Automated Segmentation Tools Research Articles

Whole brain and deep gray matter structure segmentation: Quantitative comparison between MPRAGE and MP2RAGE sequences.

T1-weighted MRI images are commonly used for volumetric assessment of brain structures. Magnetization prepared 2 rapid gradient echo (MP2RAGE) sequence offers superior gray (GM) and white matter (WM) contrast. This study aimed to quantitatively assess the agreement of whole brain tissue and deep GM (DGM) volumes obtained from MP2RAGE compared to the widely used MP-RAGE sequence. Twenty-nine healthy participants were included in this study. All subjects underwent a 3T MRI scan acquiring high-resolution 3D MP-RAGE and MP2RAGE images. Twelve participants were re-scanned after one year. The whole brain, as well as DGM segmentation, was performed using CAT12, volBrain, and FSL-FAST automatic segmentation tools based on the acquired images. Finally, contrast-to-noise ratio between WM and GM (CNRWG), the agreement between the obtained tissue volumes, as well as scan-rescan variability of both sequences were explored. Significantly higher CNRWG was detected in MP2RAGE vs. MP-RAGE (Mean ± SD = 0.97 ± 0.04 vs. 0.8 ± 0.1 respectively; p<0.0001). Significantly higher total brain GM, and lower cerebrospinal fluid‏ volumes were obtained from MP2RAGE vs. MP-RAGE based on all segmentation methods (p<0.05 in all cases). Whole-brain voxel-wise comparisons revealed higher GM tissue probability in the thalamus, putamen, caudate, lingual gyrus, and precentral gyrus based on MP2RAGE compared with MP-RAGE. Moreover, significantly higher WM probability was observed in the cerebellum, corpus callosum, and frontal-and-temporal regions in MP2RAGE vs. MP-RAGE. Finally, MP2RAGE showed a higher mean percentage of change in total brain GM compared to MP-RAGE. On the other hand, MP-RAGE demonstrated a higher overtime percentage of change in WM and DGM volumes compared to MP2RAGE. Due to its higher CNR, MP2RAGE resulted in reproducible brain tissue segmentation, and thus is a recommended method for volumetric imaging biomarkers for the monitoring of neurological diseases.

PO-1659 Clinical validation of an automatic atlas-based segmentation tool for male pelvis CT images

PO-1659 Clinical validation of an automatic atlas-based segmentation tool for male pelvis CT images

Coronary artery segmentation under class imbalance using a U-Net based architecture on computed tomography angiography images

Coronary artery disease is caused primarily by vessel narrowing. Extraction of the coronary artery area from images is the preferred procedure for diagnosing coronary diseases. In this study, a U-Net-based network architecture, 3D Dense-U-Net, was adopted to perform fully automatic segmentation of the coronary artery. The network was applied to 474 coronary computed tomography (CT) angiography scans performed at Wanfang Hospital, Taiwan. Of these, 10% were used for testing. The CT scans were divided into patches of 16 original high-resolution slices. The slices were overlapped between patches to take advantage of surrounding imaging information. However, an imbalance between the foreground and background presents a challenge in smaller-object segmentation such as with coronary arteries. The network was optimized and achieved a promising result when the focal loss concept was adopted. To evaluate the accuracy of the automatic segmentation approach, the dice similarity coefficient (DSC) was calculated, and an existing clinical tool was used. The subjective ratings of three experienced radiologists were used to compare the two ratings. The results show that the proposed approach can achieve a DSC of 0.9691, which is significantly higher than other studies using a deep learning approach. In the main trunk, the results of automatic segmentation agree with those of the clinical tool; they were significantly better in some small branches. In our study, automatic segmentation tool shows high-performance detection in coronary lumen vessels, thereby providing potential power in assisting clinical diagnosis.

Large-scale, multi-vendor, multi-protocol, quality-controlled analysis of clinical cine CMR using artificial intelligence

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Advancing Impact Award scheme of the EPSRC Impact Acceleration Account at King’s College London Background Artificial intelligence (AI) has the potential to facilitate the automation of CMR analysis for biomarker extraction. However, most AI algorithms are trained on a specific input domain (e.g., scanner vendor or hospital-tailored imaging protocol) and lack the robustness to perform optimally when applied to CMR data from other input domains. Purpose To develop and validate a robust CMR analysis tool for automatic segmentation and cardiac function analysis which achieves state-of-the-art performance for multi-vendor short-axis cine CMR images. Methods The current work is an extension of our previously published quality-controlled AI-based tool for cine CMR analysis [1]. We deployed an AI algorithm that is equipped to handle different image sizes and domains automatically - the ‘nnU-Net’ framework [2] - and retrained our tool using the UK Biobank (UKBB) cohort population (n = 4,872) and a large database of clinical CMR studies obtained from two NHS hospitals (n = 3,406). The NHS hospital data came from three different scanner types: Siemens Aera 1.5T (n = 1,419), Philips Achieva 1.5T and 3T (n = 1,160), and Philips Ingenia 1.5T (n = 827). The ‘nnU-net’ was used to segment both ventricles and the myocardium. The proposed method was evaluated on randomly selected test sets from UKBB (n = 488) and NHS (n = 331) and on two external publicly available databases of clinical CMRs acquired on Philips, Siemens, General Electric (GE), and Canon CMR scanners – ACDC (n = 100) [3] and M&amp;Ms (n = 321) [4]. We calculated the Dice scores - which measure the overlap between manual and automatic segmentations - and compared manual vs AI-based measures of biventricular volumes and function. Results Table 1 shows that the Dice scores for the NHS, ACDC, and M&amp;Ms scans are similar to those obtained in the highly controlled, single vendor and single field strength UKBB scans. Although our AI-based tool was only trained on CMR scans from two vendors (Philips and Siemens), it performs similarly in unseen vendors (GE and Canon). Furthermore, it achieves state-of-the-art performance in online segmentation challenges, without being specifically trained on these databases. Table 1 also shows good agreement between manual and automated clinical measures of ejection fraction and ventricular volume and mass. Conclusions We show that our proposed AI-based tool, which combines training on a large-scale multi-domain CMR database with a state-of-the-art AI algorithm, allows us to robustly deal with routine clinical data from multiple centres, vendors, and field strengths. This is a fundamental step for the clinical translation of AI algorithms. Moreover, our method yields a range of additional metrics of cardiac function (filling and ejection rates, regional wall motion, and strain) at no extra computational cost.

Norms for Automatic Estimation of Hippocampal Atrophy and a Step Forward for Applicability to the Italian Population.

IntroductionHippocampal volume is one of the main biomarkers of Alzheimer’s Dementia (AD). Over the years, advanced tools that performed automatic segmentation of Magnetic Resonance Imaging (MRI) T13D scans have been developed, such as FreeSurfer (FS) and ACM-Adaboost (AA). Hippocampal volume is considered abnormal when it is below the 5th percentile of the normative population. The aim of this study was to set norms, established from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) population, for hippocampal volume measured with FS v.6.0 and AA tools in the neuGRID platform (www.neugrid2.eu) and demonstrate their applicability for the Italian population.MethodsNorms were set from a large group of 545 healthy controls belonging to ADNI. For each pipeline, subjects with segmentation errors were discarded, resulting in 532 valid segmentations for FS and 421 for AA (age range 56–90 years). The comparability of ADNI and the Italian Brain Normative Archive (IBNA), representative of the Italian general population, was assessed testing clinical variables, neuropsychological scores and normalized hippocampal volumes. Finally, percentiles were validated using the Italian Alzheimer’s disease Repository Without Borders (ARWiBo) as external independent data set to evaluate FS and AA generalizability.ResultsHippocampal percentiles were checked with the chi-square goodness of fit test. P-values were not significant, showing that FS and AA algorithm distributions fitted the data well. Clinical, neuropsychological and volumetric features were similar in ADNI and IBNA (p > 0.01). Hippocampal volumes measured with both FS and AA were associated with age (p < 0.001). The 5th percentile thresholds, indicating left/right hippocampal atrophy were respectively: (i) below 3,223/3,456 mm3 at 56 years and 2,506/2,415 mm3 at 90 years for FS; (ii) below 4,583/4,873 mm3 at 56 years and 3,831/3,870 mm3 at 90 years for AA. The average volumes computed on 100 cognitively intact healthy controls (CN) selected from ARWiBo were close to the 50th percentiles, while those for 100 AD patients were close to the abnormal percentiles.DiscussionNorms generated from ADNI through the automatic FS and AA segmentation tools may be used as normative references for Italian patients with suspected AD.

3D segmentation of neuronal nuclei and cell-type identification using multi-channel information

BackgroundAnalyzing images to accurately estimate the number of different cell types in the brain using automatic methods is a major objective in neuroscience. The automatic and selective detection and segmentation of neurons would be an important step in neuroanatomical studies. New methodWe present a method to improve the 3D reconstruction of neuronal nuclei that allows their segmentation, excluding the nuclei of non-neuronal cell types. ResultsWe have tested the algorithm on stacks of images from rat neocortex, in a complex scenario (large stacks of images, uneven staining, and three different channels to visualize different cellular markers). It was able to provide a good identification ratio of neuronal nuclei and a 3D segmentation.Comparison with Existing Methods: Many automatic tools are in fact currently available, but different methods yield different cell count estimations, even in the same brain regions, due to differences in the labeling and imaging techniques, as well as in the algorithms used to detect cells. Moreover, some of the available automated software methods have provided estimations of cell numbers that have been reported to be inaccurate or inconsistent after evaluation by neuroanatomists. ConclusionsIt is critical to have a tool for automatic segmentation that allows discrimination between neurons, glial cells and perivascular cells. It would greatly speed up a task that is currently performed manually and would allow the cell counting to be systematic, avoiding human bias. Furthermore, the resulting 3D reconstructions of different cell types can be used to generate models of the spatial distribution of cells.

DEEP LEARNING FOR AUTOMATIC BUILDING DAMAGE ASSESSMENT: APPLICATION IN POST-DISASTER SCENARIOS USING UAV DATA

Abstract. During the last few years, the technical and scientific advances in the Geomatics research field have led to the validation of new mapping and surveying strategies, without neglecting already consolidated practices. The use of remote sensing data for damage assessment in post-disaster scenarios underlined, in several contexts and situations, the importance of the Geomatics applied techniques for disaster management operations, and nowadays their reliability and suitability in environmental emergencies is globally recognized. In this paper, the authors present their experiences in the framework of the 2016 earthquake in Central Italy and the 2019 Cyclone Idai in Mozambique. Thanks to the use of image-based survey techniques as the main acquisition methods (UAV photogrammetry), damage assessment analysis has been carried out to assess and map the damages that occurred in Pescara del Tronto village, using DEEP (Digital Engine for Emergency Photo-analysis) a deep learning tool for automatic building footprint segmentation and building damage classification, functional to the rapid production of cartography to be used in emergency response operations. The performed analyses have been presented, and the strengths and weaknesses of the employed methods and techniques have been outlined. In conclusion and based on the authors' experience, some operational suggestions and best practices are provided and future research perspectives within the same research topic are introduced.

Capsule Network-based architectures for the segmentation of sub-retinal serous fluid in optical coherence tomography images of central serous chorioretinopathy.

Central serous chorioretinopathy (CSCR) is a chorioretinal disorder of the eye characterized by serous detachment of the neurosensory retina at the posterior pole of the eye. CSCR results from the accumulation of subretinal fluid (SRF) due to idiopathic defects at the level of the retinal pigment epithelial (RPE) that allows serous fluid from the choriocapillaris to diffuse into the subretinal space between RPE and neurosensory retinal layers. This condition is presently investigated by clinicians using invasive angiography or non-invasive optical coherence tomography (OCT) imaging. OCT images provide a representation of the fluid underlying the retina, and in the absence of automated segmentation tools, currently only a qualitative assessment of the same is used to follow the progression of the disease. Automated segmentation of the SRF can prove to be extremely useful for the assessment of progression and for the timely management of CSCR. In this paper, we adopt an existing architecture called SegCaps, which is based on the recently introduced Capsule Networks concept, for the segmentation of SRF from CSCR OCT images. Furthermore, we propose an enhancement to SegCaps, which we have termed as DRIP-Caps, that utilizes the concepts of Dilation, Residual Connections, Inception Blocks, and Capsule Pooling to address the defined problem. The proposed model outperforms the benchmark UNet architecture while reducing the number of trainable parameters by 54.21%. Moreover, it reduces the computation complexity of SegCaps by reducing the number of trainable parameters by 37.85%, with competitive performance. The experiments demonstrate the generalizability of the proposed model, as evidenced by its remarkable performance even with a limited number of training samples. Graphical abstract is mandatory please provide.

Interactive contouring through contextual deep learning.

To investigate a deep learning approach that enables three-dimensional (3D) segmentation of an arbitrary structure of interest given a user provided two-dimensional (2D) contour for context. Such an approach could decrease delineation times and improve contouring consistency, particularly for anatomical structures for which no automatic segmentation tools exist. A series of deep learning segmentation models using a Recurrent Residual U-Net with attention gates was trained with a successively expanding training set. Contextual information was provided to the models, using a previously contoured slice as an input, in addition to the slice to be contoured. In total, 6 models were developed, and 19 different anatomical structures were used for training and testing. Each of the models was evaluated for all 19 structures, even if they were excluded from the training set, in order to assess the model's ability to segment unseen structures of interest. Each model's performance was evaluated using the Dice similarity coefficient (DSC), Hausdorff distance, and relative added path length (APL). The segmentation performance for seen and unseen structures improved when the training set was expanded by addition of structures previously excluded from the training set. A model trained exclusively on heart structures achieved a DSC of 0.33, HD of 44mm, and relative APL of 0.85 when segmenting the spleen, whereas a model trained on a diverse set of structures, but still excluding the spleen, achieved a DSC of 0.80, HD of 13mm, and relative APL of 0.35. Iterative prediction performed better compared to direct prediction when considering unseen structures. Training a contextual deep learning model on a diverse set of structures increases the segmentation performance for the structures in the training set, but importantly enables the model to generalize and make predictions even for unseen structures that were not represented in the training set. This shows that user-provided context can be incorporated into deep learning contouring to facilitate semi-automatic segmentation of CT images for any given structure. Such an approach can enable faster de-novo contouring in clinical practice.

Validating Automated Segmentation Tools in the Assessment of Caudate Atrophy in Huntington's Disease.

Background: Neuroimaging shows considerable promise in generating sensitive and objective outcome measures for therapeutic trials across a range of neurodegenerative conditions. For volumetric measures the current gold standard is manual delineation, which is unfeasible for samples sizes required for large clinical trials.Methods: Using a cohort of early Huntington’s disease (HD) patients (n = 46) and controls (n = 35), we compared the performance of four automated segmentation tools (FIRST, FreeSurfer, STEPS, MALP-EM) with manual delineation for generating cross-sectional caudate volume, a region known to be vulnerable in HD. We then examined the effect of each of these baseline regions on the ability to detect change over 15 months using the established longitudinal Caudate Boundary Shift Integral (cBSI) method, an automated longitudinal pipeline requiring a baseline caudate region as an input.Results: All tools, except Freesurfer, generated significantly smaller caudate volumes than the manually derived regions. Jaccard indices showed poorer levels of overlap between each automated segmentation and manual delineation in the HD patients compared with controls. Nevertheless, each method was able to demonstrate significant group differences in volume (p < 0.001). STEPS performed best qualitatively as well as quantitively in the baseline analysis. Caudate atrophy measures generated by the cBSI using automated baseline regions were largely consistent with those derived from a manually segmented baseline, with STEPS providing the most robust cBSI values across both control and HD groups.Conclusions: Atrophy measures from the cBSI were relatively robust to differences in baseline segmentation technique, suggesting that fully automated pipelines could be used to generate outcome measures for clinical trials.

Automated detection and segmentation of thoracic lymph nodes from CT using 3D foveal fully convolutional neural networks

BackgroundIn oncology, the correct determination of nodal metastatic disease is essential for patient management, as patient treatment and prognosis are closely linked to the stage of the disease. The aim of the study was to develop a tool for automatic 3D detection and segmentation of lymph nodes (LNs) in computed tomography (CT) scans of the thorax using a fully convolutional neural network based on 3D foveal patches.MethodsThe training dataset was collected from the Computed Tomography Lymph Nodes Collection of the Cancer Imaging Archive, containing 89 contrast-enhanced CT scans of the thorax. A total number of 4275 LNs was segmented semi-automatically by a radiologist, assessing the entire 3D volume of the LNs. Using this data, a fully convolutional neuronal network based on 3D foveal patches was trained with fourfold cross-validation. Testing was performed on an unseen dataset containing 15 contrast-enhanced CT scans of patients who were referred upon suspicion or for staging of bronchial carcinoma.ResultsThe algorithm achieved a good overall performance with a total detection rate of 76.9% for enlarged LNs during fourfold cross-validation in the training dataset with 10.3 false-positives per volume and of 69.9% in the unseen testing dataset. In the training dataset a better detection rate was observed for enlarged LNs compared to smaller LNs, the detection rate for LNs with a short-axis diameter (SAD) ≥ 20 mm and SAD 5–10 mm being 91.6% and 62.2% (p < 0.001), respectively. Best detection rates were obtained for LNs located in Level 4R (83.6%) and Level 7 (80.4%).ConclusionsThe proposed 3D deep learning approach achieves an overall good performance in the automatic detection and segmentation of thoracic LNs and shows reasonable generalizability, yielding the potential to facilitate detection during routine clinical work and to enable radiomics research without observer-bias.

Combined Atlas and Convolutional Neural Network-Based Segmentation of the Hippocampus from MRI According to the ADNI Harmonized Protocol.

Hippocampus atrophy is an early structural feature that can be measured from magnetic resonance imaging (MRI) to improve the diagnosis of neurological diseases. An accurate and robust standardized hippocampus segmentation method is required for reliable atrophy assessment. The aim of this work was to develop and evaluate an automatic segmentation tool (DeepHarp) for hippocampus delineation according to the ADNI harmonized hippocampal protocol (HarP). DeepHarp utilizes a two-step process. First, the approximate location of the hippocampus is identified in T1-weighted MRI datasets using an atlas-based approach, which is used to crop the images to a region-of-interest (ROI) containing the hippocampus. In the second step, a convolutional neural network trained using datasets with corresponding manual hippocampus annotations is used to segment the hippocampus from the cropped ROI. The proposed method was developed and validated using 107 datasets with manually segmented hippocampi according to the ADNI-HarP standard as well as 114 multi-center datasets of patients with Alzheimer’s disease, mild cognitive impairment, cerebrovascular disease, and healthy controls. Twenty-three independent datasets manually segmented according to the ADNI-HarP protocol were used for testing to assess the accuracy, while an independent test-retest dataset was used to assess precision. The proposed DeepHarp method achieved a mean Dice similarity score of 0.88, which was significantly better than four other established hippocampus segmentation methods used for comparison. At the same time, the proposed method also achieved a high test-retest precision (mean Dice score: 0.95). In conclusion, DeepHarp can automatically segment the hippocampus from T1-weighted MRI datasets according to the ADNI-HarP protocol with high accuracy and robustness, which can aid atrophy measurements in a variety of pathologies.

YeastNet: Deep-Learning-Enabled Accurate Segmentation of Budding Yeast Cells in Bright-Field Microscopy

Accurate and efficient segmentation of live-cell images is critical in maximizing data extraction and knowledge generation from high-throughput biology experiments. Despite recent development of deep-learning tools for biomedical imaging applications, great demand for automated segmentation tools for high-resolution live-cell microscopy images remains in order to accelerate the analysis. YeastNet dramatically improves the performance of the non-trainable classic algorithm, and performs considerably better than the current state-of-the-art yeast-cell segmentation tools. We have designed and trained a U-Net convolutional network (named YeastNet) to conduct semantic segmentation on bright-field microscopy images and generate segmentation masks for cell labeling and tracking. YeastNet enables accurate automatic segmentation and tracking of yeast cells in biomedical applications. YeastNet is freely provided with model weights as a Python package on GitHub.

MPRAGE to MP2RAGE UNI translation via generative adversarial network improves the automatic tissue and lesion segmentation in multiple sclerosis patients

Background and objectiveCompared to the conventional magnetization-prepared rapid gradient-echo imaging (MPRAGE) MRI sequence, the specialized magnetization prepared 2 rapid acquisition gradient echoes (MP2RAGE) shows a higher brain tissue and lesion contrast in multiple sclerosis (MS) patients. The goal of this work is to retrospectively generate realistic-looking MP2RAGE uniform images (UNI) from already acquired MPRAGE images in order to improve the automatic lesion and tissue segmentation. MethodsFor this task we propose a generative adversarial network (GAN). Multi-contrast MRI data of 12 healthy controls and 44 patients diagnosed with MS was retrospectively analyzed. Imaging was acquired at 3T using a SIEMENS scanner with MPRAGE, MP2RAGE, FLAIR, and DIR sequences. We train the GAN with both healthy controls and MS patients to generate synthetic MP2RAGE UNI images. These images were then compared to the real MP2RAGE UNI (considered as ground truth) analyzing the output of automatic brain tissue and lesion segmentation tools. Reference-based metrics as well as the lesion-wise true and false positives, Dice coefficient, and volume difference were considered for the evaluation. Statistical differences were assessed with the Wilcoxon signed-rank test. ResultsThe synthetic MP2RAGE UNI significantly improves the lesion and tissue segmentation masks in terms of Dice coefficient and volume difference (p-values < 0.001) compared to the MPRAGE. For the segmentation metrics analyzed no statistically significant differences are found between the synthetic and acquired MP2RAGE UNI. ConclusionSynthesized MP2RAGE UNI images are visually realistic and improve the output of automatic segmentation tools.

Development of Random Forest Algorithm Based Prediction Model of Alzheimer’s Disease Using Neurodegeneration Pattern

Objective Alzheimer’s disease (AD) is the most common type of dementia and the prevalence rapidly increased as the elderly population increased worldwide. In the contemporary model of AD, it is regarded as a disease continuum involving preclinical stage to severe dementia. For accurate diagnosis and disease monitoring, objective index reflecting structural change of brain is needed to correctly assess a patient’s severity of neurodegeneration independent from the patient’s clinical symptoms. The main aim of this paper is to develop a random forest (RF) algorithm-based prediction model of AD using structural magnetic resonance imaging (MRI). Methods We evaluated diagnostic accuracy and performance of our RF based prediction model using newly developed brain segmentation method compared with the Freesurfer’s which is a commonly used segmentation software. Results Our RF model showed high diagnostic accuracy for differentiating healthy controls from AD and mild cognitive impairment (MCI) using structural MRI, patient characteristics, and cognitive function (HC vs. AD 93.5%, AUC 0.99; HC vs. MCI 80.8%, AUC 0.88). Moreover, segmentation processing time of our algorithm (<5 minutes) was much shorter than of Freesurfer’s (6–8 hours). Conclusion Our RF model might be an effective automatic brain segmentation tool which can be easily applied in real clinical practice.

Feasibility of using a novel automatic cardiac segmentation algorithm in the clinical routine of lung cancer patients.

Incidental radiation exposure to the heart during lung cancer radiotherapy is associated with radiation-induced heart disease and increased rates of mortality. By considering the respiratory-induced motion of the heart it is possible to create a radiotherapy plan that results in a lower overall cardiac dose. This approach is challenging using current clinical practices: manual contouring of the heart is time consuming, and subject to inter- and intra-observer variability. In this work, we investigate the feasibility of our previously developed, atlas-based, automatic heart segmentation tool to delineate the heart in four-dimensional x-ray computed tomography (4D-CT) images. We used a dataset comprising 19 patients receiving radiotherapy for lung cancer, with 4D-CT imaging acquired at 10 respiratory phases and with a maximum intensity projection image generated from these. For each patient, one of four experienced radiation oncologists contoured the heart on each respiratory phase image and the maximum intensity image. Automatic segmentation of the heart on these same patient image sets was achieved using a leave-one-out approach, where for each patient the remaining 18 were used as an atlas set. The consistency of the automatic segmentation relative to manual contouring was evaluated using the Dice similarity coefficient (DSC) and mean absolute surface-to-surface distance (MASD). The DSC and MASD are comparable to inter-observer variability in clinically acceptable whole heart delineations (average DSC > 0.93 and average MASD < 2.0 mm in all the respiratory phases). The comparison between automatic and manual delineations on the maximum intensity images produced an overall mean DSC of 0.929 and a mean MASD of 2.07 mm. The automatic, atlas-based segmentation tool produces clinically consistent and robust heart delineations and is easy to implement in the routine care of lung cancer patients.

A Review of Publicly Available Automatic Brain Segmentation Methodologies, Machine Learning Models, Recent Advancements, and Their Comparison.

Background:The noninvasive study of the structure and functions of the brain using neuroimaging techniques is increasingly being used for its clinical and research perspective. The morphological and volumetric changes in several regions and structures of brains are associated with the prognosis of neurological disorders such as Alzheimer’s disease, epilepsy, schizophrenia, etc. and the early identification of such changes can have huge clinical significance. The accurate segmentation of three-dimensional brain magnetic resonance images into tissue types (i.e., grey matter, white matter, cerebrospinal fluid) and brain structures, thus, has huge importance as they can act as early biomarkers. The manual segmentation though considered the “gold standard” is time-consuming, subjective, and not suitable for bigger neuroimaging studies. Several automatic segmentation tools and algorithms have been developed over the years; the machine learning models particularly those using deep convolutional neural network (CNN) architecture are increasingly being applied to improve the accuracy of automatic methods.Purpose:The purpose of the study is to understand the current and emerging state of automatic segmentation tools, their comparison, machine learning models, their reliability, and shortcomings with an intent to focus on the development of improved methods and algorithms.Methods:The study focuses on the review of publicly available neuroimaging tools, their comparison, and emerging machine learning models particularly those based on CNN architecture developed and published during the last five years.Conclusion:Several software tools developed by various research groups and made publicly available for automatic segmentation of the brain show variability in their results in several comparison studies and have not attained the level of reliability required for clinical studies. The machine learning models particularly three dimensional fully convolutional network models can provide a robust and efficient alternative with relation to publicly available tools but perform poorly on unseen datasets. The challenges related to training, computation cost, reproducibility, and validation across distinct scanning modalities for machine learning models need to be addressed.

BSI whole‐brain volume change estimation initialised with fully automatic LEAP segmentation

AbstractBackgroundWhole brain atrophy as measured from T1‐wieghted MR images is a key biomarker for progression in Alzheimer's disease (AD) and employed as marker for treatment efficacy in clinical trials. The Boundary Shift Integral (BSI) methodology (Freborough;IEEE;1997) is the widely established standard and well validated method for volume change. BSI requires initialisation with a whole‐brain segmentation at both timepoints, manual segmentations remain the gold standard, however the BSI approach is robust to minor segmentation, potentially facilitating the use of automatic segmentation approaches. Furthermore, to improve sensitivity to volume change and exploit the probabilistic nature of many automatic segmentations a generalised form of the BSI (gBSI) pipeline has been proposed (Prados;NeurobiolAging;2015). The use of a fully automatic pipeline to estimate volume change provides a scalable solution for clinical trial deployment without the burden of manual segmentation or edits and associated variability introduced. Here we assess the use of LEAP (Ledig;ProcISBI;2012), a fully automatic whole brain segmentation tool providing both binary and probabilistic parcellations, to initialise the BSI, as implemented in MIDAS, and gBSI pipelines.MethodWhole brain volume change over 1 year was estimated for 48 participants, acquired as part of a clinical trial, with the BSI and gBSI methods initialised with 1) manually generated whole brain segmentations and 2) automatically estimated whole brain segmentations generate with LEAP. No manual corrections were performed; all LEAP segmentations passed a visual quality control inspection.ResultFigure 1 shows the relationship between the manually initialised BSI and the fully‐automated LEAP‐BSI approach. Table 1 shows average measures of volume change across all tested methods. All measures were significantly correlated with non‐significant differences in reported mean volume change.ConclusionLEAP offers segmentations suitable for initialisation of the BSI method providing a pipeline for fully automatic volume estimation consistent with the semi‐automated BSI approach, providing a scalable solution for clinical trial deployment without the associated variability and resource burden required for fully manual segmentations. Furthermore the probabilistic output of LEAP allows for the use of the gBSI methodology with the associated potential of increased sensitivity to volume change.

Active contour regularized semi-supervised learning for COVID-19 CT infection segmentation with limited annotations

Infection segmentation on chest CT plays an important role in the quantitative analysis of COVID-19. Developing automatic segmentation tools in a short period with limited labelled images has become an urgent need. Pseudo label-based semi-supervised method is a promising way to leverage unlabelled data to improve segmentation performance. Existing methods usually obtain pseudo labels by first training a network with limited labelled images and then inferring unlabelled images. However, these methods may generate obviously inaccurate labels and degrade the subsequent training process. To address these challenges, in this paper, an active contour regularized semi-supervised learning framework was proposed to automatically segment infections with few labelled images. The active contour regularization was realized by the region-scalable fitting (RSF) model which is embedded to the loss function of the network to regularize and refine the pseudo labels of the unlabelled images. We further designed a splitting method to separately optimize the RSF regularization term and the segmentation loss term with iterative convolution-thresholding method and stochastic gradient descent, respectively, which enable fast optimization of each term. Furthermore, we built a statistical atlas to show the infection spatial distribution. Extensive experiments on a small public dataset and a large scale dataset showed that the proposed method outperforms state-of-the-art methods with up to 5% in dice similarity coefficient and normalized surface dice, 10% in relative absolute volume difference and 8 mm in 95% Hausdorff distance. Moreover, we observed that the infections tend to occur at the dorsal subpleural lung and posterior basal segments that are not mentioned in current radiology reports and are meaningful to advance our understanding of COVID-19.

LGAN: Lung segmentation in CT scans using generative adversarial network

Lung segmentation in Computerized Tomography (CT) images plays an important role in various lung disease diagnosis. Most of the current lung segmentation approaches are performed through a series of procedures with manually empirical parameter adjustments in each step. Pursuing an automatic segmentation method with fewer steps, we propose a novel deep learning Generative Adversarial Network (GAN)-based lung segmentation schema, which we denote as LGAN. The proposed schema can be generalized to different kinds of neural networks for lung segmentation in CT images. We evaluated the proposed LGAN schema on datasets including Lung Image Database Consortium image collection (LIDC-IDRI) and Quantitative Imaging Network (QIN) collection with two metrics: segmentation quality and shape similarity. Also, we compared our work with current state-of-the-art methods. The experimental results demonstrated that the proposed LGAN schema can be used as a promising tool for automatic lung segmentation due to its simplified procedure as well as its improved performance and efficiency.