical tool for brain extraction on scans with brain tumors.

Synopsis The goal of this article is to compare resting state fMRI with task fMRI as a tool for presurgical functional mapping of the sensorimotor (SM) region. Prior to tumor resection, 38 patients were scanned using both methods. The SM area was anatomically defined using two different software tools. Overlap of anatomical regions of interest with task activation maps and resting state networks was measured in the SM region. A paired t-test showed higher overlap between resting state maps and anatomical references as compared to task activation when using a maximal overlap criterion. Resting state derived maps are more comprehensive than those derived from task fMRI.

Functional magnetic resonance imaging (fMRI) is an important tool for pre-surgical evaluation of eloquent cortex. Classic task-based paradigms require patient participation and individual imaging sequence acquisitions for each functional domain that is being assessed. Resting state fMRI (rs-fMRI), however, enables functional localization without patient participation and can evaluate numerous functional domains with a single imaging session. To date, post-processing of this resting state data has been resource intensive, which limits its widespread application for routine clinical use. Through a novel automated algorithm and advanced imaging IT structure, we report the clinical application and the large-scale integration of rs-fMRI into routine neurosurgical practice. One hundred and ninety one consecutive patients underwent a 3T rs-fMRI, 83 of whom also underwent both motor and language task-based fMRI. Data were processed using a novel, automated, multi-layer perceptron algorithm and integrated into stereotactic navigation using a streamlined IT imaging pipeline. One hundred eighty-five studies were performed for intracranial neoplasm, 14 for refractory epilepsy and 33 for vascular malformations or other neurological disorders. Failure rate of rs-fMRI of 13% was significantly better than that for task-based fMRI (38.5%,) (p <0.001). In conclusion, at Washington University in St. Louis, rs-fMRI has become an integral part of standard imaging for neurosurgical planning. Resting state fMRI can be used in all patients, and due to its lower failure rate than task-based fMRI, it is useful for patients who are unable to cooperate with task-based studies.

Background Use of functional MRI (fMRI) in pre-surgical planning is a non-invasive method for pre-operative functional mapping for patients with brain tumors, especially tumors located near eloquent cortex. Currently, this practice predominantly involves task-based fMRI (T-fMRI). Resting state fMRI (RS-fMRI) offers an alternative with several methodological advantages. Here, we compare group-level analyses of RS-fMRI vs. T-fMRI as methods for language localization. Purpose To contrast RS-fMRI vs. T-fMRI as techniques for localization of language function. Methods We analyzed data obtained in 35 patients who had both T-fMRI and RS-fMRI scans during the course of pre-surgical evaluation. The RS-fMRI data were analyzed using a previously trained resting-state network classifier. The T-fMRI data were analyzed using conventional techniques. Group-level results obtained by both methods were evaluated in terms of two outcome measures: (1) inter-subject variability of response magnitude and (2) sensitivity/specificity analysis of response topography, taking as ground truth previously reported maps of the language system based on intraoperative cortical mapping as well as meta-analytic maps of language task fMRI responses. Results Both fMRI methods localized major components of the language system (areas of Broca and Wernicke) although not with equal inter-subject consistency. Word-stem completion T-fMRI strongly activated Broca's area but also several task-general areas not specific to language. RS-fMRI provided a more specific representation of the language system. Conclusion We demonstrate several advantages of classifier-based mapping of language representation in the brain. Language T-fMRI activated task-general (i.e., not language-specific) functional systems in addition to areas of Broca and Wernicke. In contrast, classifier-based analysis of RS-fMRI data generated maps confined to language-specific regions of the brain.

Background Deep brain stimulation of the subthalamic nucleus is a widely used adjunctive therapy for motor symptoms of Parkinson's disease, but with variable motor response. Predicting motor response remains difficult, and novel approaches may improve surgical outcomes as well as the understanding of pathophysiological mechanisms. The objective of this study was to determine whether preoperative resting‐state functional connectivity MRI predicts motor response from deep brain stimulation of the subthalamic nucleus. Methods We collected preoperative resting‐state functional MRI from 70 participants undergoing subthalamic nucleus deep brain stimulation. For this cohort, we analyzed the strength of STN functional connectivity with seeds determined by stimulation‐induced (ON/OFF) 15O H2O PET regional cerebral blood flow differences in a partially overlapping group (n = 42). We correlated STN‐seed functional connectivity strength with postoperative motor outcomes and applied linear regression to predict motor outcomes. Results Preoperative functional connectivity between the left subthalamic nucleus and the ipsilateral internal globus pallidus correlated with postsurgical motor outcomes (r = −0.39, P = 0.0007), with stronger preoperative functional connectivity relating to greater improvement. Left pallidal‐subthalamic nucleus connectivity also predicted motor response to DBS after controlling for covariates. Discussion Preoperative pallidal‐subthalamic nucleus resting‐state functional connectivity predicts motor benefit from deep brain stimulation, although this should be validated prospectively before clinical application. These observations suggest that integrity of pallidal‐subthalamic nucleus circuits may be critical to motor benefits from deep brain stimulation. © 2020 International Parkinson and Movement Disorder Society