An open-source tool for longitudinal whole-brain and white matter lesion segmentation.

In this paper we describe and validate a longitudinal method for whole-brain segmentation of longitudinal MRI scans. It builds upon an existing whole-brain segmentation method that can handle multi-contrast data and robustly analyze images with white matter lesions. This method is here extended with subject-specific latent variables that encourage temporal consistency between its segmentation results, enabling it to better track subtle morphological changes in dozens of neuroanatomical structures and white matter lesions. We validate the proposed method on multiple datasets of control subjects and patients suffering from Alzheimer's disease and multiple sclerosis, and compare its results against those obtained with its original cross-sectional formulation and two benchmark longitudinal methods. The results indicate that the method attains a higher test-retest reliability, while being more sensitive to longitudinal disease effect differences between patient groups. An implementation is publicly available as part of the open-source neuroimaging package FreeSurfer.

Learn2Reg: Comprehensive Multi-Task Medical Image Registration Challenge, Dataset and Evaluation in the Era of Deep Learning.

Image registration is a fundamental medical image analysis task, and a wide variety of approaches have been proposed. However, only a few studies have comprehensively compared medical image registration approaches on a wide range of clinically relevant tasks. This limits the development of registration methods, the adoption of research advances into practice, and a fair benchmark across competing approaches. The Learn2Reg challenge addresses these limitations by providing a multi-task medical image registration data set for comprehensive characterisation of deformable registration algorithms. A continuous evaluation will be possible at https://learn2reg.grand-challenge.org. Learn2Reg covers a wide range of anatomies (brain, abdomen, and thorax), modalities (ultrasound, CT, MR), availability of annotations, as well as intra- and inter-patient registration evaluation. We established an easily accessible framework for training and validation of 3D registration methods, which enabled the compilation of results of over 65 individual method submissions from more than 20 unique teams. We used a complementary set of metrics, including robustness, accuracy, plausibility, and runtime, enabling unique insight into the current state-of-the-art of medical image registration. This paper describes datasets, tasks, evaluation methods and results of the challenge, as well as results of further analysis of transferability to new datasets, the importance of label supervision, and resulting bias. While no single approach worked best across all tasks, many methodological aspects could be identified that push the performance of medical image registration to new state-of-the-art performance. Furthermore, we demystified the common belief that conventional registration methods have to be much slower than deep-learning-based methods.

Learning the Effect of Registration Hyperparameters with HyperMorph.

We introduce HyperMorph, a framework that facilitates efficient hyperparameter tuning in learning-based deformable image registration. Classical registration algorithms perform an iterative pair-wise optimization to compute a deformation field that aligns two images. Recent learning-based approaches leverage large image datasets to learn a function that rapidly estimates a deformation for a given image pair. In both strategies, the accuracy of the resulting spatial correspondences is strongly influenced by the choice of certain hyperparameter values. However, an effective hyperparameter search consumes substantial time and human effort as it often involves training multiple models for different fixed hyperparameter values and may lead to suboptimal registration. We propose an amortized hyperparameter learning strategy to alleviate this burden by learning the impact of hyperparameters on deformation fields. We design a meta network, or hypernetwork, that predicts the parameters of a registration network for input hyperparameters, thereby comprising a single model that generates the optimal deformation field corresponding to given hyperparameter values. This strategy enables fast, high-resolution hyperparameter search at test-time, reducing the inefficiency of traditional approaches while increasing flexibility. We also demonstrate additional benefits of HyperMorph, including enhanced robustness to model initialization and the ability to rapidly identify optimal hyperparameter values specific to a dataset, image contrast, task, or even anatomical region, all without the need to retrain models. We make our code publicly available at http://hypermorph.voxelmorph.net.

SynthStrip: skull-stripping for any brain image.

The removal of non-brain signal from magnetic resonance imaging (MRI) data, known as skull-stripping, is an integral component of many neuroimage analysis streams. Despite their abundance, popular classical skull-stripping methods are usually tailored to images with specific acquisition properties, namely near-isotropic resolution and T1-weighted (T1w) MRI contrast, which are prevalent in research settings. As a result, existing tools tend to adapt poorly to other image types, such as stacks of thick slices acquired with fast spin-echo (FSE) MRI that are common in the clinic. While learning-based approaches for brain extraction have gained traction in recent years, these methods face a similar burden, as they are only effective for image types seen during the training procedure. To achieve robust skull-stripping across a landscape of imaging protocols, we introduce SynthStrip, a rapid, learning-based brain-extraction tool. By leveraging anatomical segmentations to generate an entirely synthetic training dataset with anatomies, intensity distributions, and artifacts that far exceed the realistic range of medical images, SynthStrip learns to successfully generalize to a variety of real acquired brain images, removing the need for training data with target contrasts. We demonstrate the efficacy of SynthStrip for a diverse set of image acquisitions and resolutions across subject populations, ranging from newborn to adult. We show substantial improvements in accuracy over popular skull-stripping baselines - all with a single trained model. Our method and labeled evaluation data are available at https://w3id.org/synthstrip.

TopoFit: Rapid Reconstruction of Topologically-Correct Cortical Surfaces.

Mesh-based reconstruction of the cerebral cortex is a fundamental component in brain image analysis. Classical, iterative pipelines for cortical modeling are robust but often time-consuming, mostly due to expensive procedures that involve topology correction and spherical mapping. Recent attempts to address reconstruction with machine learning methods have accelerated some components in these pipelines, but these methods still require slow processing steps to enforce topological constraints that comply with known anatomical structure. In this work, we introduce a novel learning-based strategy, TopoFit, which rapidly fits a topologically-correct surface to the white-matter tissue boundary. We design a joint network, employing image and graph convolutions and an efficient symmetric distance loss, to learn to predict accurate deformations that map a template mesh to subject-specific anatomy. This technique encompasses the work of current mesh correction, fine-tuning, and inflation processes and, as a result, offers a 150× faster solution to cortical surface reconstruction compared to traditional approaches. We demonstrate that TopoFit is 1.8× more accurate than the current state-of-the-art deep-learning strategy, and it is robust to common failure modes, such as white-matter tissue hypointensities.

A deep learning toolbox for automatic segmentation of subcortical limbic structures from MRI images.

A tool was developed to automatically segment several subcortical limbic structures (nucleus accumbens, basal forebrain, septal nuclei, hypothalamus without mammillary bodies, the mammillary bodies, and fornix) using only a T1-weighted MRI as input. This tool fills an unmet need as there are few, if any, publicly available tools to segment these clinically relevant structures. A U-Net with spatial, intensity, contrast, and noise augmentation was trained using 39 manually labeled MRI data sets. In general, the Dice scores, true positive rates, false discovery rates, and manual-automatic volume correlation were very good relative to comparable tools for other structures. A diverse data set of 698 subjects were segmented using the tool; evaluation of the resulting labelings showed that the tool failed in less than 1% of cases. Test-retest reliability of the tool was excellent. The automatically segmented volume of all structures except mammillary bodies showed effectiveness at detecting either clinical AD effects, age effects, or both. This tool will be publicly released with FreeSurfer (surfer.nmr.mgh.harvard.edu/fswiki/ScLimbic). Together with the other cortical and subcortical limbic segmentations, this tool will allow FreeSurfer to provide a comprehensive view of the limbic system in an automated way.

PSACNN: Pulse sequence adaptive fast whole brain segmentation.

With the advent of convolutional neural networks (CNN), supervised learning methods are increasingly being used for whole brain segmentation. However, a large, manually annotated training dataset of labeled brain images required to train such supervised methods is frequently difficult to obtain or create. In addition, existing training datasets are generally acquired with a homogeneous magnetic resonance imaging (MRI) acquisition protocol. CNNs trained on such datasets are unable to generalize on test data with different acquisition protocols. Modern neuroimaging studies and clinical trials are necessarily multi-center initiatives with a wide variety of acquisition protocols. Despite stringent protocol harmonization practices, it is very difficult to standardize the gamut of MRI imaging parameters across scanners, field strengths, receive coils etc., that affect image contrast. In this paper we propose a CNN-based segmentation algorithm that, in addition to being highly accurate and fast, is also resilient to variation in the input acquisition. Our approach relies on building approximate forward models of pulse sequences that produce a typical test image. For a given pulse sequence, we use its forward model to generate plausible, synthetic training examples that appear as if they were acquired in a scanner with that pulse sequence. Sampling over a wide variety of pulse sequences results in a wide variety of augmented training examples that help build an image contrast invariant model. Our method trains a single CNN that can segment input MRI images with acquisition parameters as disparate as T1-weighted and T2-weighted contrasts with only T1-weighted training data. The segmentations generated are highly accurate with state-of-the-art results (overall Dice overlap=0.94), with a fast run time (≈ 45 s), and consistent across a wide range of acquisition protocols.