Performance In Automatic Segmentation Research Articles

Spatial resolution of renal amyloid deposits through MALDI-MSI: a combined digital and molecular approach to monoclonal gammopathies

AimsIdentification and characterisation of monoclonal gammopathies of renal significance (MGRS) is critical for therapeutic purposes. Amyloidosis represents one of the most common forms of MGRS, and renal biopsy remains the...

Deep-Learning-Based Automatic Segmentation of Parotid Gland on Computed Tomography Images.

This study aims to develop an algorithm for the automatic segmentation of the parotid gland on CT images of the head and neck using U-Net architecture and to evaluate the model's performance. In this retrospective study, a total of 30 anonymized CT volumes of the head and neck were sliced into 931 axial images of the parotid glands. Ground truth labeling was performed with the CranioCatch Annotation Tool (CranioCatch, Eskisehir, Turkey) by two oral and maxillofacial radiologists. The images were resized to 512 × 512 and split into training (80%), validation (10%), and testing (10%) subgroups. A deep convolutional neural network model was developed using U-net architecture. The automatic segmentation performance was evaluated in terms of the F1-score, precision, sensitivity, and the Area Under Curve (AUC) statistics. The threshold for a successful segmentation was determined by the intersection of over 50% of the pixels with the ground truth. The F1-score, precision, and sensitivity of the AI model in segmenting the parotid glands in the axial CT slices were found to be 1. The AUC value was 0.96. This study has shown that it is possible to use AI models based on deep learning to automatically segment the parotid gland on axial CT images.

Brain segmentation in patients with perinatal arterial ischemic stroke

BackgroundPerinatal arterial ischemic stroke (PAIS) is associated with adverse neurological outcomes. Quantification of ischemic lesions and consequent brain development in newborn infants relies on labor-intensive manual assessment of brain tissues and ischemic lesions. Hence, we propose an automatic method utilizing convolutional neural networks (CNNs) to segment brain tissues and ischemic lesions in MRI scans of infants suffering from PAIS. Materials and MethodsThis single-center retrospective study included 115 patients with PAIS that underwent MRI after the stroke onset (baseline) and after three months (follow-up). Nine baseline and 12 follow-up MRI scans were manually annotated to provide reference segmentations (white matter, gray matter, basal ganglia and thalami, brainstem, ventricles, extra-ventricular cerebrospinal fluid, and cerebellum, and additionally on the baseline scans the ischemic lesions). Two CNNs were trained to perform automatic segmentation on the baseline and follow-up MRIs, respectively. Automatic segmentations were quantitatively evaluated using the Dice coefficient (DC) and the mean surface distance (MSD). Volumetric agreement between segmentations that were manually and automatically obtained was computed. Moreover, the scan quality and automatic segmentations were qualitatively evaluated in a larger set of MRIs without manual annotation by two experts. In addition, the scan quality was qualitatively evaluated in these scans to establish its impact on the automatic segmentation performance. ResultsAutomatic brain tissue segmentation led to a DC and MSD between 0.78–0.92 and 0.18–1.08 mm for baseline, and between 0.88–0.95 and 0.10–0.58 mm for follow-up scans, respectively. For the ischemic lesions at baseline the DC and MSD were between 0.72–0.86 and 1.23–2.18 mm, respectively. Volumetric measurements indicated limited oversegmentation of the extra-ventricular cerebrospinal fluid in both the follow-up and baseline scans, oversegmentation of the ischemic lesions in the left hemisphere, and undersegmentation of the ischemic lesions in the right hemisphere. In scans without imaging artifacts, brain tissue segmentation was graded as excellent in more than 85% and 91% of cases, respectively for the baseline and follow-up scans. For the ischemic lesions at baseline, this was in 61% of cases. ConclusionsAutomatic segmentation of brain tissue and ischemic lesions in MRI scans of patients with PAIS is feasible. The method may allow evaluation of the brain development and efficacy of treatment in large datasets.

Automated pancreas segmentation and volumetry using deep neural network on computed tomography

Pancreas segmentation is necessary for observing lesions, analyzing anatomical structures, and predicting patient prognosis. Therefore, various studies have designed segmentation models based on convolutional neural networks for pancreas segmentation. However, the deep learning approach is limited by a lack of data, and studies conducted on a large computed tomography dataset are scarce. Therefore, this study aims to perform deep-learning-based semantic segmentation on 1006 participants and evaluate the automatic segmentation performance of the pancreas via four individual three-dimensional segmentation networks. In this study, we performed internal validation with 1,006 patients and external validation using the cancer imaging archive pancreas dataset. We obtained mean precision, recall, and dice similarity coefficients of 0.869, 0.842, and 0.842, respectively, for internal validation via a relevant approach among the four deep learning networks. Using the external dataset, the deep learning network achieved mean precision, recall, and dice similarity coefficients of 0.779, 0.749, and 0.735, respectively. We expect that generalized deep-learning-based systems can assist clinical decisions by providing accurate pancreatic segmentation and quantitative information of the pancreas for abdominal computed tomography.

MA-SOCRATIS: An automatic pipeline for robust segmentation of the left ventricle and scar

Multi-atlas segmentation of cardiac regions and total infarct scar (MA-SOCRATIS) is an unsupervised automatic pipeline to segment left ventricular myocardium and scar from late gadolinium enhanced MR images (LGE-MRI) of the heart.We implement two different pipelines for myocardial and scar segmentation from short axis LGE-MRI. Myocardial segmentation has two steps; initial segmentation and re-estimation. The initial segmentation step makes a first estimate of myocardium boundaries by using multi-atlas segmentation techniques. The re-estimation step refines the myocardial segmentation by a combination of k-means clustering and a geometric median shape variation technique. An active contour technique determines the unhealthy and healthy myocardial wall. The scar segmentation pipeline is a combination of a Rician–Gaussian mixture model and full width at half maximum (FWHM) thresholding, to determine the intensity pixels in scar regions. Following this step a watershed method with an automatic seed-points framework segments the final scar region.MA-SOCRATIS was evaluated using two different datasets. In both datasets ground truths were based on manual segmentation of short axis images from LGE-MRI scans. The first dataset included 40 patients from the MS-CMRSeg 2019 challenge dataset (STACOM at MICCAI 2019). The second is a collection of 20 patients with scar regions that are challenging to segment. MA-SOCRATIS achieved robust and accurate performance in automatic segmentation of myocardium and scar regions without the need of training or tuning in both cohorts, compared with state-of-the-art techniques (intra-observer and inter observer myocardium segmentation: 81.9% and 70% average Dice value, and scar (intra-observer and inter observer segmentation: 70.5% and 70.5% average Dice value).

MIDeepSeg: Minimally interactive segmentation of unseen objects from medical images using deep learning.

Segmentation of organs or lesions from medical images plays an essential role in many clinical applications such as diagnosis and treatment planning. Though Convolutional Neural Networks (CNN) have achieved the state-of-the-art performance for automatic segmentation, they are often limited by the lack of clinically acceptable accuracy and robustness in complex cases. Therefore, interactive segmentation is a practical alternative to these methods. However, traditional interactive segmentation methods require a large number of user interactions, and recently proposed CNN-based interactive segmentation methods are limited by poor performance on previously unseen objects. To solve these problems, we propose a novel deep learning-based interactive segmentation method that not only has high efficiency due to only requiring clicks as user inputs but also generalizes well to a range of previously unseen objects. Specifically, we first encode user-provided interior margin points via our proposed exponentialized geodesic distance that enables a CNN to achieve a good initial segmentation result of both previously seen and unseen objects, then we use a novel information fusion method that combines the initial segmentation with only a few additional user clicks to efficiently obtain a refined segmentation. We validated our proposed framework through extensive experiments on 2D and 3D medical image segmentation tasks with a wide range of previously unseen objects that were not present in the training set. Experimental results showed that our proposed framework 1) achieves accurate results with fewer user interactions and less time compared with state-of-the-art interactive frameworks and 2) generalizes well to previously unseen objects.

Deep learning-based cardiac cine segmentation: Transfer learning application to 7T ultrahigh-field MRI.

Artificial neural networks show promising performance in automatic segmentation of cardiac MRI. However, training requires large amounts of annotated data and generalization to different vendors, field strengths, sequence parameters, and pathologies is limited. Transfer learning addresses this challenge, but specific recommendations regarding type and amount of data required is lacking. In this study, we assess data requirements for transfer learning to experimental cardiac MRI at 7T where the segmentation task can be challenging. In addition, we provide guidelines, tools, and annotated data to enable transfer learning approaches by other researchers and clinicians. A publicly available segmentation model was used to annotate a publicly available data set. This labeled data set was subsequently used to train a neural network for segmentation of left ventricle and myocardium in cardiac cine MRI. The network is used as starting point for transfer learning to 7T cine data of healthy volunteers (n = 22; 7873 images) by updating the pre-trained weights. Structured and random data subsets of different sizes were used to systematically assess data requirements for successful transfer learning. Inconsistencies in the publically available data set were corrected, labels created, and a neural network trained. On 7T cardiac cine images the model pre-trained on public imaging data, acquired at 1.5T and 3T, achieved DICELV = 0.835 and DICEMY = 0.670. Transfer learning using 7T cine data and ImageNet weight initialization improved model performance to DICELV = 0.900 and DICEMY = 0.791. Using only end-systolic and end-diastolic images reduced training data by 90%, with no negative impact on segmentation performance (DICELV = 0.908, DICEMY = 0.805). This work demonstrates and quantifies the benefits of transfer learning for cardiac cine image segmentation. We provide practical guidelines for researchers planning transfer learning projects in cardiac MRI and make data, models, and code publicly available.

Automatic segmentation of three clinical target volumes in radiotherapy using lifelong learning

Background and purposeConvolutional neural networks (CNNs) have comparable human level performance in automatic segmentation. An important challenge that CNNs face in segmentation is catastrophic forgetting. They lose performance on tasks that were previously learned when trained on task. In this study, we propose a lifelong learning method to learn multiple segmentation tasks continuously without forgetting previous tasks. Materials and methodsThe cohort included three tumors, 800 patients of which had nasopharyngeal cancer (NPC), 800 patients had breast cancer, and 800 patients had rectal cancer. The tasks included segmentation of the clinical target volume (CTV) of these three cancers. The proposed lifelong learning network adopted dilation adapter to learn three segmentation tasks one by one. Only the newly added dilation adapter (seven layers) was fine tuning for incoming new task, whereas all the other learned layers were frozen. ResultsCompared with single-task, multi-task or transfer learning, the proposed lifelong learning can achieve better or comparable segmentation accuracy with a DSC of 0.86 for NPC, 0.89 for breast cancer, and 0.87 for rectal cancer. Lifelong learning can avoid forgetting in sequential learning and yield good performance with less training data. Furthermore, it is more efficient than single-task or transfer learning, which reduced the number of parameters, size of model, and training time by ~58.8%, ~55.6%, and ~25.0%, respectively. ConclusionThe proposed method preserved the knowledge of previous tasks while learning a new one using a dilation adapter. It could yield comparable performance with much less training data, model parameters, and training time.

Development of a self-constrained 3D DenseNet model in automatic detection and segmentation of nasopharyngeal carcinoma using magnetic resonance images.

We aimed to develop a dual-task model to detect and segment nasopharyngeal carcinoma (NPC) automatically in magnetic resource images (MRI) based on deep learning method, since the differential diagnosis of NPC and atypical benign hyperplasia was difficult and the radiotherapy target contouring of NPC was labor-intensive. A self-constrained 3D DenseNet (SC-DenseNet) architecture was improved using separated training and validation sets. A total of 4100 individuals were finally enrolled and split into the training, validation and test sets at a proximate ratio of 8:1:1 using simple randomization. The diagnostic metrics of the established model against experienced radiologists was compared in the test set. The dice similarity coefficient (DSC) of manual and model-defined tumor region was used to evaluate the efficacy of segmentation. Totally, 3142 nasopharyngeal carcinoma (NPC) and 958 benign hyperplasia were included. The SC-DenseNet model showed encouraging performance in detecting NPC, attained a higher overall accuracy, sensitivity and specificity than those of the experienced radiologists (97.77% vs 95.87%, 99.68% vs 99.24% and 91.67% vs 85.21%, respectively). Moreover, the model also exhibited promising performance in automatic segmentation of tumor region in NPC, with an average DSC at 0.77±0.07 in the test set. The SC-DenseNet model showed competence in automatic detection and segmentation of NPC in MRI, indicating the promising application value as an assistant tool in clinical practice, especially in screening project.

Development and validation of an endoscopic images-based deep learning model for detection with nasopharyngeal malignancies

BackgroundDue to the occult anatomic location of the nasopharynx and frequent presence of adenoid hyperplasia, the positive rate for malignancy identification during biopsy is low, thus leading to delayed or missed diagnosis for nasopharyngeal malignancies upon initial attempt. Here, we aimed to develop an artificial intelligence tool to detect nasopharyngeal malignancies under endoscopic examination based on deep learning.MethodsAn endoscopic images-based nasopharyngeal malignancy detection model (eNPM-DM) consisting of a fully convolutional network based on the inception architecture was developed and fine-tuned using separate training and validation sets for both classification and segmentation. Briefly, a total of 28,966 qualified images were collected. Among these images, 27,536 biopsy-proven images from 7951 individuals obtained from January 1st, 2008, to December 31st, 2016, were split into the training, validation and test sets at a ratio of 7:1:2 using simple randomization. Additionally, 1430 images obtained from January 1st, 2017, to March 31st, 2017, were used as a prospective test set to compare the performance of the established model against oncologist evaluation. The dice similarity coefficient (DSC) was used to evaluate the efficiency of eNPM-DM in automatic segmentation of malignant area from the background of nasopharyngeal endoscopic images, by comparing automatic segmentation with manual segmentation performed by the experts.ResultsAll images were histopathologically confirmed, and included 5713 (19.7%) normal control, 19,107 (66.0%) nasopharyngeal carcinoma (NPC), 335 (1.2%) NPC and 3811 (13.2%) benign diseases. The eNPM-DM attained an overall accuracy of 88.7% (95% confidence interval (CI) 87.8%–89.5%) in detecting malignancies in the test set. In the prospective comparison phase, eNPM-DM outperformed the experts: the overall accuracy was 88.0% (95% CI 86.1%–89.6%) vs. 80.5% (95% CI 77.0%–84.0%). The eNPM-DM required less time (40 s vs. 110.0 ± 5.8 min) and exhibited encouraging performance in automatic segmentation of nasopharyngeal malignant area from the background, with an average DSC of 0.78 ± 0.24 and 0.75 ± 0.26 in the test and prospective test sets, respectively.ConclusionsThe eNPM-DM outperformed oncologist evaluation in diagnostic classification of nasopharyngeal mass into benign versus malignant, and realized automatic segmentation of malignant area from the background of nasopharyngeal endoscopic images.

Discriminative confidence estimation for probabilistic multi-atlas label fusion.

Quantitative neuroimaging analyses often rely on the accurate segmentation of anatomical brain structures. In contrast to manual segmentation, automatic methods offer reproducible outputs and provide scalability to study large databases. Among existing approaches, multi-atlas segmentation has recently shown to yield state-of-the-art performance in automatic segmentation of brain images. It consists in propagating the labelmaps from a set of atlases to the anatomy of a target image using image registration, and then fusing these multiple warped labelmaps into a consensus segmentation on the target image. Accurately estimating the contribution of each atlas labelmap to the final segmentation is a critical step for the success of multi-atlas segmentation. Common approaches to label fusion either rely on local patch similarity, probabilistic statistical frameworks or a combination of both. In this work, we propose a probabilistic label fusion framework based on atlas label confidences computed at each voxel of the structure of interest. Maximum likelihood atlas confidences are estimated using a supervised approach, explicitly modeling the relationship between local image appearances and segmentation errors produced by each of the atlases. We evaluate different spatial pooling strategies for modeling local segmentation errors. We also present a novel type of label-dependent appearance features based on atlas labelmaps that are used during confidence estimation to increase the accuracy of our label fusion. Our approach is evaluated on the segmentation of seven subcortical brain structures from the MICCAI 2013 SATA Challenge dataset and the hippocampi from the ADNI dataset. Overall, our results indicate that the proposed label fusion framework achieves superior performance to state-of-the-art approaches in the majority of the evaluated brain structures and shows more robustness to registration errors.

Automatic Segmentation and Quantitative Analysis of White Matter Hyperintensities on FLAIR Images Using Trimmed-Likelihood Estimator

Quantitative analysis of white matter hyperintensities (WMHs) on fluid-attenuated inversion recovery (FLAIR) images provides information for disease tracking, therapeutic efficacy assessment, and cognitive science research. This study developed an automatic segmentation method to detect and quantify WMHs on FLAIR images. This study aims to assess the accuracy and reproducibility of the proposed method. The FLAIR images of 82 patients (58-84years) with different WMH burdens were acquired with the same 3T scanner (Intera-achieva SMI-2.1; Philip Medical System, Sixth Affiliated People's Hospital, Shanghai, China). The FLAIR images were preprocessed through brain extraction and intensity inhomogeneity correction. An anatomy atlas built from a set of 20 patients with different WMH burdens (mild, 11 patients; moderate, 6 patients; and severe, 3 patients) was used to estimate a control parameter in the framework of segmentation. The general flow for WMH segmentation included classification of foreground and background regions, detection of abnormally high signals, and WMH refinement. The performance of automatic segmentation was evaluated by a volumetric comparison with manual segmentation on patients with different WMH burdens. Similarity index values for the accuracy of automatic segmentation compared to manual segmentation were consistently high for patients with different WMH burdens (mild, 0.78±0.08; moderate, 0.83±0.06; severe, 0.84±0.08; and total, 0.80±0.08). Linear regression demonstrated a strong correlation between the WMH volumes measured by the two methods in all patients (r=0.98, P=.006). Small average differences were detected between the WMH volumes obtained through manual and automatic segmentations (mild, 4.76%; moderate, 6.84%; and severe, 7.59%). The results of Bland-Altman analysis show a system bias of 0.68cm(3) and a standard deviation of 2.10cm(3) over the range of 2.58-53.9cm(3). The developed method is accurate and efficient in detecting and quantifying differently sized WMHs on FLAIR images. Automatic segmentation is a promising computer-aided diagnostic tool to study WMHs in aging and dementia in basic research and even in clinical trials.

Anatomical atlas in the context of head and neck radiotherapy and its use to automatic segmentation

In this paper we present a methodology to form an anatomical atlas based on the analysis of dense deformation fields recovered by the Morphons non-rigid registration algorithm. The methodology is based on measuring the bending energy required to register the whole database to a reference, and the atlas is the one image in the database which yields the smallest bending energy when taken as reference. The suitability of our atlas is demonstrated in the context of head and neck radiotherapy through its application to a database with thirty-one computed tomography (CT) images of the head and neck region. In head and neck radiotherapy, CT is the most frequently used modality for the segmentation of organs at risk and clinical target volumes. One challenge brought by the use of CT images is the presence of important artifacts caused by dental implants. The presence of such artifacts hinders the use of intensity averages, thus severely limiting the application of most atlas building techniques described in the literature in this context. The results presented in the paper show that our bending energy model faithfully represents the shape variability of patients in the head and neck region; they also show its good performance in segmentation of volumes of interest in radiotherapy. Moreover, when compared to other atlases of similar performance in automatic segmentation, our atlas presents the desirable feature of not being blurred after intensity averaging.

Automatic hippocampal segmentation in temporal lobe epilepsy: Impact of developmental abnormalities

In drug-resistant temporal lobe epilepsy (TLE), detecting hippocampal atrophy on MRI is important as it allows defining the surgical target. The performance of automatic segmentation in TLE has so far been considered unsatisfactory. In addition to atrophy, about 40% of patients present with developmental abnormalities (referred to as malrotation) characterized by atypical morphologies of the hippocampus and collateral sulcus. Our purpose was to evaluate the impact of malrotation and atrophy on the performance of three state-of-the-art automated algorithms. We segmented the hippocampus in 66 patients and 35 sex- and age-matched healthy subjects using a region-growing algorithm constrained by anatomical priors (SACHA), a freely available atlas-based software (FreeSurfer) and a multi-atlas approach (ANIMAL-multi). To quantify malrotation, we generated 3D models from manual hippocampal labels and automatically extracted collateral sulci. The accuracy of automated techniques was evaluated relative to manual labeling using the Dice similarity index and surface-based shape mapping, for which we computed vertex-wise displacement vectors between automated and manual segmentations. We then correlated segmentation accuracy with malrotation features and atrophy. ANIMAL-multi demonstrated similar accuracy in patients and healthy controls (p > 0.1), whereas SACHA and FreeSurfer were less accurate in patients (p < 0.05). Surface-based analysis of contour accuracy revealed that SACHA over-estimated the lateral border of malrotated hippocampi (r = 0.61; p < 0.0001), but performed well in the presence of atrophy (|r |< 0.34; p > 0.2). Conversely, FreeSurfer and ANIMAL-multi were affected by both malrotation (FreeSurfer: r = 0.57; p = 0.02, ANIMAL-multi: r = 0.50; p = 0.05) and atrophy (FreeSurfer: r = 0.78, p < 0.0001, ANIMAL-multi: r = 0.61; p < 0.0001). Compared to manual volumetry, automated procedures underestimated the magnitude of atrophy (Cohen's d: manual: 1.68; ANIMAL-multi: 1.11; SACHA: 1.10; FreeSurfer: 0.90, p < 0.0001). In addition, they tended to lateralize the seizure focus less accurately in the presence of malrotation (manual: 64%; ANIMAL-multi: 55%, p = 0.4; SACHA: 50%, p = 0.1; FreeSurfer: 41%, p = 0.05). Hippocampal developmental anomalies and atrophy had a negative impact on the segmentation performance of three state-of-the-art automated methods. These shape variants should be taken into account when designing segmentation algorithms.