Esther E. Bron

Standardized evaluation of algorithms for computer-aided diagnosis of dementia based on structural MRI: The CADDementia challenge

Algorithms for computer-aided diagnosis of dementia based on structural MRI have demonstrated high performance in the literature, but are difficult to compare as different data sets and methodology were used for evaluation. In addition, it is unclear how the algorithms would perform on previously unseen data, and thus, how they would perform in clinical practice when there is no real opportunity to adapt the algorithm to the data at hand. To address these comparability, generalizability and clinical applicability issues, we organized a grand challenge that aimed to objectively compare algorithms based on a clinically representative multi-center data set. Using clinical practice as the starting point, the goal was to reproduce the clinical diagnosis. Therefore, we evaluated algorithms for multi-class classification of three diagnostic groups: patients with probable Alzheimers disease, patients with mild cognitive impairment and healthy controls. The diagnosis based on clinical criteria was used as reference standard, as it was the best available reference despite its known limitations. For evaluation, a previously unseen test set was used consisting of 354 T1-weighted MRI scans with the diagnoses blinded. Fifteen research teams participated with a total of 29 algorithms. The algorithms were trained on a small training set (n=30) and optionally on data from other sources (e.g., the Alzheimers Disease Neuroimaging Initiative, the Australian Imaging Biomarkers and Lifestyle flagship study of aging). The best performing algorithm yielded an accuracy of 63.0% and an area under the receiver-operating-characteristic curve (AUC) of 78.8%. In general, the best performances were achieved using feature extraction based on voxel-based morphometry or a combination of features that included volume, cortical thickness, shape and intensity. The challenge is open for new submissions via the web-based framework: http://caddementia.grand-challenge.org.

Fast Parallel Image Registration on CPU and GPU for Diagnostic Classification of Alzheimer's Disease

Nonrigid image registration is an important, but time-consuming task in medical image analysis. In typical neuroimaging studies, multiple image registrations are performed, i.e., for atlas-based segmentation or template construction. Faster image registration routines would therefore be beneficial. In this paper we explore acceleration of the image registration package elastix by a combination of several techniques: (i) parallelization on the CPU, to speed up the cost function derivative calculation; (ii) parallelization on the GPU building on and extending the OpenCL framework from ITKv4, to speed up the Gaussian pyramid computation and the image resampling step; (iii) exploitation of certain properties of the B-spline transformation model; (iv) further software optimizations. The accelerated registration tool is employed in a study on diagnostic classification of Alzheimers disease and cognitively normal controls based on T1-weighted MRI. We selected 299 participants from the publicly available Alzheimers Disease Neuroimaging Initiative database. Classification is performed with a support vector machine based on gray matter volumes as a marker for atrophy. We evaluated two types of strategies (voxel-wise and region-wise) that heavily rely on nonrigid image registration. Parallelization and optimization resulted in an acceleration factor of 4–5x on an 8-core machine. Using OpenCL a speedup factor of 2 was realized for computation of the Gaussian pyramids, and 15–60 for the resampling step, for larger images. The voxel-wise and the region-wise classification methods had an area under the receiver operator characteristic curve of 88 and 90%, respectively, both for standard and accelerated registration. We conclude that the image registration package elastix was substantially accelerated, with nearly identical results to the non-optimized version. The new functionality will become available in the next release of elastix as open source under the BSD license.

Diagnostic classification of arterial spin labeling and structural MRI in presenile early stage dementia

Because hypoperfusion of brain tissue precedes atrophy in dementia, the detection of dementia may be advanced by the use of perfusion information. Such information can be obtained noninvasively with arterial spin labeling (ASL), a relatively new MR technique quantifying cerebral blood flow (CBF). Using ASL and structural MRI, we evaluated diagnostic classification in 32 prospectively included presenile early stage dementia patients and 32 healthy controls. Patients were suspected of Alzheimers disease (AD) or frontotemporal dementia. Classification was based on CBF as perfusion marker, gray matter (GM) volume as atrophy marker, and their combination. These markers were each examined using six feature extraction methods: a voxel‐wise method and a region of interest (ROI)‐wise approach using five ROI‐sets in the GM. These ROI‐sets ranged in number from 72 brain regions to a single ROI for the entire supratentorial brain. Classification was performed with a linear support vector machine classifier. For validation of the classification method on the basis of GM features, a reference dataset from the AD Neuroimaging Initiative database was used consisting of AD patients and healthy controls. In our early stage dementia population, the voxelwise feature‐extraction approach achieved more accurate results (area under the curve (AUC) range = 86 − 91%) than all other approaches (AUC = 57 − 84%). Used in isolation, CBF quantified with ASL was a good diagnostic marker for dementia. However, our findings indicated only little added diagnostic value when combining ASL with the structural MRI data (AUC = 91%), which did not significantly improve over accuracy of structural MRI atrophy marker by itself. Hum Brain Mapp 35:4916–4931, 2014.

Feature Selection Based on the SVM Weight Vector for Classification of Dementia

Computer-aided diagnosis of dementia using a support vector machine (SVM) can be improved with feature selection. The relevance of individual features can be quantified from the SVM weights as a significance map (p-map). Although these p-maps previously showed clusters of relevant voxels in dementia-related brain regions, they have not yet been used for feature selection. Therefore, we introduce two novel feature selection methods based on p-maps using a direct approach (filter) and an iterative approach (wrapper). To evaluate these p-map feature selection methods, we compared them with methods based on the SVM weight vector directly, t-statistics, and expert knowledge. We used MRI data from the Alzheimers disease neuroimaging initiative classifying Alzheimers disease (AD) patients, mild cognitive impairment (MCI) patients who converted to AD (MCIc), MCI patients who did not convert to AD (MCInc), and cognitively normal controls (CN). Features for each voxel were derived from gray matter morphometry. Feature selection based on the SVM weights gave better results than t-statistics and expert knowledge. The p-map methods performed slightly better than those using the weight vector. The wrapper method scored better than the filter method. Recursive feature elimination based on the p-map improved most for AD-CN: the area under the receiver-operating-characteristic curve (AUC) significantly increased from 90.3% without feature selection to 92.0% when selecting 1.5%-3% of the features. This feature selection method also improved the other classifications: AD-MCI 0.1% improvement in AUC (not significant), MCI-CN 0.7%, and MCIc-MCInc 0.1% (not significant). Although the performance improvement due to feature selection was limited, the methods based on the p-map generally had the best performance, and were therefore better in estimating the relevance of individual features.

Is T1ρ Mapping an Alternative to Delayed Gadolinium-enhanced MR Imaging of Cartilage in the Assessment of Sulphated Glycosaminoglycan Content in Human Osteoarthritic Knees? An in Vivo Validation Study

PURPOSE To determine if T1ρ mapping can be used as an alternative to delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC) in the quantification of cartilage biochemical composition in vivo in human knees with osteoarthritis. MATERIALS AND METHODS This study was approved by the institutional review board. Written informed consent was obtained from all participants. Twelve patients with knee osteoarthritis underwent dGEMRIC and T1ρ mapping at 3.0 T before undergoing total knee replacement. Outcomes of dGEMRIC and T1ρ mapping were calculated in six cartilage regions of interest. Femoral and tibial cartilages were harvested during total knee replacement. Cartilage sulphated glycosaminoglycan (sGAG) and collagen content were assessed with dimethylmethylene blue and hydroxyproline assays, respectively. A four-dimensional multivariate mixed-effects model was used to simultaneously assess the correlation between outcomes of dGEMRIC and T1ρ mapping and the sGAG and collagen content of the articular cartilage. RESULTS T1 relaxation times at dGEMRIC showed strong correlation with cartilage sGAG content (r = 0.73; 95% credibility interval [CI] = 0.60, 0.83) and weak correlation with cartilage collagen content (r = 0.40; 95% CI: 0.18, 0.58). T1ρ relaxation times did not correlate with cartilage sGAG content (r = 0.04; 95% CI: -0.21, 0.28) or collagen content (r = -0.05; 95% CI = -0.31, 0.20). CONCLUSION dGEMRIC can help accurately measure cartilage sGAG content in vivo in patients with knee osteoarthritis, whereas T1ρ mapping does not appear suitable for this purpose. Although the technique is not completely sGAG specific and requires a contrast agent, dGEMRIC is a validated and robust method for quantifying cartilage sGAG content in human osteoarthritis subjects in clinical research.

Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) shows no change in cartilage structural composition after viscosupplementation in patients with early-stage knee osteoarthritis

Introduction Viscosupplementation with hyaluronic acid (HA) of osteoarthritic (OA) knee joints has a well-established positive effect on clinical symptoms. This effect, however, is only temporary and the working mechanism of HA injections is not clear. It was suggested that HA might have disease modifying properties because of its beneficial effect on cartilage sulphated glycosaminoglycan (sGAG) content. Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) is a highly reproducible, non-invasive surrogate measure for sGAG content and hence composition of cartilage. The aim of this study was to assess whether improvement in cartilage structural composition is detected using dGEMRIC 14 weeks after 3 weekly injections with HA in patients with early-stage knee OA. Methods In 20 early-stage knee OA patients (KLG I-II), 3D dGEMRIC at 3T was acquired before and 14 weeks after 3 weekly injections with HA. To evaluate patient symptoms, the knee injury and osteoarthritis outcome score (KOOS) and a numeric rating scale (NRS) for pain were recorded. To evaluate cartilage composition, six cartilage regions in the knee were analyzed on dGEMRIC. Outcomes of dGEMRIC, KOOS and NRS before and after HA were compared using paired t-testing. Since we performed multiple t-tests, we applied a Bonferroni-Holm correction to determine statistical significance for these analyses. Results All KOOS subscales (‘pain’, ‘symptoms’, ‘daily activities’, ‘sports’ and ’quality of life’) and the NRS pain improved significantly 14 weeks after Viscosupplementation with HA. Outcomes of dGEMRIC did not change significantly after HA compared to baseline in any of the cartilage regions analyzed in the knee. Conclusions Our results confirm previous findings reported in the literature, showing persisting improvement in symptomatic outcome measures in early-stage knee OA patients 14 weeks after Viscosupplementation. Outcomes of dGEMRIC, however, did not change after Viscosupplementation, indicating no change in cartilage structural composition as an explanation for the improvement of clinical symptoms.

No Difference on Quantitative Magnetic Resonance Imaging in Patellofemoral Cartilage Composition Between Patients With Patellofemoral Pain and Healthy Controls

Background: Retropatellar cartilage damage has been suggested as an etiological factor for patellofemoral pain (PFP), a common knee condition among young and physically active individuals. To date, there is no conclusive evidence for an association between cartilage defects and PFP. Nowadays, advanced quantitative magnetic resonance imaging (MRI) techniques enable estimation of cartilage composition. Purpose: To investigate differences in patellofemoral cartilage composition between patients with PFP and healthy control subjects using quantitative MRI. Study Design: Cross-sectional study; Level of evidence, 3. Methods: Patients with PFP and healthy control subjects underwent 3.0-T MRI including delayed gadolinium-enhanced MRI of cartilage and T1ρ and T2 mapping. Differences in relaxation times of patellofemoral cartilage were compared between groups by linear regression analyses, adjusted for age, body mass index, sex, sports participation, and time of image acquisition. Results: This case-control study included 64 patients and 70 controls. The mean (±SD) age was 23.2 ± 6.4 years and the mean body mass index was 22.9 ± 3.4 kg/m2; 56.7% were female. For delayed gadolinium-enhanced MRI of cartilage, the mean T1GD relaxation times of patellar (657.8 vs 669.4 ms) and femoral cartilage (661.6 vs 659.8 ms) did not significantly differ between patients and controls. In addition, no significant difference was found in mean T1ρ relaxation times of patellar (46.9 vs 46.0 ms) and femoral cartilage (50.8 vs 50.2 ms) and mean T2 relaxation times of patellar (33.2 vs 32.9 ms) and femoral cartilage (36.7 vs 36.6 ms) between patients and controls. Analysis of prespecified medial and lateral subregions within the patellofemoral cartilage also revealed no significant differences. Conclusion: There was no difference in composition of the patellofemoral cartilage, estimated with multiple quantitative MRI techniques, between patients with PFP and healthy control subjects. However, clinically relevant differences could not be ruled out for T1ρ in the adolescent population. Retropatellar cartilage damage has long been hypothesized as an important factor in the pathogenesis of PFP, but study findings suggest that diminished patellofemoral cartilage composition is not associated with PFP.

Multiparametric computer-aided differential diagnosis of Alzheimer’s disease and frontotemporal dementia using structural and advanced MRI

ObjectivesTo investigate the added diagnostic value of arterial spin labelling (ASL) and diffusion tensor imaging (DTI) to structural MRI for computer-aided classification of Alzheimers disease (AD), frontotemporal dementia (FTD), and controls.MethodsThis retrospective study used MRI data from 24 early-onset AD and 33 early-onset FTD patients and 34 controls (CN). Classification was based on voxel-wise feature maps derived from structural MRI, ASL, and DTI. Support vector machines (SVMs) were trained to classify AD versus CN (AD-CN), FTD-CN, AD-FTD, and AD-FTD-CN (multi-class). Classification performance was assessed by the area under the receiver-operating-characteristic curve (AUC) and accuracy. Using SVM significance maps, we analysed contributions of brain regions.ResultsCombining ASL and DTI with structural MRI resulted in higher classification performance for differential diagnosis of AD and FTD (AUC = 84%; p = 0.05) than using structural MRI by itself (AUC = 72%). The performance of ASL and DTI themselves did not improve over structural MRI. The classifications were driven by different brain regions for ASL and DTI than for structural MRI, suggesting complementary information.ConclusionsASL and DTI are promising additions to structural MRI for classification of early-onset AD, early-onset FTD, and controls, and may improve the computer-aided differential diagnosis on a single-subject level.Key points• Multiparametric MRI is promising for computer-aided diagnosis of early-onset AD and FTD.• Diagnosis is driven by different brain regions when using different MRI methods.• Combining structural MRI, ASL, and DTI may improve differential diagnosis of dementia.

Age-Dependent Effects of Methylphenidate on the Human Dopaminergic System in Young vs Adult Patients With Attention-Deficit/Hyperactivity Disorder: A Randomized Clinical Trial

IMPORTANCE Although numerous children receive methylphenidate hydrochloride for the treatment of attention-deficit/hyperactivity disorder (ADHD), little is known about age-dependent and possibly lasting effects of methylphenidate on the human dopaminergic system. OBJECTIVES To determine whether the effects of methylphenidate on the dopaminergic system are modified by age and to test the hypothesis that methylphenidate treatment of young but not adult patients with ADHD induces lasting effects on the cerebral blood flow response to dopamine challenge, a noninvasive probe for dopamine function. DESIGN, SETTING, AND PARTICIPANTS A randomized, double-blind, placebo-controlled trial (Effects of Psychotropic Drugs on Developing Brain-Methylphenidate) among ADHD referral centers in the greater Amsterdam area in the Netherlands between June 1, 2011, and June 15, 2015. Additional inclusion criteria were male sex, age 10 to 12 years or 23 to 40 years, and stimulant treatment-naive status. INTERVENTIONS Treatment with either methylphenidate or a matched placebo for 16 weeks. MAIN OUTCOMES AND MEASURES Change in the cerebral blood flow response to an acute challenge with methylphenidate, noninvasively assessed using pharmacological magnetic resonance imaging, between baseline and 1 week after treatment. Data were analyzed using intent-to-treat analyses. RESULTS Among 131 individuals screened for eligibility, 99 patients met DSM-IV criteria for ADHD, and 50 participants were randomized to receive methylphenidate and 49 to placebo. Sixteen weeks of methylphenidate treatment increased the cerebral blood flow response to methylphenidate within the thalamus (mean difference, 6.5; 95% CI, 0.4-12.6; P = .04) of children aged 10 to 12 years old but not in adults or in the placebo group. In the striatum, the methylphenidate condition differed significantly from placebo in children but not in adults (mean difference, 7.7; 95% CI, 0.7-14.8; P = .03). CONCLUSIONS AND RELEVANCE We confirm preclinical data and demonstrate age-dependent effects of methylphenidate treatment on human extracellular dopamine striatal-thalamic circuitry. Given its societal relevance, these data warrant replication in larger groups with longer follow-up. TRIAL REGISTRATION identifier: NL34509.000.10 and trialregister.nl identifier: NTR3103.

Feature Selection Based on SVM Significance Maps for Classification of Dementia

Support vector machine significance maps (SVM p-maps) previously showed clusters of significantly different voxels in dementia-related brain regions. We propose a novel feature selection method for classification of dementia based on these p-maps. In our approach, the SVM p-maps are calculated on the training set with a time-efficient analytic approximation. The features that are most significant on the p-map are selected for classification with an SVM classifier. We validated our method using MRI data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI), classifying Alzheimer’s disease (AD) patients, mild cognitive impairment (MCI) patients who converted to AD within 18 months, MCI patients who did not convert to AD, and cognitively normal controls (CN). The voxel-wise features were based on gray matter morphometry. We compared p-map feature selection to classification without feature selection and feature selection based on t-tests and expert knowledge. Our method obtained in all experiments similar or better performance and robustness than classification without feature selection with a substantially reduced number of features. In conclusion, we proposed a novel and efficient feature selection method with promising results.