For the past 25 years, the field of neuroimaging has witnessed the development of several software packages for processing multi-parametric magnetic resonance imaging (mpMRI) to study the brain. ...These software packages are now routinely used by researchers and clinicians, and have contributed to important breakthroughs for the understanding of brain anatomy and function. However, no software package exists to process mpMRI data of the spinal cord. Despite the numerous clinical needs for such advanced mpMRI protocols (multiple sclerosis, spinal cord injury, cervical spondylotic myelopathy, etc.), researchers have been developing specific tools that, while necessary, do not provide an integrative framework that is compatible with most usages and that is capable of reaching the community at large. This hinders cross-validation and the possibility to perform multi-center studies. In this study we introduce the Spinal Cord Toolbox (SCT), a comprehensive software dedicated to the processing of spinal cord MRI data. SCT builds on previously-validated methods and includes state-of-the-art MRI templates and atlases of the spinal cord, algorithms to segment and register new data to the templates, and motion correction methods for diffusion and functional time series. SCT is tailored towards standardization and automation of the processing pipeline, versatility, modularity, and it follows guidelines of software development and distribution. Preliminary applications of SCT cover a variety of studies, from cross-sectional area measures in large databases of patients, to the precise quantification of mpMRI metrics in specific spinal pathways. We anticipate that SCT will bring together the spinal cord neuroimaging community by establishing standard templates and analysis procedures.
Display omitted
•SCT (Spinal Cord Toolbox): Software package for processing spinal cord MRI data.•Features Templates & atlases of spinal cord, gray matter and white matter tracts.•State-of-the-art segmentation, registration and atlas-based analysis methods.•Open-source, extensive testing framework, documentation and support via forum.•Enables standardized, automatic, robust and reproducible multi-center studies of large datasets.
Linear registration to a standard space is one of the major steps in processing and analyzing magnetic resonance images (MRIs) of the brain. Here we present an overview of linear stereotaxic MRI ...registration and compare the performance of 5 publicly available and extensively used linear registration techniques in medical image analysis.
A set of 9693 T1-weighted MR images were obtained for testing from 4 datasets: ADNI, PREVENT-AD, PPMI, and HCP, two of which have multi-center and multi-scanner data and three of which have longitudinal data. Each individual native image was linearly registered to the MNI ICBM152 average template using five versions of MRITOTAL from MINC tools, FLIRT from FSL, two versions of Elastix, spm_affreg from SPM, and ANTs linear registration techniques. Quality control (QC) images were generated from the registered volumes and viewed by an expert rater to assess the quality of the registrations. The QC image contained 60 sub-images (20 of each of axial, sagittal, and coronal views at different levels throughout the brain) overlaid with contours of the ICBM152 template, enabling the expert rater to label the registration as acceptable or unacceptable. The performance of the registration techniques was then compared across different datasets. In addition, the effect of image noise, intensity non-uniformity, age, head size, and atrophy on the performance of the techniques was investigated by comparing differences between age, scaling factor, ventricle volume, brain volume, and white matter hyperintensity (WMH) volumes between passed and failed cases for each method.
The average registration failure rate among all datasets was 27.41%, 27.14%, 12.74%, 13.03%, 0.44% for the five versions of MRITOTAL techniques, 8.87% for ANTs, 11.11% for FSL, 12.35% for Elastix Affine, 24.40% for Elastix Similarity, and 30.66% for SPM. There were significant effects of signal to noise ratio, image intensity non-uniformity estimates, as well as age, head size, and atrophy related changes between passed and failed registrations.
Our experiments show that the Revised BestLinReg had the best performance among the evaluated registration techniques while all techniques performed worse for images with higher levels of noise and non-uniformity as well as atrophy related changes.
•Comparison of 5 publicly available linear registration methods•9693 T1-weighted volumes used for evaluation•4 different multi-center and multi-scanner datasets•Effect of age, atrophy, image SNR, and non-uniformity on registration evaluated
Abstract Optimized magnetic resonance imaging (MRI)–based biomarkers of Alzheimer's disease (AD) may allow earlier detection and refined prediction of the disease. In addition, they could serve as ...valuable tools when designing therapeutic studies of individuals at risk of AD. In this study, we combine (1) a novel method for grading medial temporal lobe structures with (2) robust cortical thickness measurements to predict AD among subjects with mild cognitive impairment (MCI) from a single T1-weighted MRI scan. Using AD and cognitively normal individuals, we generate a set of features potentially discriminating between MCI subjects who convert to AD and those who remain stable over a period of 3 years. Using mutual information-based feature selection, we identify 5 key features optimizing the classification of MCI converters. These features are the left and right hippocampi gradings and cortical thicknesses of the left precuneus, left superior temporal sulcus, and right anterior part of the parahippocampal gyrus. We show that these features are highly stable in cross-validation and enable a prediction accuracy of 72% using a simple linear discriminant classifier, the highest prediction accuracy obtained on the baseline Alzheimer's Disease Neuroimaging Initiative first phase cohort to date. The proposed structural features are consistent with Braak stages and previously reported atrophic patterns in AD and are easy to transfer to new cohorts and to clinical practice.
Template-based analysis of multi-parametric MRI data of the spinal cord sets the foundation for standardization and reproducibility, thereby helping the discovery of new biomarkers of spinal-related ...diseases. While MRI templates of the spinal cord have been recently introduced, none of them cover the entire spinal cord. In this study, we introduced an unbiased multimodal MRI template of the spinal cord and the brainstem, called PAM50, which is anatomically compatible with the ICBM152 brain template and uses the same coordinate system. The PAM50 template is based on 50 healthy subjects, covers the full spinal cord (C1 to L2 vertebral levels) and the brainstem, is available for T1-, T2-and T2*-weighted MRI contrasts and includes a probabilistic atlas of the gray matter and white matter tracts. Template creation accuracy was assessed by computing the mean and maximum distance error between each individual spinal cord centerline and the PAM50 centerline, after registration to the template. Results showed high accuracy for both T1- (mean = 0.37 ± 0.06 mm; max = 1.39 ± 0.58 mm) and T2-weighted (mean = 0.11 ± 0.03 mm; max = 0.71 ± 0.27 mm) contrasts. Additionally, the preservation of the spinal cord topology during the template creation process was verified by comparing the cross-sectional area (CSA) profile, averaged over all subjects, and the CSA profile of the PAM50 template. The fusion of the PAM50 and ICBM152 templates will facilitate group and multi-center studies of combined brain and spinal cord MRI, and enable the use of existing atlases of the brainstem compatible with the ICBM space.
•PAM50 is an MRI template of the full spinal cord and brainstem.•PAM50 is available for T1-, T2-and T2*-weighted MRI contrast.•PAM50 is compatible with the ICBM152 brain template, allowing cerebrospinal studies.•PAM50 includes atlases of white matter pathways and gray matter subregions.•PAM50 is available in SCT, an open-source software for processing spinal cord MRI data.
Background
Harmonized protocols to collect imaging data must be devised, employed, and maintained in multicentric studies to reduce interscanner variability in subsequent analyses.
Purpose
To present ...a standardized protocol for multicentric research on dementia linked to neurodegeneration in aging, harmonized on all three major vendor platforms. The protocol includes a common procedure for qualification, quality control, and quality assurance and feasibility in large‐scale studies.
Study Type
Prospective.
Subjects
The study involved a geometric phantom, a single individual volunteer, and 143 cognitively healthy, mild cognitively impaired, and Alzheimer's disease participants in a large‐scale, multicentric study.
Field Strength/Sequences
MRI was perform with 3T scanners (GE, Philips, Siemens) and included 3D T1w, PD/T2w,
T2*, T2w‐FLAIR, diffusion, and BOLD resting state acquisitions.
Assessment
Measures included signal‐ and contrast‐to‐noise ratios (SNR and CNR, respectively), total brain volumes, and total scan time.
Statistical Tests
SNR, CNR, and scan time were compared between scanner vendors using analysis of variance (ANOVA) and Tukey tests, while brain volumes were tested using linear mixed models.
Results
Geometric phantom T1w SNR was significantly (P < 0.001) higher in Philips (mean: 71.4) than Siemens (29.5), while no significant difference was observed between vendors for T2w (32.0 and 37.2, respectively, P = 0.243). Single individual volunteer T1w CNR was higher in subcortical regions for Siemens (P < 0.001), while Philips had higher cortical CNR (P = 0.044). No significant difference in brain volumes was observed between vendors (P = 0.310/0.582/0.055). The average scan time was 41.0 minutes (SD: 2.8) and was not significantly different between sites (P = 0.071) and cognitive groups (P = 0.853).
Data Conclusion
The harmonized Canadian Dementia Imaging Protocol suits the needs of studies that need to ensure quality MRI data acquisition for the measurement of brain changes across adulthood, due to aging, neurodegeneration, and other etiologies. A detailed description, exam cards, and operators' manual are freely available at the following site: www.cdip-pcid.ca.
Level of Evidence: 2
Technical Efficacy: Stage 2
J. Magn. Reson. Imaging 2019;49:456–465.
White matter hyperintensities (WMHs) are areas of abnormal signal on magnetic resonance images (MRIs) that characterize various types of histopathological lesions. The load and location of WMHs are ...important clinical measures that may indicate the presence of small vessel disease in aging and Alzheimer's disease (AD) patients. Manually segmenting WMHs is time consuming and prone to inter-rater and intra-rater variabilities. Automated tools that can accurately and robustly detect these lesions can be used to measure the vascular burden in individuals with AD or the elderly population in general. Many WMH segmentation techniques use a classifier in combination with a set of intensity and location features to segment WMHs, however, the optimal choice of classifier is unknown.
We compare 10 different linear and nonlinear classification techniques to identify WMHs from MRI data. Each classifier is trained and optimized based on a set of features obtained from co-registered MR images containing spatial location and intensity information. We further assess the performance of the classifiers using different combinations of MRI contrast information. The performances of the different classifiers were compared on three heterogeneous multi-site datasets, including images acquired with different scanners and different scan-parameters. These included data from the ADC study from University of California Davis, the NACC database and the ADNI study. The classifiers (naïve Bayes, logistic regression, decision trees, random forests, support vector machines, k-nearest neighbors, bagging, and boosting) were evaluated using a variety of voxel-wise and volumetric similarity measures such as Dice Kappa similarity index (SI), Intra-Class Correlation (ICC), and sensitivity as well as computational burden and processing times. These investigations enable meaningful comparisons between the performances of different classifiers to determine the most suitable classifiers for segmentation of WMHs. In the spirit of open-source science, we also make available a fully automated tool for segmentation of WMHs with pre-trained classifiers for all these techniques.
Random Forests yielded the best performance among all classifiers with mean Dice Kappa (SI) of 0.66±0.17 and ICC=0.99 for the ADC dataset (using T1w, T2w, PD, and FLAIR scans), SI=0.72±0.10, ICC=0.93 for the NACC dataset (using T1w and FLAIR scans), SI=0.66±0.23, ICC=0.94 for ADNI1 dataset (using T1w, T2w, and PD scans) and SI=0.72±0.19, ICC=0.96 for ADNI2/GO dataset (using T1w and FLAIR scans). Not using the T2w/PD information did not change the performance of the Random Forest classifier (SI=0.66±0.17, ICC=0.99). However, not using FLAIR information in the ADC dataset significantly decreased the Dice Kappa, but the volumetric correlation did not drastically change (SI=0.47±0.21, ICC=0.95).
Our investigations showed that with appropriate features, most off-the-shelf classifiers are able to accurately detect WMHs in presence of FLAIR scan information, while Random Forests had the best performance across all datasets. However, we observed that the performances of most linear classifiers and some nonlinear classifiers drastically decline in absence of FLAIR information, with Random Forest still retaining the best performance.
Blood‐flow artifacts present a serious challenge for most, if not all, volumetric analytical approaches. We utilize T1‐weighted data with prominent blood‐flow artifacts from the Autism Brain Imaging ...Data Exchange (ABIDE) multisite agglomerative dataset to assess the impact that such blood‐flow artifacts have on registration of T1‐weighted data to a template. We use a heuristic approach to identify the blood‐flow artifacts in these data; we use the resulting blood masks to turn the underlying voxels to the intensity of the cerebro‐spinal fluid, thus mimicking the effect of blood suppression. We then register both the original data and the deblooded data to a common T1‐weighted template, and compare the quality of those registrations to the template in terms of similarity to the template. The registrations to the template based on the deblooded data yield significantly higher similarity values compared with those based on the original data. Additionally, we measure the nonlinear deformations needed to transform the data from the position achieved by registering the original data to the template to the position achieved by registering the deblooded data to the template. The results indicate that blood‐flow artifacts may seriously impact data processing that depends on registration to a template, that is, most all data processing.
Blood‐flow artifacts present a serious challenge for most, if not all, volumetric analytical approaches, and even surface‐based approaches. We utilize T1‐weighted data with prominent blood‐flow artifacts from the Autism Brain Imaging Data Exchange (ABIDE) multisite agglomerative dataset to assess the impact that such blood‐flow artifacts have on registration of T1‐weighted data to a template. We use a heuristic approach to identify the blood‐flow artifacts in these data; we then use the resulting blood masks to mimick the effect of blood suppression. The registrations to the template based on the deblooded data yield significantly higher similarity values compared with those based on the original data. The results indicate that blood‐flow artifacts may seriously impact data processing that depends on registration to a template, that is, most all data processing.
Identifying individuals destined to develop Alzheimer's dementia within time frames acceptable for clinical trials constitutes an important challenge to design studies to test emerging ...disease-modifying therapies. Although amyloid-β protein is the core pathologic feature of Alzheimer's disease, biomarkers of neuronal degeneration are the only ones believed to provide satisfactory predictions of clinical progression within short time frames. Here, we propose a machine learning–based probabilistic method designed to assess the progression to dementia within 24 months, based on the regional information from a single amyloid positron emission tomography scan. Importantly, the proposed method was designed to overcome the inherent adverse imbalance proportions between stable and progressive mild cognitive impairment individuals within a short observation period. The novel algorithm obtained an accuracy of 84% and an under-receiver operating characteristic curve of 0.91, outperforming the existing algorithms using the same biomarker measures and previous studies using multiple biomarker modalities. With its high accuracy, this algorithm has immediate applications for population enrichment in clinical trials designed to test disease-modifying therapies aiming to mitigate the progression to Alzheimer's disease dementia.
Linear registration to stereotaxic space is a common first step in many automated image-processing tools for analysis of human brain MRI scans. This step is crucial for the success of the subsequent ...image-processing steps. Several well-established algorithms are commonly used in the field of neuroimaging for this task, but none have a 100% success rate. Manual assessment of the registration is commonly used as part of quality control. To reduce the burden of this time-consuming step, we propose Deep Automated Registration Qc (DARQ), a fully automatic quality control method based on deep learning that can replace the human rater and accurately perform quality control assessment for stereotaxic registration of T1w brain scans.
In a recently published study from our group comparing linear registration methods, we used a database of 9325 MRI scans and 64476 registrations from several publicly available datasets and applied seven linear registration tools to them. In this study, the resulting images that were assessed and labeled by a human rater are used to train a deep neural network to detect cases when registration failed. We further validated the results on an independent dataset of patients with multiple sclerosis, with manual QC labels available (n=1200).
In terms of agreement with a manual rater, our automated QC method was able to achieve 89% accuracy and 85% true negative rate (equivalently 15% false positive rate) in detecting scans that should pass quality control in a balanced cross-validation experiments, and 96.1% accuracy and 95.5% true negative rate (or 4.5% FPR) when evaluated in a balanced independent sample, similar to manual QC rater (test-retest accuracy of 93%).
The results show that DARQ is robust, fast, accurate, and generalizable in detecting failure in linear stereotaxic registrations and can substantially reduce QC time (by a factor of 20 or more) when processing large datasets.
In this paper, we propose an innovative approach to robustly and accurately detect Alzheimer's disease (AD) based on the distinction of specific atrophic patterns of anatomical structures such as ...hippocampus (HC) and entorhinal cortex (EC). The proposed method simultaneously performs segmentation and grading of structures to efficiently capture the anatomical alterations caused by AD. Known as SNIPE (Scoring by Non-local Image Patch Estimator), the novel proposed grading measure is based on a nonlocal patch-based frame-work and estimates the similarity of the patch surrounding the voxel under study with all the patches present in different training populations. In this study, the training library was composed of two populations: 50 cognitively normal subjects (CN) and 50 patients with AD, randomly selected from the ADNI database. During our experiments, the classification accuracy of patients (CN vs. AD) using several biomarkers was compared: HC and EC volumes, the grade of these structures and finally the combination of their volume and their grade. Tests were completed in a leave-one-out framework using discriminant analysis. First, we showed that biomarkers based on HC provide better classification accuracy than biomarkers based on EC. Second, we demonstrated that structure grading is a more powerful measure than structure volume to distinguish both populations with a classification accuracy of 90%. Finally, by adding the ages of subjects in order to better separate age-related structural changes from disease-related anatomical alterations, SNIPE obtained a classification accuracy of 93%.
► A new patch-based biomarker is proposed for automatic patient's classification. ► Nonlocal estimator is used to simultaneously segment and grade anatomical structures. ► Validation is carried out on entorhinal cortex and hippocampus of 100 subjects. ► Comparison of several biomarkers demonstrates advantages of grading measure. ► Grading enables accurate detection of structural modifications caused by a disease.