This PDF file contains the front matter associated with SPIE Proceedings Volume 10975, including the Title Page, Copyright information, Table of Contents, Author and Conference Committee lists.

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by a set of progressive motor disabilities knows as shuffling gait patterns. The diagnosis and treatment of parkinsonian patients at different stages is typically supported by a Kinematic analysis. In clinical routine, such analysis is related with the quantitative and qualitative description of body segment displacements, computed from a reduced set of markers. Nevertheless, classical markers-based analysis has strong limitations to capture local and regional dynamic relationships associated with shuffling gait patterns. Particularly, the sparse set of markers lost sensitivity to detect progression of disease and commonly this kinematic characterization is restricted only to advanced stages. This work introduces a new hierarchical parkinsonian gait descriptor that coded kinematics at local and regional levels. At local level, a Spatial Kinematic Pattern (SKP) is computed as circular binary occurrence vectors, along trajectories. Regionally, such local vectors are grouped to describe body segments motions. Each of these regions coarsely correspond to the head, trunk and limbs. From each independent region is possible to describe kinematic patterns associated with the disease. The proposed approach was validated into a classification scheme to differentiate among regional parkinsonian patterns w.r.t to control patterns. Hence, each coding region descriptor was mapped to a support vector machine model. The proposed method was evaluated from a set of 84 gait videos of control and parkinsonian patients, achieving an average accuracy of 84, 52%.

Resting hand tremor is one of the most important biomarkers in Parkinson’s disease (PD). This indicator is mainly described as periodic oscillatory movements when hands are completely supported, i.e., without voluntary muscle contraction. Such characterization is however very difficult to observe in standard clinical analysis, due to the imperceptible low tremor amplitude. Furthermore, in early stages of PD those motions are commonly misclassified as control patterns. Common clinical practice often suggests a physical tremor magnification by forcing postural hand configurations, dealing with natural strain motions that might disturb tremor behavior. In this work was introduced a video characterization that highlights hand tremor patterns from resting and postural setups. Initially, each of videos are represented as a bank of spatial and temporal filters. Then, specific spatio-temporal bands are amplified to stand out tremor patterns. A set of anatomical points of interest was fixed to be quantitatively assessed along the magnified sequence. Temporal variance of these points were associated with tremor recorded in videos. The proposed approach was evaluated in a total of 80 videos recording hands in resting and postural configurations. Variance analysis was performed to measure temporal amplitude differences of tremor in PD and control videos. In resting validation, a gain of 7.76 dB was achieved in parkinsonian and control comparison by using amplified videos. While physical magnification obtains a F-test of 5.19, the proposed optical magnification yields a F-test of 8.19, allowing a better quantification of the disease.

In this paper we present a new non-contact strategy to estimate the breathing rate based on the Eulerian motion magnification technique and a system based on different images processing steps. After the magnification procedure, a ROI is selected manually, an enhancement algorithm based on an adaptive histogram equalization is applied and finally the frames are binarized using the Otsu algorithm. Morphological operations are carry out on the video frames and a tracking temporal strategy is implemented to estimate the breathing rate. The magnification procedure was carried out using an Hermite decomposition. We have tested the method on three subjects in four positions (seat, lying face down, lying face up and lying in fetal position). The motion magnification approach is compared to the Laplacian decomposition strategy computing the mean absolute error.

Background: Nowadays, cardiovascular diseases (CVD) remain the main cause of death worldwide. A heart sound signal or phonocardiogram (PCG) is the most simple, economical and non-invasive tool to detect CVDs. Advances in technology and signal processing allow the design of computer-aided systems for heart illnesses detection from PCG signals. Purpose: The paper proposes a pipeline and benchmark for binary heart sounds classification. The features extraction architecture is focused on the use of Matching Pursuit time-frequency decomposition using Gabor dictionaries and the Linear Predictive Coding method of a residual. We compare seven classifiers with two different approaches: feature averaging and cycle averaging. Methods: We test our proposal on the PhysioNet/CinC challenge 2016 database, which comprises a wide variety of heart sounds recorded from patients with normal and different pathological heart conditions. We conduct a 10-fold stratified cross-validation method to evaluate the performance of different classification algorithms. The feature sets were also tested when using an oversampling method for balancing. Results: The benchmark identified systems showing a satisfying performance in terms of accuracy, sensitivity, and Matthews correlation coefficient. Results can be improved when using feature averaging and an oversampling strategy.

Emotions are affective states accompanied by physiological reactions that affect cognition processes such as decision making, perception, and learning. Emotion detection can be helpful in fields like education, sports and accident prevention. In this pilot study, we used biosensors to measure heart rate and galvanic skin response of twenty-eight volunteers (fourteen male, fourteen female). They were asked to watch video clips to elicit two target emotions: amusement and anger. The purpose of this study was to determine the relationship between mean values of biosignals and emotional states (including amusement, anger and neutral state). From the analysis of variance, Fisher least significant difference and Multiple Range test, it was observed that emotions elicited with video clips influence mean values and other features of physiological signals with a confidence level of 90%.

Intimate Partner Violence (IPV) against women is a major problem in Colombia. Nowadays the question about the effects of violence on women and the identification of latent risks that affect their health, is increasingly relevant. This article describes a pilot study that aimed to measure electrophysiological signals corresponding to the emotional neurophysiological response of women who had experienced IPV in contrast to those who did not. Six healthy female adults, ranged in age from 18 to 55 years old enrolled in this study. For the measurement we used the International Affective Picture System (IAPS) and an Auditory and Visual Emotional Memory Test (avEMT), and we recorded the EEG signal with a g.Nautilus 32 g.LadyBird. EEG signals acquired from baseline and during the tests were compared. As a result of IAPS test, we found for all the participants a higher power spectrum at low EEG frequencies and a decrease in power as the frequencies increase for baseline and emotional pictures. For the avMET, both groups show a higher power spectrum in the different phases of the task compared with the baseline, with an exception of one participant from the WIPV group who show the opposite tendency. Also, two machine learning models were trained and an accuracy of more than 85% were achieved to classify EEG signals from women who experienced IPV and women who had not. This research is an approach to the phenomenon of violence against women and broadens the understanding of the effects on emotional response and electrophysiological activation in women who have experienced this type of situation.

Obstructive sleep apnea (OSA) is a disorder that primarily affects the respiratory system and occurs during the sleep period. (OSA) is a respiratory disorder related to common sleep, characterized by repetitive obstruction of the upper airway, intermittent hypoxemia and recurrent awakenings during sleep. The absence of this stimulus generates a lack of respiratory effort, because the diaphragm muscle is not being stimulated. People with this disorder increase the risk of suffering fatigue, daytime sleepiness, psychological disorders, cardiovascular problems and respiratory and physical weakness, among many others. For this reason, the present project presents an electronic device that allows the measurement and / or detection of physical variables related to the diagnosis of OSA, using LabVIEW software. The developed device is non-invasive, designed to measure the distance of the thoracic cavity, the nasal temperature, the respiratory flow and the sound produced by the snoring originated mainly in the trachea.

Currently, the deaf women’s community must face the lack of several resources related to their personal and professional development: (i) the lack of bilingualism (sign language and written language), (ii) the lack of appropriate knowledge about the local language sign; and (iii) the scarcity of educational plans on Sexual and Reproductive Health (SRH). This situation triggers several problems, such as the transmission of sexually trans- mitted diseases, unwanted pregnancy in adolescence, sexual violence and complications during pregnancy. For these reasons, in this document, we present an intelligent ecosystem and an educational methodology aimed at improving access to SRH programs in Ecuadorian Sign Language (ESL), with the aim of improving access to information and knowledge in SRH through methodologies based on interactive and dynamic learning environments capable of responding to the specific needs of deaf people. Our proposal relies on an expert system based on rules-reasoning, and an educational web-environment. This approach was tested with 30 volunteers (26 deaf women and 4 interpreters in ESL), evaluating the contents that are housed within the system (content quality, the coherence of the applied methodology, access to the environment,...). The results are encouraging and have shown that it is necessary to implement interactive programs for the support of the SRH.

Several authors such as Piaget, suggest that sensory and motor experiences are considered the base of intellectual functioning during the first two years of a person’s life. In this line, the fine and gross motor skills are tightly related with the cognitive development of a child. For these reasons, in this paper, we present an interactive system aimed at supporting exercises and activities for both rehabilitation and development of motor skills in children with disabilities. Our proposal relies on a Kinect device and an expert system that provides a guide of activities that must be carried out by a child according to his/her profile. We have tested the system in two stages, one to determine how well the expert system provides suggestions on the base of one first level of granularity: areas and general interactive activities configuration. Moreover, in the second stage, we worked in real therapy sessions with six children with different types of disabilities. The results show high levels of motivation by children side, and 90% of precision in the generation of therapy plans.

Abdominal electrocardiography (AECG) is an indirect method for obtaining a continuous reading of fetal heart rate and is widely used during pregnancy as a method for assessing fetal well-being. Information obtained by AECG is used for early identification of fetal risk and may help in the anticipation of future complications; however, improper interpretation of the AECG recordings, related with inter- and intra-individual variability, may lead to inadequate treatments that can cause the death of the fetus. A set of 33 indices (4 maternal, 5 temporals, 23 time-frequency and 1 non-linear), extracted from AECG recordings and maternal information, were tested with a Random Forest (RF) classification method for the identification of normal fetuses and fetuses with intrauterine growth restriction. Because RFs may perform poorly when confronted with a high number of features compared to the number of training data available, a Genetic Algorithm (GA) was used to select the minimum set of features that improves the outcome of the RF. The accuracy of the RF method using the 33 indices was of 60%. After a run of the GA, the best individual in the last generation had an accuracy value of 85% and reduced the number of used indices from 33 to 11.

This study proposes a new tool based on Virtual Reality (VR) as a complement in the treatment of people diagnosed with Autism Spectrum Disorder (ASD). VR tools have been stablished in last years as a new option in learning and practising new skills during the treatment. In this work, a VR application is developed simulating several environments corresponding to different types of emotions according to the Gestalt school of psychology. The VR application was tested in five male teenagers diagnosed with ASD of level one according to the DSM-5 during the therapy sessions. A qualitative evaluation of the VR application is carried out by the therapist during the session. It is observed and annotated which emotions have been detonated by the VR application giving to the therapist new information for the subsequent sessions.

The comet assay is a commonly used technique in molecular and cell biology fields, for studies in which the DNA damage of a cell is measured. For instance, it is useful to analyze whenever a carcinogenic cell is affected by chemical agents, helping with oncology research. Traditionally, in order to evaluate the damage of a cell, an expert observes the morphology and the intensity (brightness) of the resulting comet. However, taking into account that a large number of images have to be analyzed, this task may demand a lot of time to be done manually. In recent years, the comet assay analysis has been implemented semi-automatically and automatically with the rise of new image processing algorithms. Although these new algorithms reduce the time invested in the image analysis, some problems in comet identification and accurate measure of their components need to be improved. This project aimed to develop an algorithm and an interface, named CometLab, for flexible automatic comet segmentation. Its performance was assessed with a set of images and compared against an open source, available software called OpenComet. It was found that only 1 of the 15 features that were extracted by both algorithms was not statistically correlated (head diameter), meaning that the designed application is suitable; therefore, this research helped to obtain information about the performance of CometLab in comparison to OpenComet, which serves as setpoint for future works in which it would be possible to decide which algorithm is better.

New evidence suggests 25% - 26% of colon polyps may be missed during a routine colonoscopy[1, 2, 3, 4, 5]. These polyps or hyperplastic lesions are currently considered as pre-neoplastic lesions that must be detected. In this context, automatic strategies are appealing as second readers or diagnostic supporting tools. However, this task is challenging because of the huge variability and multiple sources of noise. This paper introduces a strategy for automatic detection of polyps larger than 5 mm. The underlying idea is that polyps in a sequence of frames are those locations with smaller frame-to-frame variance. The method starts by segmenting an input frame into a set of superpixels, i.e., clusters of neighbor pixels with minimal luminance variance. Each of these superpixels in characterized by a concatenated vector of 57 features collecting texture, shape, and color. A Support Vector Machine with a linear and Radial Basis Function (RBF) kernel was used as a supervised learning model. The evaluation was carried out using a set of 39 cases belonging to two datasets (6.594 frames: 3.123 with polyps and 3.471 without polyps) under a Leave-One-Out Cross Validation scheme and obtaining a 0.73 of accuracy. In addition, the data set was split into 70%-30% between train and test respectively and obtaining a 0.87 of accuracy.

The non-small cell lung cancer (NSCLC) is the most frequent with about 80% of new cases and it is subdivided into adenocarcinoma, squamous cell and large cell carcinomas. Several studies have demonstrated the relevance of identifying NSCLC cancer subtype for prognosis and treatment. This work presents a classification approach for NSCLC subtypes in computed tomography images based on a multi-scale texture analysis. For doing so, gradients over the difference between multi-scale homogeneity textures was computed to build feature descriptors. Binary classifications were performed for the three NSCLC cancer subtypes under a 10-fold cross-validation scheme, and the best results were obtained for adenocarcinoma vs. squamous cell carcinoma, with an area under the curve of 80% and an accuracy of 77; 4%. The results demonstrate that CT is an useful source of information for extracting patterns that allow to identify tissue changes and correlate them with histological outcome.

Image-based diagnosis becoming one of the most important areas in medicine, as the diversity and sophistication of imaging techniques are being increasingly used in hospitals and medical centers. This, however, raises the issue of having image analysis capabilities that go with this trend, to be able to use medical imagery to provide fast and accurate diagnosis. In andrology in particular, the spermiogram analysis is considered the most significant study to evaluate the male reproductive capacity. Spermiograms can be produced with relatively little effort and cost, since they require only standard procedures for sample treatment. However, an adequate assessment of sperm quality requires the careful inspection by higly trained specialists, requiring time, and being prone to high inter- and intra-specialist variances. In this paper we present a system for automatic spermiogram analysis using image processing and machine learning techniques. The system was trained using a repository of spermiograms and the opinion of several experts in andrology and in human reproduction, using different information sources and classification criteria. The results are aimed to develop a SaaS CASA (Computer Assisted Sperm Analysis) system that can provide results over the Internet.

This paper presents a quantitative imaging method and software technology to predict the risk and assess the severity of respiratory diseases in premature babies by fusing information from multiple sources: non-invasive low-radiation chest X-ray (CXR) imaging and clinical parameters. Prematurity is the largest single cause of death in children under five in the world. Lower respiratory tract infections (LRTI) are the top cause of hospitalization and mortality in prematurity. However, there is no objective clinical marker to predict and prevent severe LRTI in the 15 million babies born prematurely every year worldwide. Traditionally, imaging biomarkers of lung disease from computed tomography have been successfully used in adults, but they entail heightened risks for children due to cumulative radiation and the need for sedation. The proposed technology is the first approach that uses low-radiation CXR imaging to predict hospitalization due to LRTI in prematurity. The method uses deep learning to quantify heterogeneous patterns (air trapping and irregular opacities) in the chest, which are combined with clinical parameters to predict the risk of LRTI. Our preliminary results obtained using a data obtained from ten premature subjects with LRTI showed high correlation between our imaging biomarkers and the rehospitalization of these subjects R²=0.98).

Age-related macular degeneration is a common cause of vision loss in people aging 55 and older. The condition affects the light-sensing cells in the macula limiting the sharp and central vision. On the other hand, Spectral Domain Optical Coherence Tomography (SD-OCT) allow highlighting abnormalities and thickness in the retinal layers which are useful for age-related macular degeneration diagnosis and follow up. The Neurosensory retina (NSR) map is defined as the thickness between the inner limiting membrane layer and the inner aspect of the retinal pigment epithelium complex. Additionally, the NSR map has been used to differentiate between healthy and subjects with macular problems, but the plotting of the retinal thickness map depends critically on additional manufacturer interpretation software to automatically drawing. Therefore, this paper presents an end-to-end 3D convolutional neural network to automatically extract nine thickness mean values to draw the NSR map from an SD-OCT.

Fat deposition in the human body can be imaged using MR. MR fat imaging is used in studies ranging from examining body habitus to evaluating organ composition. A detailed method of segmenting and quantifying fat by post-processing MR image volumes using Synapse 3D and custom Matlab softwares is presented.

Voxel-based morphometry (VBM) is a technique used to detect and localize morphological differences in brain anatomy. Its design includes the adjustment of several parameters that influence final results, including the spatial normalization method used during preprocessing. VBM is widely applied in human studies, yet its usage for experimental animal models is scarce in the literature. Moreover, there is no a quantitative analysis nor validation of the method’s results when applied to small animal brain images. In this work we evaluated results of different VBM workflows by simulating tissue changes on the mouse brain. Using an automatic method we simulated atrophies on specific areas of the mouse brain cortex and cerebellum and created two experimental groups. We applied different VBM configurations to detect tissue changes between control group and each experimental group and compare its results with the location and extension of the atrophies simulated. Finally, we varied the size of the experimental group by randomly selecting sets of 9, 8 and 7 individuals and applied optimized VBM between control group and each experimental subset. Our results empirically demonstrated that optimized VBM was the most effective method to detect differences between groups. Detection was accurate on the brain cortex, while on the cerebellum the accuracy was reduced being the extension of findings smaller than expected and variable along different experimental group conformations. Our results suggested that a wider smoothing kernel is needed in this brain region, in order to minimize misalignments between individuals and obtain more consistent findings.

Polygenic risk scores (PRSs) may be used to investigate the effects of genetic risk for disease on complex human traits. Here we set out to determine how the overall genetic risk for Alzheimer’s disease (AD) shapes brain structure in non-AD populations. PRS scores were computed using results from the International Genomics of Alzheimer's Project (IGAP) study. PRSs were computed at 14 different significance thresholds in the IGAP results. The effect of PRS as a predictor of brain morphometry was mapped voxelwise on brain structure as determined by tensor-based morphometry (TBM) in three cohorts: ADNI1, ADNI2, and HUNT. Our multi-cohort TBM framework first tests associations in each cohort individually, then meta-analyzes findings in a common space. Higher PRS for AD was associated with greater ventricular and lower hippocampal volumes. Associations remained after removing the major AD risk gene, APOE, from the PRS. This cumulative influence of common genetic variants on brain-wide structural variation in nondemented individuals may pinpoint genetic and neurological pathways that contribute to the preclinical assessment of disease risk.

The corpus callosum (CC) is the main neural pathway that communicates information between the brain’s hemispheres. Impairment of this pathway is evident in neurogenetic and developmental disorders, neurodegenerative diseases, and in many major psychiatric disorders, making the CC the focus of intense study. Prior studies often require manual input for segmentation, or have been single site, single modality, or fail to report on the reliability and generalizability of segmentations. We develop a Robust Automatic Corpus Callosum Analysis Toolkit (RACCAT) that segments the midsagittal CC from T1-weighted images, guided by diffusion MRI where available, to facilitate large-scale multimodal CC studies of its global, regional, and local (pointwise) structure. RACCAT was applied to data from 772 individuals aged 3-21 from the Pediatric Imaging, Neurocognition, and Genetics study, a developmental cohort imaged using multiple scanners and imaging protocols. CC area and fractional anisotropy were associated with age but also with site and scanner manufacturer; CC curvature also showed significant age associations but showed no detectable association with the scanning site, making it a robust developmental biomarker for multisite studies.

Medical Image Registration (MIR) is a challenge that arises in many image processing applications when multiple images must be aligned. We deal with the case of multi-modalities that is approached as an optimization problem. In MIR deterministic algorithms are used mainly, the disadvantage is that many of them get trapped in local optima, especially in multi-modal registration. This work aims to overcome this disadvantage using Scatter Search and Particle Swarm as optimization algorithms, with the mutual information approach proposed by Mattes et. al. The proposed optimizers were tested and contrasted with Reference Algorithms. A multi-modal rigid 3D / 3D of brain medical image registration scheme was implemented, and it was validated in RIRE project. The qualitative and quantitative validation of the results was satisfactory; the results demonstrated the accuracy and applicability of the proposed methods in comparison to conventional methods, as well as not being trapped in local optima.

Diffusion-weighted MRI (dMRI) offers a range of measures that are sensitive to brain aging and neurodegeneration. Here we analyzed data from 318 participants (mean age: 75.4±7.9 years; 143 men/175 women) from the third phase of the Alzheimer’s Disease Neuroimaging Initiative (ADNI3), who were each scanned with one of six different diffusion MRI protocols using scanners from three different manufacturers. We computed 4 standard diffusion tensor imaging (DTI) anisotropy and diffusivity indices, and one advanced anisotropy index based on the tensor distribution function (TDF), in 24 white matter regions of interest. Modeling protocol effects, we ranked the diffusion indices for their strength of correlation with 3 standard clinical measures of cognitive impairment: the ADAS-Cog, MMSE, and sum-of-boxes Clinical Dementia Rating. Across all dMRI indices and cognitive measures, the cingulum-hippocampal region and the uncinate showed some of the strongest associations with cognitive impairment; largest effect sizes were detected with axial diffusivity (Ax^DDTI). While fractional anisotropy (FA) derived from the DTI model was the weakest in detecting associations with cognitive measures, FA derived from the TDF detected widespread, robust associations. Protocol differences affected dMRI indices; however by modeling protocol effects, we were able to pool dMRI data from multiple acquisition protocols and detect consistent associations with cognitive impairment and age. dMRI indices computed from the upgraded scanning protocols in ADNI3 were sensitive to cognitive impairment in brain aging, offering a benchmark to compare to future multi-shell or multi-compartment diffusion indices.

Traumatic brain injury (TBI) is a major public health issue around the world. Pediatric TBI patients are at risk of long-term disabilities, as a brain injury sustained during development can affect on-going maturational processes. The white matter (WM) in particular is vulnerable, as myelination continues into the third decade of life and beyond, and poor myelination of tracts can result in decreased integration within brain networks. In addition, variability and heterogeneity are hallmarks of TBI, e.g., injury-related variables and symptoms. These issues combined with small sample sizes limit the power and generalizability of individual studies. In the present study, we employed a meta-analytic approach, combining data across 4 pediatric TBI samples resulting in 104 TBI (75M/29F) and 114 control participants (70M/44F) between 7-18 years, using harmonized processing and analysis as part of the ENIGMA consortium (Enhancing NeuroImaging Genetics through Meta-Analysis). We report lower fractional anisotropy (FA) values in TBI patients across several post-injury windows, particularly in central WM tracts. Within the TBI patient group, we also report marginally significant results of lower FA in younger TBI patients, patients scanned closer to time of injury, and female patients. Although this meta-analytic approach yielded the largest sample size reported yet in pediatric moderate/severe TBI (msTBI) neuroimaging, our trends indicate that larger sample sizes are needed in further studies. As additional cohorts join the ENIGMA Pediatric moderate/severe TBI (msTBI) effort, more robust effects will be revealed.

Voxel Based Morphometry (VBM) is a methodology for medical image analysis that can be used to detect GM decrease in brain images at a group level. The performance of VBM with DARTEL registration method has already been studied, but not in a single subject. Classic VBM is not expected to detect GM decrease in healthy patients, but, at single subject level some differences could be detected, due to the complexity of sulcal patterns and inter and intra-individual anatomical variability. Thus, this study assesses the effects of DARTEL method in single subject VBM for healthy subjects. We applied a 2 sample T-test with 2 covariates, age and gender. The p-values were corrected using False Discovery Rate (FDR) (p<0.05), using a cluster extent threshold of k>15. We used 3 group sizes, 20, 50 and 100, and four different smoothing kernel of 3, 5, 8 and 10 mm. Each subject was compared against the rest of its group. We observed that the single subject VBM detected GM decrease in healthy patients for all the group sizes and smoothing. In spite of that, detected voxels were different for almost all the patients. For most of the cases, the maximum percentage of subjects with the same voxel detected was less than 5% and of all the detections, less than 5% were presented in more than 1 subject.

This paper aims to automatically describe changes occurring in brain regions of patients from the OASIS database as follows: 66 control patients (CN), 20 patients diagnosed with mild Alzheimer’s disease (AD) and 50 with mild cognitive impairment (MCI). A regional-based morphometry method is proposed to explore the location of anatomical differences functionally connected between AD, MCI and CN subjects. A first step provides a set of regions with statistically significant volume differences which are then challenged by a classification task, providing a second set of regions that demonstrate better performance when separating the groups. Afterward, connectivity between these regions is analyzed to establish how functionally connected these regions are. Results demonstrate the disease follows functional patterns rather than anatomical ones.

Alzheimer’s disease is heterogeneous and despite some consistent neuropathological hallmarks, different clinical forms have been identified, including non-amnestic presentations. Even in amnestic forms, the presentation of the disease can differ across individuals, in terms of age of onset, dynamics of progression and specific impairment profiles. Different distributions of neurofibrillary tangles exist in AD, and these are linked with structural differences detectable on ante-mortem MRI , but these are hard to identify in the earlier stages of disease. In the present work, we validate and test a previously proposed method for identifying subtypes of cortical atrophy in AD, based on MRI data from an independent case/control study of individuals defined by pathophysiological biomarkers. We implemented a clustering method based on the Louvain modularity method, and tested it across a range of pre-processing parameters. Our cohort of participants was comprised of 111 participants (mean age: 67.7 year; range: 51-91), including 37 cognitively normal controls, 43 prodromal AD, and 31 demented AD patients. We identified 4 patient clusters with distinct atrophy patterns either predominantly in the temporal lobes (groups 0 and 1), in the parietal and temporal lobes (group 2), or in the frontal and temporal lobes (group 3). Further evaluation of neuro-psychological characteristics of each patient cluster will be carried out in the future. In conclusion, the modularity-based clustering method may help to identify specific subtypes of atrophy in neurological diseases such as AD.

Parkinson's disease (PD) is a progressive neurodegenerative disorder in which patients show progressively worsening motor symptoms, often followed by cognitive impairment and dementia. Brain MRI can be used to identify patterns of neurodegeneration that are characteristic of PD, but the spatial pattern of brain abnormalities is still not well understood. “Sulcus-based morphometry” provides measures of the cortical fissures of the brain that reflect degenerative changes in relation to neuropsychiatric disease. Extracting sulci requires good contrast between the gray matter and the CSF, and less well-defined sulci may be difficult to extract reliably. Before embarking on a study of sulcal abnormalities in PD, we set out to determine the reliability of measures from 123 sulci, defined by an existing atlas, using publicly available test-retest data from 8 cohorts. Of the 123 atlas-defined sulci, several major sulci were broken down into smaller regions (e.g., the superior temporal sulcus was divided into the main STS, the anterior terminal ascending branch of STS and the posterior terminal ascending branch of STS); we assessed reliability in each individually, and after merging the portions of the sulci together, in a newly defined, concatenated atlas. For 467 subjects from the PPMI cohort (http://www.ppmiinfo. org ;age range: 61.5 ± 10.1 years), we segmented and labeled major sulci and extracted 4 shape descriptors for each: length, depth, surface area, and width. We then aimed to establish the profile of case-control differences for 3 candidate sulci of interest: the central sulcus, superior temporal sulcus and the calcarine fissure. These sulci were among the more robust in terms of reproducibility; we found that the calcarine width was associated with PD, offering new features for genetic and interventional studies of PD.

Diffusion MRI (dMRI) is widely used to study the brain’s white matter (WM) microstructure in a range of psychiatric and neurological diseases. As the diffusion tensor model has limitations in brain regions with crossing fibers, novel diffusion MRI reconstruction models may offer more accurate measures of tissue properties, and a better understanding of the brain abnormalities in specific diseases. Here we studied a large sample of 249 participants with 22q11.2 deletion syndrome (22q11DS), a neurogenetic condition associated with high rates of developmental neuropsychiatric disorders, and 224 age-matched healthy controls (HC) (age range: 8-35 years). Participants were scanned with dMRI at eight centers worldwide. Using a meta-analytic approach, we assessed the profile of group differences in four diffusion anisotropy measures to better understand the patterns of WM microstructural abnormalities and evaluate their consistency across alternative measures. When assessed in atlas-defined regions of interest, we found statistically significant differences for all anisotropy measures, all showing a widespread but not always coinciding pattern of effects. The tensor distribution function fractional anisotropy (TDF-FA) showed largest effect sizes all in the same direction (greater anisotropy in 22q11DS than HC). Fractional anisotropy based on the tensor model (FA) showed the second largest effect sizes after TDF-FA; some regions showed higher mean values in 22q11DS, but others lower. Generalized fractional anisotropy (GFA) showed the opposite pattern to TDF-FA with most regions showing lower anisotropy in 22q11DS versus HC. Anisotropic power maps (AP) showed the lowest effect sizes also with a mixed pattern of effects across regions. These results were also consistent across skeleton projection methods, with few differences when projecting anisotropy values from voxels sampled on the FA map or projecting values from voxels sampled from each anisotropy map. This study highlights that different mathematical definitions of anisotropy may lead to different profiles of group differences, even in large, well-powered population studies. Further studies of biophysical models derived from multi-shell dMRI and histological validations may help to understand the sources of these differences. 22q11DS is a promising model to study differences among novel anisotropy/dMRI measures, as group differences are relatively large and there exist animal models suitable for histological validation.

Advances in the treatment of cancer, including surgery, chemotherapy and radiation therapy, have led to an increase in the survival rate of children with brain tumors. However, the efficacy of these therapies is often overshadowed by the long term neurological consequences of treatment-induced injuries. Diffusion weighted imaging, a magnetic resonance imaging technique, allows us to measure changes in white matter in a population of posterior fossa brain tumor survivors who had two different treatment schemes: surgery + chemotherapy (S+C) and surgery, chemotherapy + cranial irradiation (S+C+R). The results of our analysis reveal significantly lower mean diffusivity (MD) and lower radial diffusivity (RD) values in the posterior thalamic radiation in the S+C+R group, which may indicate more myelin or more axonal damage in the S+C group compared to the S+C+R group. While it is possible that this may be related to a more intensive chemotherapeutic regimen in the S+C group, more work will be forthcoming to produce a clearer picture of treatment-related injury in survivors of posterior fossa tumors in childhood. These preliminary findings will be further analyzed to include demographic factors, neuropsychological data, and radiation dose values.

The Convolutional neural networks (CNN) have been shown to be able to learn the relevant visual features for different computer vision tasks from large amounts of annotated data. Hence, the performance of CNNs can vary depending on the training data set and associated model architecture. This article presents a comparative analysis of the robustness and sensitivity of different CNN architectures to classify invasive breast cancer tissue. Our experiment involved a comparison of six CNN architectures with different depths (number of layers), specifically trained to detect invasive breast cancer from digitized pathology images. Additionally, the pre-trained model VGG 16 (trained to classify natural images) was added as the seventh architecture. Each of the models was trained with two different data sets: a cohort of 239 breast cancer slide images from the Hospital of the University of Pennsylvania (HUP), and another with 172 digitized breast cancer images from the Cancer Genome Atlas (TCGA). In addition, in each case the training was validated with 40 breast cancer slide images from the New Jersey Cancer Institute (CINJ). The last layer of the VGG 16 model was modified to allow classification of the binary problem (presence or absence of invasive ductal carcinoma). The experimental results show a performance of greater than 93% in terms of AUC (Area Under the ROC Curve) for the CNNs trained specifically with cases of invasive breast cancer from the TCGA. However, we also note that VGG-CNN-16 achieves an AUC of 92.43% and 86.87% respectively, despite the fact that it was trained for a different domain.

Automatic detection and quantification of glands in gastric cancer may contribute to objectively measure the lesion severity, to develop strategies for early diagnosis, and most importantly to improve the patient categorization. This article presents an entire framework for automatic detection of glands in gastric cancer images. This approach starts by selecting gland candidates from a binarized version of the hematoxylin channel. Next, the gland’s shape and nuclei are characterized using local features which feed a Monte Carlo Cross validation method classifier trained previously with manually labeled images. Validation was carried out using a dataset with 1330 annotated structures (2372 glands) from seven fields of view extracted from gastric cancer whole slide images. Results showed an accuracy of 93% using a simple linear classifier. The presented strategy is quite simple, flexible and easily adapted to an actual pathology laboratory.

Prostate cancer diagnosis is performed by pathologists through the analysis of tissue samples from the prostate gland using a microscope. The development of automatic acquisition and digitalization technologies has allowed the construction of large collections of digitalized histopathology slides, that are usually accompanied by clinical information and other types of metadata. This collection of cases, along with the metadata, has the potential to be an invaluable resource for the analysis of new challenging cases supporting diagnosis, prognosis, and theragnosis decision tasks. This paper presents a multimodal retrieval system based on a supervised multimodal kernel semantic embedding model that supports the search of relevant cases in a multimodal database, combining both images, i.e. histopathology slides, and text, i.e. pathologist’s reports. The system was tested in a multimodal prostate adenocarcinoma dataset, composed of whole slide images of tissue samples, pathologist’s reports and gradation information using the Gleason score. The system shows a high performance for multimodal information retrieval with a Mean Average Precision of 0.6263.

Dermatopathology education meaningfully relies on consultation of books, which are expensive, quickly outdated and have limited possibilities. In recent years, virtual microscopy, a method that enables examination of digitized microscopy samples, has earn interest because of its possibilities in terms of interaction, availability, usability, low costs and adaptation to multiple clinic scenarios. This work introduces a customized low-cost system for consultation of dermatopathology samples. First, physical slides are digitized using an optical microscope coupled to a digital camera controlled by a custom-motorized scanner. Then, digitized images are automatically stitched to assembly the Whole Slide Image (WSI). A web application, developed using open source tools, gives access to such WSI and allows users to interact with the content by panning and zooming. The application also allows to hand-free annotate specific regions. A set of 100 dermatopathology slides, provided by the Pathology Department of the Universidad Nacional de Colombia, representing basic lesions and inflammatory skin diseases (based on Ackerman patterns) were digitized. Each WSI contains diagnosis and annotations of relevant regions. The platform is currently being used by trainees who highlight the benefits of this kind of tools that complement their training and help to improve their diagnostic skills.

An indirect method of tissue consistency measurement is proposed, based on intensity and texture features of conventional ultrasound (US) cervix images. Calibration and validation were carried out in five phantoms simulating different cervical firmness, as well as in short and long cervices. Several image features attributed to the histogram, the co–occurrence matrix and the run–length encoding matrix were extracted and analyzed to evaluate their ability to distinguish between degrees of phantoms’ firmness. The most indicative of firmness indices were selected by correlating their values with the phantoms’ elasticities determined through Young’s moduli. Also, a random forest classifier was implemented, allowing to identify the features that contribute the most to class separation between phantoms. Using both tests, six features were selected: mean, standard deviation, entropy, skewness and two RLE-matrix features. A 6–fold cross validation was used to evaluate the model, obtaining a 98.9±0.79% accuracy. Finally, a preliminary case study was conducted upon closed and opened cervical US images, classifying them between both groups using a random forest model, obtaining an 84.34% accuracy. The indicated tests show that intensity and texture features extracted from conventional US images provide indirect and less–invasive information than other methods regarding tissue consistency, and therefore may be used to measure changes in cervical firmness.

Ultrasound (US) images are necessary in obstetrics because they provide the most important clinical parameters for fetal health assessment during the second and third trimesters: head circumference, biparietal diameter, abdominal circumference and femur length. These fetometric indices are helpful for gestational age and fetal weight estimation; they are also helpful for obstetricians to diagnose fetal development abnormalities. However, these indices are obtained manually, which provokes high intra and interobserver variability and lack of repeatability. A fully automatic method to segment and measure femur’s length is presented in this paper. The proposed methodology incorporates texture information and introduces a novel curvature analysis to adequately detect the femur. It consists on pre–processing US images with an anisotropic diffusion filter, followed by morphological operations and thresholding to isolate femur–candidate regions. A normalized metric composed of intensity, length, centroid position and entropy is assigned to each region in order to select the most probable candidate to be femur. This selected region is afterwards thinned to a one–pixel line, whose curvature is analyzed with an angle threshold criterion to accurately locate femur’s extrema. The method was tested on 64 US images (20 taken on the second and 44 on the third trimester of pregnancy); a correlation coefficient of 0.984 and an error of 1.016±2.764 mm were achieved between expert–obtained manual measures and automatically calculated indices. Results are consistent, outperform those reported previously by other authors and show a high correlation with measures obtained by experts; therefore, the developed method is suitable to be adapted for clinical use.

In this work we present a combination of segmentation and motion estimation methods applied to left ventricle evaluation in fetal echocardiographic images which are used for prenatal diagnosis. In our proposed scheme, several features of the ultrasound images are computed and used for both algorithms. A multiresolution framework is employed for the segmentation and motion estimation tasks. The segmentation is achieved using a multi-texture active appearance model based on the Hermite transform. The analysis is performed using the appearance models provided by Hermite coefficients up to third order. The multiresolution approach allows to obtain a robust segmentation to extract the shape of the left ventricle. The obtained results in the segmentation step are used for the motion estimation algorithm. The left ventricle is the structure used for evaluation. The main goal is to determinate the heart movement of fetal heart which can be used for disease detection, characterization and further analysis. Results of the motion estimation process are analyzed and compared with other techniques applied to heart ultrasound data.

In this paper we propose a semi-automatic method to segment the fetal cerebellum in ultrasound images. The method is based on an active shape model which includes profiles of Hermite features. In order to fit the shape model we used a PCA of Hermite features. This model was tested on ultrasound images of the fetal brain taken from 20 pregnant women with gestational weeks varying from 18 to 24. Segmentation results compared to manual annotation show a mean Hausdorff distance of 6.85 mm using a conventional active shape model trained with gray profiles, and a mean Hausdorff distance of 5.67 mm using Hermite profiles. We conclude that the Hermite profile model is more robust in segmenting fetal cerebellum in ultrasound images.

Accurate volume quantification in magnetic resonance imaging (MRI) provides an important cardiac indicator in congenital heart diseases and furthermore, it is crucial for any surgical planning of congenital surgery. This paper presents an automatic segmentation of the left ventricle (LV) in congenital heart diseases. The proposed approach is basically the suite of three steps: first, a radial saliency analysis coarsely approximates the myocardium boundary. Second, this boundary is refined by choosing, among the candidate points, those ones that follow the most ellipsoidal closed curve. Third, these points serve as the external energy of a conventional snake that is evolved to approximate the inner myocardium boundary. This method requires a minimum parameterization and demands low computational power, in fact, a whole case is segmented in 80 s. The strategy was evaluated using 10 cardiac MRI volumes of actual congenital diseases provided by the HVSMR 2016 challenges, achieving an average Dice of 0.77.

Rotational X-ray coronary angiography is a medical imaging technique safe and effective in identifying of the luminal disease, which considers a significant reduction in radiation exposure and contrast medium volume compared to conventional angiography. The main objective of this research is to propose a computational approach to automatically extract a description of the morphopatological shape of the left coronary artery by means the centerlines of this vessel. The proposal is based on a sequential design which involves image enhancement, identification of all the types of vascular points belonging to the vascular system, construction of the coronary tree and tracking of the centerlines along the rotational angiography sequence. Some results obtained after applying this method to monoplane rotational X–ray image sequences are presented.

This paper presents a novel method for the automatic design of binary descriptors for the detection of coronary arteries in X-ray angiograms. The method is divided into two different steps for detection and segmentation. In the step of automatic vessel detection, the metaheuristic of iterated local search (ILS) is used for the design of optimal binary descriptors for detecting vessel-like structures by using the top-hat transform in the spatial image domain. The detection results are compared with those obtained by five state-of-the-art vessel enhancement methods. The proposed method obtained the highest detection results in terms of the area (A_z ) under the ROC curve (A_z = 0.9635) using a training set of 50 angiograms, and A_z = 0.9544 with an independent test set of 50 X-ray images. In the segmentation step, the inter-class variance thresholding method was applied to classify vessel and nonvessel pixels from the top-hat filter response obtained from the binary descriptor. According to the experimental results, the vessel detection by using an automatically generated binary descriptor can be highly suitable for computer-aided diagnosis.

Analysis of cardiac images has become fundamental for heart evaluation. It is well known that cardiac affections constitute one of the main causes of death in developed and developing countries. An early detection of cardiac diseases might substantially contribute to find the correct treatment. In this work, we propose a variational framework for left ventricle (LV) segmentation of cardiac MR and CT volumes. The method is based on a multiscale scheme provided by the Hermite transform (HT) which is used for local image feature coding. The variational approach includes several functional terms embedded into a level set - based framework in which geometrical and image features computed from coefficients of the HT are processed. Methods based on level sets are commonly configured using a set of parameters which are frequently selected experimentally. In this paper, we present an automatic mechanism for parameters selection using the contrast information obtained from the input data. The method was evaluated on several cardiac CT and MR volumes. Distance metrics were used for evaluation by comparing with manual segmentations.