Search

The healthcare model is shifting towards integrated care approaches. This new model requires patients to be more closely involved. The iCARE-PD project aims to address this need by developing a technology-enabled, home-based, and community-centered integrated care paradigm. A central part of this project is the codesign process of the model of care, exemplified by the active participation of patients in the design and iterative evaluation of three sensor-based technological solutions. We proposed a codesign methodology used for testing the usability and acceptability of these digital technologies and present initial results for one of them, MooVeo. Our results show the usefulness of this approach in testing the usability and acceptability as well as the opportunity to incorporate patients’ feedback into the development. This initiative will hopefully help other groups incorporate a similar codesign approach and develop tools that are well adapted to patients’ and care teams’ needs.

Over the past few years, neuroimaging studies have been performed in young adults with perinatally acquired HIV (PHIV) to study the impact of HIV infection on the central nervous system (CNS), but no recent review have been published. This review aims to identify brain areas where PHIV eems to have greater impact taking into account demographic, behavioral, and clinical characteristics in PHIV infected patients. For this purpose, PubMed and Medline searches were carried out which included studies from 2010 to April 2020. We performed a systematic review and included 26 articles using structural (brain morphometry and diffusion tensor imaging) and functional magnetic resonance imaging methods involving 1182 PHIV-infected participants. Ample evidence has been provided of HIV effects on underlying brain structure. However, information recorded in the studies is commonly incomplete and results sometimes contradictory. In addition to future improvements and dissemination of tools for the developing brain MRI processing and analysis, the inclusion of data related to HIV infection itself (including clinical and immunovirological characteristics as well as detailed information about antiretroviral treatment such as age at ART initiation) may be of vital importance to the better understanding of the impact of the disease on CNS.

Attenuation correction remains a challenge in pelvic PET/MRI. In addition to the segmentation/model-based approaches, deep learning methods have shown promise in synthesizing accurate pelvic attenuation maps (μ-maps). However, these methods often misclassify air pockets in the digestive tract, potentially introducing bias in the reconstructed PET images. The aims of this work were to develop deep learning–based methods to automatically segment air pockets and generate pseudo-CT images from CAIPIRINHA-accelerated MR Dixon images. Methods: A convolutional neural network (CNN) was trained to segment air pockets using 3-dimensional CAIPIRINHA-accelerated MR Dixon datasets from 35 subjects and was evaluated against semiautomated segmentations. A separate CNN was trained to synthesize pseudo-CT μ-maps from the Dixon images. Its accuracy was evaluated by comparing the deep learning–, model-, and CT-based μ-maps using data from 30 of the subjects. Finally, the impact of different μ-maps and air pocket segmentation methods on the PET quantification was investigated. Results: Air pockets segmented using the CNN agreed well with semiautomated segmentations, with a mean Dice similarity coefficient of 0.75. The volumetric similarity score between 2 segmentations was 0.85 ± 0.14. The mean absolute relative changes with respect to the CT-based μ-maps were 2.6% and 5.1% in the whole pelvis for the deep learning–based and model-based μ-maps, respectively. The average relative change between PET images reconstructed with deep learning–based and CT-based μ-maps was 2.6%. Conclusion: We developed a deep learning–based method to automatically segment air pockets from CAIPIRINHA-accelerated Dixon images, with accuracy comparable to that of semiautomatic segmentations. The μ-maps synthesized using a deep learning–based method from CAIPIRINHA-accelerated Dixon images were more accurate than those generated with the model-based approach available on integrated PET/MRI scanners.

The relationship between psychotic symptoms and global measures of functioning has been widely studied. No previous study has assessed so far the interplay between specific clinical symptoms and particular areas of functioning in first-episode psychosis (FEP) using network analysis methods. A total of 191 patients with FEP (age 24.45 ± 6.28 years, 64.9% male) participating in an observational and longitudinal study (AGES-CM) comprised the study sample. Functioning problems were assessed with the WHO Disability Assessment Schedule (WHODAS), whereas the Positive and Negative Syndrome Scale (PANSS) was used to assess symptom severity. Network analysis were conducted with the aim of analysing the patterns of relationships between the different dimensions of functioning and PANSS symptoms and factors at baseline. According to our results, the most important nodes were “conceptual disorganization”, “emotional withdrawal”, “lack of spontaneity and flow of conversation”, “delusions”, “unusual thought content”, “dealing with strangers” and “poor rapport”. Our findings suggest that these symptoms and functioning dimensions should be prioritized in the clinical assessment and management of patients with FEP. These areas may also become targets of future early intervention strategies, so as to improve quality of life in this population, This work was supported by the Madrid Regional Government (R&D activities in Biomedicine (grant number S2017/BMD-3740 - AGES-CM 2-CM)) and Structural Funds of the European Union. Ana Izquierdo’s work is supported by the PFIS predoctoral program (FI17/00138) from the Instituto de Salud Carlos III (Spain) and co-funded by the European Union (ERDF/ESF, "A way to make Europe”/ “Investing in your future”) and The Biomedical Research Foundation of La Princesa University Hospital. Angela Ib´a˜nez thanks the support of CIBERSAM and of the Spanish Ministry of Science, Innovation and Universities. Instituto de Salud Carlos III (PI16/00834 and PI19/01295) co-financed by ERDF Funds from the European Commission. Covadonga M. Díaz-Caneja holds a Juan Rod´es Grant from Instituto de Salud Carlos III (JR19/00024). Celso Arango was supported by the Spanish Ministry of Science and Innovation. Instituto de Salud Carlos III (SAM16PE07CP1, PI16/02012, PI19/ 024), co-financed by ERDF Funds from the European Commission, “A way of making Europe”, CIBERSAM. Madrid Regional Government (B2017/BMD-3740 AGES-CM-2), European Union Structural Funds. European Union Seventh Framework Program under grant agreements FP7-4-HEALTH-2009-2.2.1-2-241909 (Project EU-GEI), FP7- HEALTH- 2013-2.2.1-2-603196 (Project PSYSCAN) and FP7- HEALTH-2013- 2.2.1-2-602478 (Project METSY); and European Union H2020 Program under the Innovative Medicines Initiative 2 Joint Undertaking (grant agreement No 115916, Project PRISM, and grant agreement No 777394, Project AIMS-2-TRIALS), Fundaci´on Familia Alonso, Fundaci´on Alicia Koplowitz and Fundaci´on Mutua Madrile˜na

Social functioning is severely affected in psychotic disorders. Negative symptoms and social cognition seem to play an important role in social functioning, although the preponderance and relationship between these three domains is not clear. In this study, we sought to assess the interrelation between social cognition, social functioning, and the expressiveness and experiential factors of negative symptoms in first-episode psychosis (FEP).216 patients, participants in a multicentre study (AGES-CM), comprised our study sample. The WHO Disability Assessment Schedule (WHODAS 2.0) was used to assess functioning, whereas the Positive and Negative Schizophrenia Syndrome Scale (PANSS) was used to measure the severity of negative symptoms, and the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) was applied to assess the emotional processing component of social cognition. Network analyses were conducted with the aim of analysing the patterns of relationships between social cognition, social functioning, and the expressiveness and experiential factors of negative symptoms.Our findings suggest that there is a direct relationship between social cognition and social functioning (weight = -.077), but also an indirect connection between them, mediated by the experiential (but not the expressiveness) factor of negative symptoms (weight = 0.300).The importance of the affectation of subdomains of social cognition, as well as the role of negative symptoms, specifically the experiential factor, in the functioning of patients with FEP seems to be relevant. The inclusion of these factors in prevention and treatment programs would thus allow us to reduce their impact on the social functioning of these patients.

Medical imaging is considered one of the most important advances in the history of medicine and has become an essential part of the diagnosis and treatment of patients. Earlier prediction and treatment have been driving the acquisition of higher image resolutions as well as the fusion of different modalities, raising the need for sophisticated hardware and software systems for medical image registration, storage, analysis, and processing. In this scenario and given the new clinical pipelines and the huge clinical burden of hospitals, these systems are often required to provide both highly accurate and real-time processing of large amounts of imaging data. Additionally, lowering the prices of each part of imaging equipment, as well as its development and implementation, and increasing their lifespan is crucial to minimize the cost and lead to more accessible healthcare. This paper focuses on the evolution and the application of different hardware architectures (namely, CPU, GPU, DSP, FPGA, and ASIC) in medical imaging through various specific examples and discussing different options depending on the specific application. The main purpose is to provide a general introduction to hardware acceleration techniques for medical imaging researchers and developers who need to accelerate their implementations.

When a seismometer network records an earthquake, operators will manually review the waveforms and identify the wave phases, a task known as phase picking. Manual phase picking is a time-consuming process that can be automated using machine learning; however, automatic methods have not yet achieved human-level performance, and open-source implementations of state-of-the-art algorithms are not always available. Convolutional networks have revolutionized the field of image processing, where the large amounts of readily available data make possible near-human performance in tasks such as classification and segmentation. Fortunately, phase picking is also an area where thousands of phases are manually picked, which makes convolutional networks a good fit for the processing of this type of data. In this paper, we describe Cospy, an open-source convolutional phase picker that uses a two-stage analysis in which the first stage segments a rough area around the phase, and the second stage regresses the precise location. Our approach was evaluated on the Northern California Earthquake Data Center (NCEDC) data set and, when targeting picks closer than 0.1 s, it achieved an $F_{1}$ -score of 93.13% for P phases and 91.07% for S phases. Our results show that convolutional networks are on track to achieve human-level performance on the task of seismic phase picking and can contribute to decreasing the need for manual analysis. An open-source implementation of the proposed approach, pretrained on the NCEDC data set, can be downloaded at https://github.com/stbnps/cospy .

Objective: The purpose of this paper is to prove that computer-vision techniques allow synthesizing water-fat separation maps for local specific absorption rate (SAR) estimation, when patient-specific water-fat images are not available. Methods: We obtained ground truth head models by using patient-specific water-fat images. We obtained two different label-fusion water-fat models generating a water-fat multiatlas and applying the STAPLE and local-MAP-STAPLE label-fusion methods. We also obtained patch-based water-fat models applying a local group-wise weighted combination of the multiatlas. Electromagnetic (EM) simulations were performed, and B1+ magnitude and 10 g averaged SAR maps were generated. Results: We found local approaches provide a high DICE overlap (72.6 ± 10.2% fat and 91.6 ± 1.5% water in local-MAP-STAPLE, and 68.8 ± 8.2% fat and 91.1 ± 1.0% water in patch-based), low Hausdorff distances (18.6 ± 7.7 mm fat and 7.4 ± 11.2 mm water in local-MAP-STAPLE, and 16.4 ± 8.5 mm fat and 7.2 ± 11.8 mm water in patch-based) and a low error in volume estimation (15.6 ± 34.4% fat and 5.6 ± 4.1% water in the local-MAP-STAPLE, and 14.0 ± 17.7% fat and 4.7 ± 2.8% water in patch-based). The positions of the peak 10 g-averaged local SAR hotspots were the same for every model. Conclusion: We have created patient-specific head models using three different computer-vision-based water-fat separation approaches and compared the predictions of B1+ field and SAR distributions generated by simulating these models. Our results prove that a computer-vision approach can be used for patient-specific water-fat separation, and utilized for local SAR estimation in high-field MRI. Significance: Computer-vision approaches can be used for patient-specific water-fat separation and for patient specific local SAR estimation, when water-fat images of the patient are not available.

Botvinik-Nezer, Rotem et al., Data analysis workflows in many scientific domains have become increasingly complex and flexible. Here we assess the effect of this flexibility on the results of functional magnetic resonance imaging by asking 70 independent teams to analyse the same dataset, testing the same 9 ex-ante hypotheses. The flexibility of analytical approaches is exemplified by the fact that no two teams chose identical workflows to analyse the data. This flexibility resulted in sizeable variation in the results of hypothesis tests, even for teams whose statistical maps were highly correlated at intermediate stages of the analysis pipeline. Variation in reported results was related to several aspects of analysis methodology. Notably, a meta-analytical approach that aggregated information across teams yielded a significant consensus in activated regions. Furthermore, prediction markets of researchers in the field revealed an overestimation of the likelihood of significant findings, even by researchers with direct knowledge of the dataset. Our findings show that analytical flexibility can have substantial effects on scientific conclusions, and identify factors that may be related to variability in the analysis of functional magnetic resonance imaging. The results emphasize the importance of validating and sharing complex analysis workflows, and demonstrate the need for performing and reporting multiple analyses of the same data. Potential approaches that could be used to mitigate issues related to analytical variability are discussed., European Union’s Horizon 2020 research and innovation programme (https://ec.europa.eu/programmes/horizon2020/en) under the Marie Skłodowska-Curie grant agreement no. 665501. A. Eed was supported by a predoctoral fellowship La Caixa-Severo Ochoa from Obra Social La Caixa. S.B.E. was supported by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement no. 785907 (HBP SGA2).

Background Gadolinium-perfusion magnetic resonance (MR) identifies gray matter abnormalities in early multiple sclerosis (MS), even in the absence of structural differences. These perfusion changes could be related to the cognitive disability of these patients, especially in the working memory. Arterial spin labeling (ASL) is a relatively recent perfusion technique that does not require intravenous contrast, making the technique especially attractive for clinical research. Purpose To verify the perfusion alterations in early MS, even in the absence of cerebral volume changes. To introduce the ASL sequence as a suitable non-invasive method in the monitoring of these patients. Material and Methods Nineteen healthy controls and 28 patients were included. The neuropsychological test EDSS and SDMT were evaluated. Cerebral blood flow and bolus arrival time were collected from the ASL study. Cerebral volume and cortical thickness were obtained from the volumetric T1 sequence. Spearman’s correlation analyzed the correlation between EDSS and SDMT tests and perfusion data. Differences were considered significant at a level of P

Typically, pseudo-Computerized Tomography (CT) synthesis schemes proposed in the literature rely on complete atlases acquired with the same field of view (FOV) as the input volume. However, clinical CTs are usually acquired in a reduced FOV to decrease patient ionization. In this work, we present the Franken-CT approach, showing how the use of a non-parametric atlas composed of diverse anatomical overlapping Magnetic Resonance (MR)-CT scans and deep learning methods based on the U-net architecture enable synthesizing extended head and neck pseudo-CTs. Visual inspection of the results shows the high quality of the pseudo-CT and the robustness of the method, which is able to capture the details of the bone contours despite synthesizing the resulting image from knowledge obtained from images acquired with a completely different FOV. The experimental Zero-Normalized Cross-Correlation (ZNCC) reports a 0.9367 ± 0.0138 (mean ± SD) and 95% confidence interval (0.9221, 0.9512), the experimental Mean Absolute Error (MAE) reports 73.9149 ± 9.2101 HU and 95% confidence interval (66.3383, 81.4915), the Structural Similarity Index Measure (SSIM) reports 0.9943 ± 0.0009 and 95% confidence interval (0.9935, 0.9951), and the experimental Dice coefficient for bone tissue reports 0.7051 ± 0.1126 and 95% confidence interval (0.6125, 0.7977). The voxel-by-voxel correlation plot shows an excellent correlation between pseudo-CT and ground-truth CT Hounsfield Units (m = 0.87, adjusted R2 = 0.91, p &lt, 0.001). The Bland–Altman plot shows that the average of the differences is low (−38.6471 ± 199.6100, 95% CI (−429.8827, 352.5884)). This work serves as a proof of concept to demonstrate the great potential of deep learning methods for pseudo-CT synthesis and their great potential using real clinical datasets.

The synthesis of patient data of a certain medical image modality by applying an image processing pipeline starting from other modality is receiving a lot of interest recently, as it allows to save acquisition time and sometimes avoid radiation to the patient. An example of this is the creation of computerized tomography volumes from magnetic resonance imaging data, which can be useful for several applications such as electromagnetic simulations, cranial morphometry and attenuation correction in PET/MR systems. We present a fast patch-based algorithm for this purpose, implemented using graphics processing unit computing techniques and gaining up to $$\times$$×15.9 of speedup against a multicore CPU solution and up to about $$\times$$×75 against a single core CPU solution.

We present a study of multiple sclerosis segmentation algorithms conducted at the international MICCAI 2016 challenge. This challenge was operated using a new open-science computing infrastructure. This allowed for the automatic and independent evaluation of a large range of algorithms in a fair and completely automatic manner. This computing infrastructure was used to evaluate thirteen methods of MS lesions segmentation, exploring a broad range of state-of-theart algorithms, against a high-quality database of 53 MS cases coming from four centers following a common definition of the acquisition protocol. Each case was annotated manually by an unprecedented number of seven different experts. Results of the challenge highlighted that automatic algorithms, including the recent machine learning methods (random forests, deep learning, …), are still trailing human expertise on both detection and delineation criteria. In addition, we demonstrate that computing a statistically robust consensus of the algorithms performs closer to human expertise on one score (segmentation) although still trailing on detection scores This work was partly funded by France Life Imaging (grant ANR-11-INBS-0006 from the French “Investissements d’Avenir” program) for funding and sponsoring the challenge. This work has also been partly supported by a grant (OFSEP) provided by the French State and handled by the “Agence nationale de la recherche”, within the framework of the “Investissements d’Avenir” program, under the reference ANR-10-COHO-002. We also thank the French national cohort OFSEP (a French “Investissements d’Avenir” program), and particularly the imaging group inside this cohort consortium for their constant support, fruitful discussions on the challenge and providing the MR images

Multicolor in situ hybridization (mFISH) is a karyotyping technique used to detect major chromosomal alterations using fluorescent probes and imaging techniques. Manual interpretation of mFISH images is a time consuming step that can be automated using machine learning; in previous works, pixel or patch wise classification was employed, overlooking spatial information which can help identify chromosomes. In this work, we propose a fully convolutional semantic segmentation network for the interpretation of mFISH images, which uses both spatial and spectral information to classify each pixel in an end-to-end fashion. The semantic segmentation network developed was tested on samples extracted from a public dataset using cross validation. Despite having no labeling information of the image it was tested on, our algorithm yielded an average correct classification ratio (CCR) of 87.41%. Previously, this level of accuracy was only achieved with state of the art algorithms when classifying pixels from the same image in which the classifier has been trained. These results provide evidence that fully convolutional semantic segmentation networks may be employed in the computer aided diagnosis of genetic diseases with improved performance over the current image analysis methods. © 2018 International Society for Advancement of Cytometry.

Whole-body attenuation correction (AC) is still challenging in combined PET/MR scanners. We describe Dixon-VIBE Deep Learning (DIVIDE), a deep-learning network that allows synthesizing pelvis pseudo-CT maps based only on the standard Dixon volumetric interpolated breath-hold examination (Dixon-VIBE) images currently acquired for AC in some commercial scanners. Methods: We propose a network that maps between the four 2-dimensional (2D) Dixon MR images (water, fat, in-phase, and out-of-phase) and their corresponding 2D CT image. In contrast to previous methods, we used transposed convolutions to learn the up-sampling parameters, we used whole 2D slices to provide context information, and we pretrained the network with brain images. Twenty-eight datasets obtained from 19 patients who underwent PET/CT and PET/MR examinations were used to evaluate the proposed method. We assessed the accuracy of the μ-maps and reconstructed PET images by performing voxel- and region-based analysis comparing the SUVs (in g/mL) obtained after AC using the Dixon-VIBE (PET(Dixon)), DIVIDE (PET(DIVIDE)), and CT-based (PET(CT)) methods. Additionally, the bias in quantification was estimated in synthetic lesions defined in the prostate, rectum, pelvis, and spine. Results: Absolute mean relative change values relative to CT AC were lower than 2% on average for the DIVIDE method in every region of interest except for bone tissue, where it was lower than 4% and 6.75 times smaller than the relative change of the Dixon method. There was an excellent voxel-by-voxel correlation between PET(CT) and PET(DIVIDE) (R(2) = 0.9998, P < 0.01). The Bland–Altman plot between PET(CT) and PET(DIVIDE) showed that the average of the differences and the variability were lower (mean PET(CT)–PET(DIVIDE) SUV, 0.0003; PET(CT)–PET(DIVIDE) SD, 0.0094; 95% confidence interval, [−0.0180,0.0188]) than the average of differences between PET(CT) and PET(Dixon) (mean PET(CT)–PET(Dixon) SUV, 0.0006; PET(CT)–PET(Dixon) SD, 0.0264; 95% confidence interval, [−0.0510,0.0524]). Statistically significant changes in PET data quantification were observed between the 2 methods in the synthetic lesions, with the largest improvement in femur and spine lesions. Conclusion: The DIVIDE method can accurately synthesize a pelvis pseudo-CT scan from standard Dixon-VIBE images, allowing for accurate AC in combined PET/MR scanners. Additionally, our implementation allows rapid pseudo-CT synthesis, making it suitable for routine applications and even allowing retrospective processing of Dixon-VIBE data.

PURPOSE: MRI-based skull segmentation is a useful procedure for many imaging applications. This study describes a methodology for automatic segmentation of the complete skull from a single T1-weighted volume. METHODS: The skull is estimated using a multi-atlas segmentation approach. Using a whole head computed tomography (CT) scan database, the skull in a new MRI volume is detected by nonrigid image registration of the volume to every CT, and combination of the individual segmentations by label-fusion. We have compared Majority Voting, Simultaneous Truth and Performance Level Estimation (STAPLE), Shape Based Averaging (SBA), and the Selective and Iterative Method for Performance Level Estimation (SIMPLE) algorithms. RESULTS: The pipeline has been evaluated quantitatively using images from the Retrospective Image Registration Evaluation database (reaching an overlap of 72.46 ± 6.99%), a clinical CT-MR dataset (maximum overlap of 78.31 ± 6.97%), and a whole head CT-MRI pair (maximum overlap 78.68%). A qualitative evaluation has also been performed on MRI acquisition of volunteers. CONCLUSION: It is possible to automatically segment the complete skull from MRI data using a multi-atlas and label fusion approach. This will allow the creation of complete MRI-based tissue models that can be used in electromagnetic dosimetry applications and attenuation correction in PET/MR. Magn Reson Med, 2015. © 2015 Wiley Periodicals, Inc.

Magnetic resonance imaging (MRI) is a medical technique used in radiology to obtain anatomical images of healthy and pathological tissues. Due to hardware limitations and clinical protocols, MRI data are often acquired with low-resolution. For this reason, the scientific community has been developing super-resolution (SR) methodologies in order to enhance spatial resolution through post-processing of 2D multi-slice images. The enhancement of spatial resolution in magnetic resonance (MR) images improves clinical procedures such as tissue segmentation, registration and disease diagnosis. Several methods to perform SR-MR images have been proposed. However, they present different drawbacks: sensitivity to noise, high computational cost, and complex optimization algorithms. In this paper, we develop a supervised learning methodology to perform SR-MR images using a patch-based Gaussian process regression (GPR) method. We compare our approach with nearest-neighbor interpolation, B-splines and a SR-GPR scheme based on nearest-neighbors. We test our SR-GPR algorithm in MRI-T1 and MRI-T2 studies, evaluating the performance through error metrics and morphological validation (tissue segmentation). Results obtained with our methodology outperform the other alternatives for all validation protocols.

Fetal health is critically dependent on placental function, especially placental transport of oxygen from mother to fetus. When fetal growth is compromised, placental insufficiency must be distinguished from modest genetic growth potential. If placental insufficiency is present, the physician must trade off the risk of prolonged fetal exposure to placental insufficiency against the risks of preterm delivery. Current ultrasound methods to evaluate the placenta are indirect and insensitive. We propose to use Blood-Oxygenation-Level-Dependent (BOLD) MRI with maternal hyperoxia to quantitatively assess mismatch in placental function in seven monozygotic twin pairs naturally matched for genetic growth potential. In-utero BOLD MRI time series were acquired at 29 to 34 weeks gestational age. Maps of oxygen Time-To-Plateau (TTP) were obtained in the placentas by voxel-wise fitting of the time series. Fetal brain and liver volumes were measured based on structural MR images. After delivery, birth weights were obtained and placental pathological evaluations were performed. Mean placental TTP negatively correlated with fetal liver and brain volumes at the time of MRI as well as with birth weights. Mean placental TTP positively correlated with placental pathology. This study demonstrates the potential of BOLD MRI with maternal hyperoxia to quantify regional placental function in vivo.

Reconstruction of source neural activity has an increasing importance due to its high time resolution, promoting its application in diagnosis of neurodegenerative and cognitive tasks. To improve the accuracy of reconstruction, we study the influence of anisotropic blood vessels in the EEG-based source localization solution, including several tissues. To this end, we develop a model that reflects physical properties of the head volume based on collected angiographic data. From obtained results of real data, we find that omission of the anisotropic blood vessels within the forward modeling may result in potential discrepancies larger than 35 \(\upmu \)V and dipole localization errors greater than 15 mm, especially, in deep brain areas.

© 2017 International Society for Magnetic Resonance in Medicine Purpose: To present a method for spatiotemporal alignment of in-utero magnetic resonance imaging (MRI) time series acquired during maternal hyperoxia for enabling improved quantitative tracking of blood oxygen level-dependent (BOLD) signal changes that characterize oxygen transport through the placenta to fetal organs. Materials and Methods: The proposed pipeline for spatiotemporal alignment of images acquired with a single-shot gradient echo echo-planar imaging includes 1) signal nonuniformity correction, 2) intravolume motion correction based on nonrigid registration, 3) correction of motion and nonrigid deformations across volumes, and 4) detection of the outlier volumes to be discarded from subsequent analysis. BOLD MRI time series collected from 10 pregnant women during 3T scans were analyzed using this pipeline. To assess pipeline performance, signal fluctuations between consecutive timepoints were examined. In addition, volume overlap and distance between manual region of interest (ROI) delineations in a subset of frames and the delineations obtained through propagation of the ROIs from the reference frame were used to quantify alignment accuracy. A previously demonstrated rigid registration approach was used for comparison. Results: The proposed pipeline improved anatomical alignment of placenta and fetal organs over the state-of-the-art rigid motion correction methods. In particular, unexpected temporal signal fluctuations during the first normoxia period were significantly decreased (P < 0.01) and volume overlap and distance between region boundaries measures were significantly improved (P < 0.01). Conclusion: The proposed approach to align MRI time series enables more accurate quantitative studies of placental function by improving spatiotemporal alignment across placenta and fetal organs. Level of Evidence: 1. Technical Efficacy: Stage 1. J. MAGN. RESON. IMAGING 2017;46:403–412.

Eyelid identification provides key data that can be used in several application such as controlling gaze-based HMIs (human machine interfaces), the design of new diagnostic tools for brain diseases, improving driver safety, drowsiness detection, research on advertisement, etc. We propose a novel eyetracking algorithm by learning a deep deconvolutional neural network. To train and test our method, we use several data sets with hand-labeled eye images from real-world tasks. Our method outperforms previous eye tracking methods, improving the results of the current state of the art in a 19%.

Fluorescent cytometry refers to the quantification of cell physical properties and surface biomarkers using fluorescently-tagged antibodies. The generally preferred techniques to perform such measurements are flow cytometry, which performs rapid single cell analysis by flowing cells one-by-one through a channel, and microscopy, which eliminates the complexity of the flow channel, offering multi-cell analysis at a lesser throughput. Low-magnification image-based cytometers, also called “cell astronomy” systems, hold promise of simultaneously achieving both instrumental simplicity and high throughput. In this magnification regime, a single cell is mapped to a handful of pixels in the image. While very attractive, this idea has, so far, not been proven to yield quantitative results of cell-labeling, mainly due to the poor signal-to-noise ratio present in those images and to partial volume effects. In this work we present a cell astronomy system that, when coupled with custom-developed algorithms, is able to quantify cell intensities and diameters reliably. We showcase the system using calibrated MESF beads and fluorescently stained leukocytes, achieving good population identification in both cases. The main contribution of the proposed system is in the development of a novel algorithm, H-EM, that enables inter-cluster separation at a very low magnification regime (2x). Such algorithm provides more accurate brightness estimates than DAOSTORM when compared to manual analysis, while fitting cell location, brightness, diameter, and background level concurrently. The algorithm first performs Fisher discriminant analysis to detect bright spots. From each spot an expectation-maximization algorithm is initialized over a heterogeneous mixture model (H-EM), this algorithm recovers both the cell fluorescence and diameter with sub-pixel accuracy while discriminating the background noise. Finally, a recursive splitting procedure is applied to discern individual cells in cell clusters.

The pathogenesis of orthostatic tremor (OT) remains unclear, although some evidence points to dysfunction in the brainstem or cerebellum. We used single voxel proton magnetic resonance spectroscopy (1H-MRS) (3 T) to investigate whether neurochemical changes underlie abnormal cerebellar or cortical function in OT. Fourteen OT patients and 14 healthy controls underwent 1H-MRS studies with voxels placed in midparietal gray matter and cerebellum (vermis and central white matter). Spectral analysis was analyzed using the software package LCModel (version 6.3). The absolute metabolite concentrations and ratios of total N-acetylaspartate + N-acetylaspartyl glutamate (NAA), choline-containing compounds, myoinositol, and glutamate + glutamine to creatine were calculated. In midparietal gray matter spectra, we found a significant decrease in the absolute concentration of NAA in OT patients versus healthy controls (7.76 ± 0.25 vs 8.11 ± 0.45, P = 0.017). A similar decrease in NAA was seen in the cerebellar vermis (7.33 ± 0.61 vs 8.55 ± 1.54, P = 0.014) and cerebellar white matter (8.54 ± 0.79 vs 9.95 ± 1.57, P = 0.010). No differences in the other metabolites or their ratios were observed. Reductions in both cerebral cortical and cerebellar NAA suggest that there is neuronal damage or loss in OT, raising the intriguing question as to whether OT is a neurodegenerative disease. Along with clinical history and electrophysio0logical examination, 1H-MRS could serve as a useful diagnostic aid for OT.

Cardiovascular disease is one of the leading causes of death worldwide. Epicardial adipose tissue (EAT) has emerged as an independent predictor of high cardiometabolic risk. Cardiovascular MRI has proven to be a feasible and reproducible method to assess EAT quantitatively. We present a novel approach for the automated quantification of EAT using "a priori" anatomical information. We extracted a region of interest (ROI) in the end-diastolic heart phase followed by a GVF-snake algorithm to smooth it. For the EAT and endocardial boundary detection, a Law's texture filter is applied. Left and right ventricle are localized using spatial prior information. Then, thresholding is applied to quantify the cardiac muscle. For the EAT, it is differentiated from the paracardial fat by K-cosine curvature analysis. Results for 10 morbidly obese patients show no significant differences between manual and automatic quantification with a remarkable time and effort saving between them.

MRI-based bone segmentation is a challenging task because bone tissue and air both present low signal intensity on MR images, making it difficult to accurately delimit the bone boundaries. However, estimating bone from MRI images may allow decreasing patient ionization by removing the need of patient-specific CT acquisition in several applications. In this work, we propose a fast GPU-based pseudo-CT generation from a patient-specific MRI T1-weighted image using a group-wise patch-based approach and a limited MRI and CT atlas dictionary. For every voxel in the input MR image, we compute the similarity of the patch containing that voxel with the patches of all MR images in the database, which lie in a certain anatomical neighborhood. The pseudo-CT is obtained as a local weighted linear combination of the CT values of the corresponding patches. The algorithm was implemented in a GPU. The use of patch-based techniques allows a fast and accurate estimation of the pseudo-CT from MR T1-weighted images, with a similar accuracy as the patient-specific CT. The experimental normalized cross correlation reaches 0.9324±0.0048 for an atlas with 10 datasets. The high NCC values indicate how our method can accurately approximate the patient-specific CT. The GPU implementation led to a substantial decrease in computational time making the approach suitable for real applications.

136 Objectives It is largely recognized that tumor microenvironment influences the efficacy of therapy and affects the outcome of different neoplasia. Therefore, we investigated the impact of metabolic information on FDG-PET in resectable non-small cell lung cancer (NSCLC), by comparing the imaging findings with the tissue expression of immune-checkpoints and other markers of tumor-related immunity. Methods All patients referred to our Institution for upfront surgical resection of NSCLC, who were investigated with FDG-PET prior to surgery (n=130) were consecutively enrolled in the study. From January 2010 to May 2014, 55 patients (stage IA-IIIB; M:F=42:13; mean age 68.9 years) were investigated. Sampled surgical tumor specimens were analyzed with immunohistochemistry (IHC) for CD68-TAMs (tumor-associated macrophages), CD8-TILs (tumor infiltrating Lymphocytes), PD-1 and PD-L1. The immunoreactivity was quantitatively evaluated and the scores compared with imaging findings. FDG-PET images were visualized on a GE Adw4.6 workstation and used to define semi-quantitative parameters (cut-off threshold 42%): i.e. SUVmax and SUVmean. Metabolic information on FDG-PET were correlated with tissue markers expression, and both with disease-free survival (DFS) considering a median follow-up of 16.2 months. Results Overall, 36 adenocarcinomas (ADK), 18 squamous cell carcinomas (SCCs), and 1 sarcomatoid carcinoma were analyzed. All investigated tumors resulted positive at FDG-PET: median SUVmax 11.3 (range: 2.3-32.5) and SUVmean 6.4 (range: 1.5-13), both significantly higher in SCCs compared to other histotypes (p=0.007 and 0.048, respectively). IHC demonstrated a median immunoreactive area covered by CD68-TAMs of 5.41% (range: 0.84-14.01%), CD8-TILs of 2.9% (range: 0.11-11.92%), PD-1 of 0.65% (range: 0.02-5.87%), and PD-L1 of 0.7% (range: 0.03-10.29%). We found a statistically significant correlation between SUVmax and SUVmean with the expression of CD68-TAMs (rho=0.31; p=0.027), CD8 TILs (rho=0.55; p>0.001), and PD-1 (rho=0.33; p=0.017 and 0.009 respectively). The other tissue markers correlated as follows: CD8 TILs and PD-1 (rho=0.45; p=0.001), CD8 TILs and PD-L1 (rho=0.41; p=0.003), CD68-TAMs and PD-L1 (rho=0.30; p=0.027), PD-1 and PD-L1 (rho=0.26; p=0.059). With respect to the outcome, only SUVmax, SUVmean and stage showed a statistically significant correlation with DFS (p=0.002, 0.004 and Conclusions The present research shows a direct association between metabolic parameters on FDG-PET and the tumor expression of immune-checkpoints and other markers of tumor-related immunity, such as PD-1 and PD-L1, CD8-TILs, and CD68-TAMs.

In each new generation of PET scanners, the axial field-of-view (FOV) has been increased as a way to improve their sensitivity and obtain better images, faster and with lower injected doses. Nevertheless, in neurological studies, the use of larger axial FOV can introduce errors in the PET attenuation correction (AC), as possible movements of the patient’s mouth are not currently considered. This effect happens when oblique PET sinograms containing both the jaw and the brain are used. In this work we have studied the bias that mouth motion can introduce in the PET AC, for two different scanners. The average effect over large regions in the brain is around 1 % for a moderate mouth motion, although some voxels show differences of up to 18 %. These results indicate that if scanners with large axial FOV and acceptance angles are going to be used in the future, possible movements of patient’s mouth should be considered in the AC procedure.