Automated segmentation of midbrain structures with high iron content

Harnessing routine MRI for the early screening of Parkinson's disease: a multicenter machine learning study using T2-weighted FLAIR imaging

OBJECTIVE

To explore the potential of radiomics features derived from T2-weighted fluid-attenuated inversion recovery (T2W FLAIR) images to distinguish idiopathic Parkinson's disease (PD) patients from healthy controls (HCs).

METHODS

T2W FLAIR images from 1727 subjects were retrospectively obtained from five cohorts and divided into a training set (395 PD/574 HC), an internal test set (99 PD/144 HC) and an external test set (295 PD/220 HC). Regions of interest (ROIs), including bilateral globus pallidus (GP), putamen (PU), substantia nigra (SN), and red nucleus (RN), were manually delineated. The radiomics features were extracted from ROIs. Six independent machine learning (ML) classifiers were trained on the training set, and validated on the internal and external test sets.

RESULTS

A selection of five, two, three, and ten highly correlated diagnostic features were identified from SN, RN, GP, and PU regions, respectively. Six ML classifiers were implemented based on the combined 20 radiomics features. In the internal test cohort, the six models achieved AUC of 0.96-0.98 with the accuracy ranging from 0.80 to 0.90. In the external test cohort, the multilayer perceptron model demonstrated the highest AUC of 0.85 (95% CI: 0.80-0.89) with an accuracy of 0.78.

CONCLUSION

ML models based on the conventional T2W FLAIR images demonstrated promising diagnostic performance for PD and those models may serve as a basis for future investigations on PD diagnosis with the aid of ML methods.

CRITICAL RELEVANCE STATEMENT

Our study confirmed that early screening of Parkinson's Disease based on the conventional T2W FLAIR images was feasible with the aid of machine learning algorithms in a large multicenter cohort and those models had certain generalization.

KEY POINTS

Conventional head MRI is routinely performed in Parkinson's disease (PD) but exhibits inadequate specificity for diagnosis. Machine learning (ML) models based on conventional T2W FLAIR images showed favorable accuracy for PD diagnosis. ML algorithm enables early screening of PD on routine T2W FLAIR sequence.

Quantitative susceptibility mapping of brain iron in healthy aging and cognition

Quantitative susceptibility mapping (QSM) is a magnetic resonance imaging (MRI) technique that can assess the magnetic properties of cerebral iron in vivo. Although brain iron is necessary for basic neurobiological functions, excess iron content disrupts homeostasis, leads to oxidative stress, and ultimately contributes to neurodegenerative disease. However, some degree of elevated brain iron is present even among healthy older adults. To better understand the topographical pattern of iron accumulation and its relation to cognitive aging, we conducted an integrative review of 47 QSM studies of healthy aging, with a focus on five distinct themes. The first two themes focused on age-related increases in iron accumulation in deep gray matter nuclei versus the cortex. The overall level of iron is higher in deep gray matter nuclei than in cortical regions. Deep gray matter nuclei vary with regard to age-related effects, which are most prominent in the putamen, and age-related deposition of iron is also observed in frontal, temporal, and parietal cortical regions during healthy aging. The third theme focused on the behavioral relevance of iron content and indicated that higher iron in both deep gray matter and cortical regions was related to decline in fluid (speed-dependent) cognition. A handful of multimodal studies, reviewed in the fourth theme, suggest that iron interacts with imaging measures of brain function, white matter degradation, and the accumulation of neuropathologies. The final theme concerning modifiers of brain iron pointed to potential roles of cardiovascular, dietary, and genetic factors. Although QSM is a relatively recent tool for assessing cerebral iron accumulation, it has significant promise for contributing new insights into healthy neurocognitive aging.

An automatic interpretable deep learning pipeline for accurate Parkinson's disease diagnosis using quantitative susceptibility mapping and T1-weighted images

Parkinson's disease (PD) diagnosis based on magnetic resonance imaging (MRI) is still challenging clinically. Quantitative susceptibility maps (QSM) can potentially provide underlying pathophysiological information by detecting the iron distribution in deep gray matter (DGM) nuclei. We hypothesized that deep learning (DL) could be used to automatically segment all DGM nuclei and use relevant features for a better differentiation between PD and healthy controls (HC). In this study, we proposed a DL-based pipeline for automatic PD diagnosis based on QSM and T1-weighted (T1W) images. This consists of (1) a convolutional neural network model integrated with multiple attention mechanisms which simultaneously segments caudate nucleus, globus pallidus, putamen, red nucleus, and substantia nigra from QSM and T1W images, and (2) an SE-ResNeXt50 model with an anatomical attention mechanism, which uses QSM data and the segmented nuclei to distinguish PD from HC. The mean dice values for segmentation of the five DGM nuclei are all >0.83 in the internal testing cohort, suggesting that the model could segment brain nuclei accurately. The proposed PD diagnosis model achieved area under the the receiver operating characteristic curve (AUCs) of 0.901 and 0.845 on independent internal and external testing cohorts, respectively. Gradient-weighted class activation mapping (Grad-CAM) heatmaps were used to identify contributing nuclei for PD diagnosis on patient level. In conclusion, the proposed approach can potentially be used as an automatic, explainable pipeline for PD diagnosis in a clinical setting.

Spatial profiles provide sensitive MRI measures of the midbrain micro- and macrostructure

The midbrain is the rostral-most part of the brainstem. It contains numerous nuclei and white matter tracts, which are involved in motor, auditory and visual processing, and changes in their structure and function have been associated with aging, as well as neurodegenerative disorders. Current tools for estimating midbrain subregions and their structure with MRI require high resolution and multi-parametric quantitative MRI measures. We propose an approach that relies on morphology to calculate profiles along the midbrain and show these profiles are sensitive to the underlying macrostructure of the midbrain. First, we show that the midbrain structure can be sampled, within subject space, along three main axes of the left and right midbrain, producing profiles that are similar across subjects. We use two data sets with different field strengths, that contain R1, R2* and QSM maps and show that the profiles are highly correlated both across subjects and between datasets. Next, we compare profiles of the midbrain that sample ROIs, and show that the profiles along the first two axes sample the midbrain in a way that reliably separates the main structures, i.e., the substantia nigra, the red nucleus, and periaqueductal gray. We further show that age differences which are localized to specific nuclei, are reflected in the profiles. Finally, we generalize the same approach to calculate midbrain profiles on a third clinically relevant dataset using HCP subjects, with metrics such as the diffusion tensor and semi-quantitative data such as T1w/T2w maps. Our results suggest that midbrain profiles, both of quantitative and semi-quantitative estimates are sensitive to the underlying macrostructure of the midbrain. The midbrain profiles are calculated in native space, and rely on simple measurements. We show that it is robust and can be easily expanded to different datasets, and as such we hope that it will be of great use to the community and to the study of the midbrain in particular.

Automated Segmentation of Midbrain Structures in High-Resolution Susceptibility Maps Based on Convolutional Neural Network and Transfer Learning

Background

Accurate delineation of the midbrain nuclei, the red nucleus (RN), substantia nigra (SN) and subthalamic nucleus (STN), is important in neuroimaging studies of neurodegenerative and other diseases. This study aims to segment midbrain structures in high-resolution susceptibility maps using a method based on a convolutional neural network (CNN).

Methods

The susceptibility maps of 75 subjects were acquired with a voxel size of 0.83 × 0.83 × 0.80 mm³ on a 3T MRI system to distinguish the RN, SN, and STN. A deeply supervised attention U-net was pre-trained with a dataset of 100 subjects containing susceptibility maps with a voxel size of 0.63 × 0.63 × 2.00 mm³ to provide initial weights for the target network. Five-fold cross-validation over the training cohort was used for all the models’ training and selection. The same test cohort was used for the final evaluation of all the models. Dice coefficients were used to assess spatial overlap agreement between manual delineations (ground truth) and automated segmentation. Volume and magnetic susceptibility values in the nuclei extracted with automated CNN delineation were compared to those extracted by manual tracing. Consistencies of volume and magnetic susceptibility values by different extraction strategies were assessed by Pearson correlation coefficients and Bland-Altman analyses.

Results

The automated CNN segmentation method achieved mean Dice scores of 0.903, 0.864, and 0.777 for the RN, SN, and STN, respectively. There were no significant differences between the achieved Dice scores and the inter-rater Dice scores (p > 0.05 for each nucleus). The overall volume and magnetic susceptibility values of the nuclei extracted by the automatic CNN method were significantly correlated with those by manual delineation (p < 0.01).

Conclusion

Midbrain structures can be precisely segmented in high-resolution susceptibility maps using a CNN-based method.

Stability of AI-Enabled Diagnosis of Parkinson's Disease: A Study Targeting Substantia Nigra in Quantitative Susceptibility Mapping Imaging

Purpose: Parkinson's disease (PD) diagnosis algorithms based on quantitative susceptibility mapping (QSM) and image algorithms rely on substantia nigra (SN) labeling. However, the difference between SN labels from different experts (or segmentation algorithms) will have a negative impact on downstream diagnostic tasks, such as the decrease of the accuracy of the algorithm or different diagnostic results for the same sample. In this article, we quantify the accuracy of the algorithm on different label sets and then improve the convolutional neural network (CNN) model to obtain a high-precision and highly robust diagnosis algorithm. Methods: The logistic regression model and CNN model were first compared for classification between PD patients and healthy controls (HC), given different sets of SN labeling. Then, based on the CNN model with better performance, we further proposed a novel "gated pooling" operation and integrated it with deep learning to attain a joint framework for image segmentation and classification. Results: The experimental results show that, with different sets of SN labeling that mimic different experts, the CNN model can maintain a stable classification accuracy at around 86.4%, while the conventional logistic regression model yields a large fluctuation ranging from 78.9 to 67.9%. Furthermore, the "gated pooling" operation, after being integrated for joint image segmentation and classification, can improve the diagnosis accuracy to 86.9% consistently, which is statistically better than the baseline. Conclusion: The CNN model, compared with the conventional logistic regression model using radiomics features, has better stability in PD diagnosis. Furthermore, the joint end-to-end CNN model is shown to be suitable for PD diagnosis from the perspectives of accuracy, stability, and convenience in actual use.

Multi-contrast anatomical subcortical structures parcellation

The human subcortex is comprised of more than 450 individual nuclei which lie deep in the brain. Due to their small size and close proximity, up until now only 7% have been depicted in standard MRI atlases. Thus, the human subcortex can largely be considered as terra incognita. Here, we present a new open-source parcellation algorithm to automatically map the subcortex. The new algorithm has been tested on 17 prominent subcortical structures based on a large quantitative MRI dataset at 7 Tesla. It has been carefully validated against expert human raters and previous methods, and can easily be extended to other subcortical structures and applied to any quantitative MRI dataset. In sum, we hope this novel parcellation algorithm will facilitate functional and structural neuroimaging research into small subcortical nuclei and help to chart terra incognita.

MRI biomarkers of motor and non-motor symptoms in Parkinson's disease

Parkinson's disease is a heterogeneous disorder with both motor and non-motor symptoms that contribute to functional impairment. To develop effective, disease modifying treatments for these symptoms, biomarkers are necessary to detect neuropathological changes early in the disease course and monitor changes over time. Advances in MRI scan sequences and analytical techniques present numerous promising metrics to detect changes within the nigrostriatal system, implicated in the cardinal motor symptoms of the disease, and detect broader dysfunction involved in the non-motor symptoms, such as cognitive impairment. There is emerging evidence that iron sensitive, neuromelanin sensitive, diffusion sensitive, and resting state functional magnetic imaging measures can capture changes within the nigrostriatal system. Iron, neuromelanin, and diffusion sensitive measures demonstrate high specificity and sensitivity in distinguishing Parkinson's disease relative to controls, with inconsistent results differentiating Parkinson's disease relative to atypical parkinsonian disorders. They may also serve as useful monitoring biomarkers, with each possibly detecting different aspects of the disease course (early nigrosome changes versus broader substantia nigra changes). Investigations of non-motor symptoms, such as cognitive impairment, require careful consideration of the nature of cognitive deficits to characterize regional and network specific impairment. While the early, executive dysfunction observed is consistent with nigrostriatal degeneration, the memory and visuospatial impairments, the harbingers of a dementia process reflect dopaminergic independent dysfunction involving broader regions of the brain.

Multi-atlas tool for automated segmentation of brain gray matter nuclei and quantification of their magnetic susceptibility

Quantification of tissue magnetic susceptibility using MRI offers a non-invasive measure of important tissue components in the brain, such as iron and myelin, potentially providing valuable information about normal and pathological conditions during aging. Despite many advances made in recent years on imaging techniques of quantitative susceptibility mapping (QSM), accurate and robust automated segmentation tools for QSM images that can help generate universal and sharable susceptibility measures in a biologically meaningful set of structures are still not widely available. In the present study, we developed an automated process to segment brain nuclei and quantify tissue susceptibility in these regions based on a susceptibility multi-atlas library, consisting of 10 atlases with T1-weighted images, gradient echo (GRE) magnitude images and QSM images of brains with different anatomic patterns. For each atlas in this library, 10 regions of interest in iron-rich deep gray matter structures that are better defined by QSM contrast were manually labeled, including caudate, putamen, globus pallidus internal/external, thalamus, pulvinar, subthalamic nucleus, substantia nigra, red nucleus and dentate nucleus in both left and right hemispheres. We then tested different pipelines using different combinations of contrast channels to bring the set of labels from the multi-atlases to each target brain and compared them with the gold standard manual delineation. The results showed that the segmentation accuracy using dual contrasts QSM/T1 pipeline outperformed other dual-contrast or single-contrast pipelines. The dice values of 0.77 ± 0.09 using the QSM/T1 multi-atlas pipeline rivaled with the segmentation reliability obtained from multiple evaluators with dice values of 0.79 ± 0.07 and gave comparable or superior performance in segmenting subcortical nuclei in comparison with standard FSL FIRST or recent multi-atlas package of volBrain. The segmentation performance of the QSM/T1 multi-atlas was further tested on QSM images acquired using different acquisition protocols and platforms and showed good reliability and reproducibility with average dice of 0.79 ± 0.08 to manual labels and 0.89 ± 0.04 in an inter-protocol manner. The extracted quantitative magnetic susceptibility values in the deep gray matter nuclei also correlated well between different protocols with inter-protocol correlation constants all larger than 0.97. Such reliability and performance was ultimately validated in an external dataset acquired at another study site with consistent susceptibility measures obtained using the QSM/T1 multi-atlas approach in comparison to those using manual delineation. In summary, we designed a susceptibility multi-atlas tool for automated and reliable segmentation of QSM images and for quantification of magnetic susceptibilities. It is publicly available through our cloud-based platform (www.mricloud.org). Further improvement on the performance of this multi-atlas tool is expected by increasing the number of atlases in the future.

Automatic localization of the subthalamic nucleus on patient-specific clinical MRI by incorporating 7 T MRI and machine learning: Application in deep brain stimulation

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) has shown clinical potential for relieving the motor symptoms of advanced Parkinson's disease. While accurate localization of the STN is critical for consistent across-patients effective DBS, clear visualization of the STN under standard clinical MR protocols is still challenging. Therefore, intraoperative microelectrode recordings (MER) are incorporated to accurately localize the STN. However, MER require significant neurosurgical expertise and lengthen the surgery time. Recent advances in 7 T MR technology facilitate the ability to clearly visualize the STN. The vast majority of centers, however, still do not have 7 T MRI systems, and fewer have the ability to collect and analyze the data. This work introduces an automatic STN localization framework based on standard clinical MRIs without additional cost in the current DBS planning protocol. Our approach benefits from a large database of 7 T MRI and its clinical MRI pairs. We first model in the 7 T database, using efficient machine learning algorithms, the spatial and geometric dependency between the STN and its adjacent structures (predictors). Given a standard clinical MRI, our method automatically computes the predictors and uses the learned information to predict the patient-specific STN. We validate our proposed method on clinical T2 W MRI of 80 subjects, comparing with experts-segmented STNs from the corresponding 7 T MRI pairs. The experimental results show that our framework provides more accurate and robust patient-specific STN localization than using state-of-the-art atlases. We also demonstrate the clinical feasibility of the proposed technique assessing the post-operative electrode active contact locations.

Optimization and comparative evaluation of nonlinear deformation algorithms for atlas-based segmentation of DBS target nuclei

Nonlinear registration of individual brain MRI scans to standard brain templates is common practice in neuroimaging and multiple registration algorithms have been developed and refined over the last 20 years. However, little has been done to quantitatively compare the available algorithms and much of that work has exclusively focused on cortical structures given their importance in the fMRI literature. In contrast, for clinical applications such as functional neurosurgery and deep brain stimulation (DBS), proper alignment of subcortical structures between template and individual space is important. This allows for atlas-based segmentations of anatomical DBS targets such as the subthalamic nucleus (STN) and internal pallidum (GPi). Here, we systematically evaluated the performance of six modern and established algorithms on subcortical normalization and segmentation results by calculating over 11,000 nonlinear warps in over 100 subjects. For each algorithm, we evaluated its performance using T1-or T2-weighted acquisitions alone or a combination of T1-, T2-and PD-weighted acquisitions in parallel. Furthermore, we present optimized parameters for the best performing algorithms. We tested each algorithm on two datasets, a state-of-the-art MRI cohort of young subjects and a cohort of subjects age- and MR-quality-matched to a typical DBS Parkinson's Disease cohort. Our final pipeline is able to segment DBS targets with precision comparable to manual expert segmentations in both cohorts. Although the present study focuses on the two prominent DBS targets, STN and GPi, these methods may extend to other small subcortical structures like thalamic nuclei or the nucleus accumbens.

Human brain imaging of nicotinic acetylcholine α4β2* receptors using [18 F]Nifene: Selectivity, functional activity, toxicity, aging effects, gender effects, and extrathalamic pathways

Nicotinic acetylcholinergic receptors (nAChR's) have been implicated in several brain disorders, including addiction, Parkinson's disease, Alzheimer's disease and schizophrenia. Here we report in vitro selectivity and functional properties, toxicity in rats, in vivo evaluation in humans, and comparison across species of [¹⁸ F]Nifene, a fast acting PET imaging agent for α4β2* nAChRs. Nifene had subnanomolar affinities for hα2β2 (0.34 nM), hα3β2 (0.80 nM) and hα4β2 (0.83 nM) nAChR but weaker (27-219 nM) for hβ4 nAChR subtypes and 169 nM for hα7 nAChR. In functional assays, Nifene (100 μM) exhibited 14% agonist and >50% antagonist characteristics. In 14-day acute toxicity in rats, the maximum tolerated dose (MTD) and the no observed adverse effect level (NOAEL) were estimated to exceed 40 μg/kg/day (278 μg/m² /day). In human PET studies, [¹⁸ F]Nifene (185 MBq; <0.10 μg) was well tolerated with no adverse effects. Distribution volume ratios (DVR) of [¹⁸ F]Nifene in white matter thalamic radiations were ∼1.6 (anterior) and ∼1.5 (superior longitudinal fasciculus). Habenula known to contain α3β2 nAChR exhibited low levels of [¹⁸ F]Nifene binding while the red nucleus with α2β2 nAChR had DVR ∼1.6-1.7. Females had higher [¹⁸ F]Nifene binding in all brain regions, with thalamus showing >15% than males. No significant aging effect was observed in [¹⁸ F]Nifene binding over 5 decades. In all species (mice, rats, monkeys, and humans) thalamus showed highest [¹⁸ F]Nifene binding with reference region ratios >2 compared to extrathalamic regions. Our findings suggest that [¹⁸ F]Nifene PET may be used to study α4β2* nAChRs in various CNS disorders and for translational research.