Multiatlas-based segmentation with preregistration atlas selection

Deep convolutional neural network for hippocampus segmentation with boundary region refinement

Accurately segmenting the hippocampus from magnetic resonance (MR) brain images is a crucial step in studying brain disorders. However, this task is challenging due to the low signal contrast of hippocampal images, the irregular shape, and small structural size of the hippocampi. In recent years, several deep convolutional networks have been proposed for hippocampus segmentation, which have achieved state-of-the-art performance. These methods typically use large image patches for training the network, as larger patches are beneficial for capturing long-range contextual information. However, this approach increases the computational burden and overlooks the significance of the boundary region. In this study, we propose a deep learning-based method for hippocampus segmentation with boundary region refinement. Our method involves two main steps. First, we propose a convolutional network that takes large image patches as input for initial segmentation. Then, we extract small image patches around the hippocampal boundary for training the second convolutional neural network, which refines the segmentation in the boundary regions. We validate our proposed method on a publicly available dataset and demonstrate that it significantly improves the performance of convolutional neural networks that use single-size image patches as input. In conclusion, our study proposes a novel method for hippocampus segmentation, which improves upon the current state-of-the-art methods. By incorporating a boundary refinement step, our approach achieves higher accuracy in hippocampus segmentation and may facilitate research on brain disorders.

Auto-segmentation for total marrow irradiation

Purpose

To evaluate the accuracy and efficiency of Artificial-Intelligence (AI) segmentation in Total Marrow Irradiation (TMI) including contours throughout the head and neck (H&N), thorax, abdomen, and pelvis.

Methods

An AI segmentation software was clinically introduced for total body contouring in TMI including 27 organs at risk (OARs) and 4 planning target volumes (PTVs). This work compares the clinically utilized contours to the AI-TMI contours for 21 patients. Structure and image dicom data was used to generate comparisons including volumetric, spatial, and dosimetric variations between the AI- and human-edited contour sets. Conventional volume and surface measures including the Sørensen-Dice coefficient (Dice) and the 95th% Hausdorff Distance (HD95) were used, and novel efficiency metrics were introduced. The clinical efficiency gains were estimated by the percentage of the AI-contour-surface within 1mm of the clinical contour surface. An unedited AI-contour has an efficiency gain=100%, an AI-contour with 70% of its surface<1mm from a clinical contour has an efficiency gain of 70%. The dosimetric deviations were estimated from the clinical dose distribution to compute the dose volume histogram (DVH) for all structures.

Results

A total of 467 contours were compared in the 21 patients. In PTVs, contour surfaces deviated by >1mm in 38.6% ± 23.1% of structures, an average efficiency gain of 61.4%. Deviations >5mm were detected in 12.0% ± 21.3% of the PTV contours. In OARs, deviations >1mm were detected in 24.4% ± 27.1% of the structure surfaces and >5mm in 7.2% ± 18.0%; an average clinical efficiency gain of 75.6%. In H&N OARs, efficiency gains ranged from 42% in optic chiasm to 100% in eyes (unedited in all cases). In thorax, average efficiency gains were >80% in spinal cord, heart, and both lungs. Efficiency gains ranged from 60-70% in spleen, stomach, rectum, and bowel and 75-84% in liver, kidney, and bladder. DVH differences exceeded 0.05 in 109/467 curves at any dose level. The most common 5%-DVH variations were in esophagus (86%), rectum (48%), and PTVs (22%).

Conclusions

AI auto-segmentation software offers a powerful solution for enhanced efficiency in TMI treatment planning. Whole body segmentation including PTVs and normal organs was successful based on spatial and dosimetric comparison.

Automated major psoas muscle volumetry in computed tomography using machine learning algorithms

Purpose

The psoas major muscle (PMM) volume serves as an opportunistic imaging marker in cross-sectional imaging datasets for various clinical applications. Since manual segmentation is time consuming, two different automated segmentation methods, a generative adversarial network architecture (GAN) and a multi-atlas segmentation (MAS), as well as a combined approach of both, were investigated in terms of accuracy of automated volumetrics in given CT datasets.

Materials and methods

The bilateral PMM was manually segmented by a radiologist in 34 abdominal CT scans, resulting in 68 single 3D muscle segmentations as training data. Three different methods were tested for their ability to generate automated image segmentations: a GAN- and MAS-based approach and a combined approach of both methods (COM). Bilateral PMM volume (PMMV) was calculated in cm³ by each algorithm for every CT. Results were compared to the corresponding ground truth using the Dice similarity coefficient (DSC), Spearman’s correlation coefficient and Wilcoxon signed-rank test.

Results

Mean PMMV was 239 ± 7.0 cm³ and 308 ± 9.6 cm³, 306 ± 9.5 cm³ and 243 ± 7.3 cm³ for the CNN, MAS and COM, respectively. Compared to the ground truth the CNN and MAS overestimated the PMMV significantly (+ 28.9% and + 28.0%, p < 0.001), while results of the COM were quite accurate (+ 0.7%, p = 0.33). Spearman’s correlation coefficients were 0.38, 0.62 and 0.73, and the DSCs were 0.75 [95%CI: 0.56–0.88], 0.73 [95%CI: 0.54–0.85] and 0.82 [95%CI: 0.65–0.90] for the CNN, MAS and COM, respectively.

Conclusion

The combined approach was able to efficiently exploit the advantages of both methods (GAN and MAS), resulting in a significantly higher accuracy in PMMV predictions compared to the isolated implementations of both methods. Even with the relatively small set of training data, the segmentation accuracy of this hybrid approach was relatively close to that of the radiologist.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11548-021-02539-2.

SOMA: Subject-, object-, and modality-adapted precision atlas approach for automatic anatomy recognition and delineation in medical images

PURPOSE

In the multi-atlas segmentation (MAS) method, a large enough atlas set, which can cover the complete spectrum of the whole population pattern of the target object will benefit the segmentation quality. However, the difficulty in obtaining and generating such a large set of atlases and the computational burden required in the segmentation procedure make this approach impractical. In this paper, we propose a method called SOMA to select subject-, object-, and modality-adapted precision atlases for automatic anatomy recognition in medical images with pathology, following the idea that different regions of the target object in a novel image can be recognized by different atlases with regionally best similarity, so that effective atlases have no need to be globally similar to the target subject and also have no need to be overall similar to the target object.

METHODS

The SOMA method consists of three main components: atlas building, object recognition, and object delineation. Considering the computational complexity, we utilize an all-to-template strategy to align all images to the same image space belonging to the root image determined by the minimum spanning tree (MST) strategy among a subset of radiologically near-normal images. The object recognition process is composed of two stages: rough recognition and refined recognition. In rough recognition, subimage matching is conducted between the test image and each image of the whole atlas set, and only the atlas corresponding to the best-matched subimage contributes to the recognition map regionally. The frequency of best match for each atlas is recorded by a counter, and the atlases with the highest frequencies are selected as the precision atlases. In refined recognition, only the precision atlases are examined, and the subimage matching is conducted in a nonlocal manner of searching to further increase the accuracy of boundary matching. Delineation is based on a U-net-based deep learning network, where the original gray scale image together with the fuzzy map from refined recognition compose a two-channel input to the network, and the output is a segmentation map of the target object.

RESULTS

Experiments are conducted on computed tomography (CT) images with different qualities in two body regions - head and neck (H&N) and thorax, from 298 subjects with nine objects and 241 subjects with six objects, respectively. Most objects achieve a localization error within two voxels after refined recognition, with marked improvement in localization accuracy from rough to refined recognition of 0.6-3 mm in H&N and 0.8-4.9 mm in thorax, and also in delineation accuracy (Dice coefficient) from refined recognition to delineation of 0.01-0.11 in H&N and 0.01-0.18 in thorax.

CONCLUSIONS

The SOMA method shows high accuracy and robustness in anatomy recognition and delineation. The improvements from rough to refined recognition and further to delineation, as well as immunity of recognition accuracy to varying image and object qualities, demonstrate the core principles of SOMA where segmentation accuracy increases with precision atlases and gradually refined object matching.

Machine Learning for Head and Neck Cancer: A Safe Bet?-A Clinically Oriented Systematic Review for the Radiation Oncologist

BACKGROUND AND PURPOSE

Machine learning (ML) is emerging as a feasible approach to optimize patients' care path in Radiation Oncology. Applications include autosegmentation, treatment planning optimization, and prediction of oncological and toxicity outcomes. The purpose of this clinically oriented systematic review is to illustrate the potential and limitations of the most commonly used ML models in solving everyday clinical issues in head and neck cancer (HNC) radiotherapy (RT).

MATERIALS AND METHODS

Electronic databases were screened up to May 2021. Studies dealing with ML and radiomics were considered eligible. The quality of the included studies was rated by an adapted version of the qualitative checklist originally developed by Luo et al. All statistical analyses were performed using R version 3.6.1.

RESULTS

Forty-eight studies (21 on autosegmentation, four on treatment planning, 12 on oncological outcome prediction, 10 on toxicity prediction, and one on determinants of postoperative RT) were included in the analysis. The most common imaging modality was computed tomography (CT) (40%) followed by magnetic resonance (MR) (10%). Quantitative image features were considered in nine studies (19%). No significant differences were identified in global and methodological scores when works were stratified per their task (i.e., autosegmentation).

DISCUSSION AND CONCLUSION

The range of possible applications of ML in the field of HN Radiation Oncology is wide, albeit this area of research is relatively young. Overall, if not safe yet, ML is most probably a bet worth making.

Automated Motion Analysis of Bony Joint Structures from Dynamic Computer Tomography Images: A Multi-Atlas Approach

Dynamic computer tomography (CT) is an emerging modality to analyze in-vivo joint kinematics at the bone level, but it requires manual bone segmentation and, in some instances, landmark identification. The objective of this study is to present an automated workflow for the assessment of three-dimensional in vivo joint kinematics from dynamic musculoskeletal CT images. The proposed method relies on a multi-atlas, multi-label segmentation and landmark propagation framework to extract bony structures and detect anatomical landmarks on the CT dataset. The segmented structures serve as regions of interest for the subsequent motion estimation across the dynamic sequence. The landmarks are propagated across the dynamic sequence for the construction of bone embedded reference frames from which kinematic parameters are estimated. We applied our workflow on dynamic CT images obtained from 15 healthy subjects on two different joints: thumb base (n = 5) and knee (n = 10). The proposed method resulted in segmentation accuracies of 0.90 ± 0.01 for the thumb dataset and 0.94 ± 0.02 for the knee as measured by the Dice score coefficient. In terms of motion estimation, mean differences in cardan angles between the automated algorithm and manual segmentation, and landmark identification performed by an expert were below 1°. Intraclass correlation (ICC) between cardan angles from the algorithm and results from expert manual landmarks ranged from 0.72 to 0.99 for all joints across all axes. The proposed automated method resulted in reproducible and reliable measurements, enabling the assessment of joint kinematics using 4DCT in clinical routine.

Multi-class medical image segmentation using one-vs-rest graph cuts and majority voting

Purpose: Semi-automatic image segmentation is still a valuable tool in clinical applications since it retains the expert oversights legally required. However, semi-automatic methods for simultaneous multi-class segmentation are difficult to be clinically implemented due to the complexity of underlining algorithms. We purpose an efficient one-vs-rest graph cut approach of which the complexity only grows linearly as the number of classes increases. Approach: Given an image slice, we construct multiple one-vs-rest graphs, each for a tissue class, for inference of a conditional random field (CRF). The one-vs-rest graph cut is to minimize the CRF energy derived from regional and boundary class probabilities estimated from random forests to obtain a one-vs-rest segmentation. The final segmentation is obtained by fusing from those one-vs-rest segmentations based on majority voting. We compare our method to a well-used multi-class graph cut method, alpha-beta swap, and a fully connected CRF (FCCRF) method, in brain tumor segmentation of 20 high-grade tumor cases in 2013 MICCAI dataset. Results: Our method achieved mean Dice score of 0.83 for whole tumor, compared to 0.80 by alpha-beta swap and 0.79 by FCCRF. There was a performance improvement over alpha-beta swap by a factor of five. Conclusions: Our method utilizes the probabilistic-based CRF which can be estimated from any machine learning technique. Comparing to traditional multi-class graph cut, the purposed one-vs-rest approach has complexity that grows only linearly as the number of classes increases, therefore, our method can be applicable for both online semi-automatic and offline automatic segmentation in clinical applications.

A consistent deep registration network with group data modeling

Medical image registration is a critical process for automated image computing, and ideally, the deformation field from one image to another should be smooth and inverse-consistent in order to bidirectionally align anatomical structures and to preserve their topology. Consistent registration can reduce bias caused by the order of input images, increase robustness, and improve reliability of subsequent quantitative analysis. Rigorous differential geometry constraints have been used in traditional methods to enforce the topological consistency but require comprehensive optimization and are time consuming. Recent studies show that deep learning-based registration methods can achieve comparable accuracy and are much faster than traditional registration. However, the estimated deformation fields do not necessarily possess inverse consistency when the order of two input images is swapped. To tackle this problem, we propose a new deep registration algorithm by employing the inverse consistency training strategy, so the forward and backward deformations of a pair of images can consistently align anatomical structures. In addition, since fine-tuned deformations among the training images reflect variability of shapes and appearances in a high-dimensional space, we formulate a group prior data modeling framework so that such statistics can be used to improve accuracy and consistency for registering new input image pairs. Specifically, we implement the wavelet principle component analysis (w-PCA) model of deformation fields and incorporate such prior constraints into the inverse-consistent deep registration network. We refer the proposed algorithm as consistent deep registration with group data modeling. Experiments on 3D brain magnetic resonance (MR) images showed that the unsupervised consistent deep registration and data modeling strategy yield consistent deformations after switching the input images and tolerated image variations well.

Clinical Target Volume Auto-Segmentation of Esophageal Cancer for Radiotherapy After Radical Surgery Based on Deep Learning

Radiotherapy plays an important role in controlling the local recurrence of esophageal cancer after radical surgery. Segmentation of the clinical target volume is a key step in radiotherapy treatment planning, but it is time-consuming and operator-dependent. This paper introduces a deep dilated convolutional U-network to achieve fast and accurate clinical target volume auto-segmentation of esophageal cancer after radical surgery. The deep dilated convolutional U-network, which integrates the advantages of dilated convolution and the U-network, is an end-to-end architecture that enables rapid training and testing. A dilated convolution module for extracting multiscale context features containing the original information on fine texture and boundaries is integrated into the U-network architecture to avoid information loss due to down-sampling and improve the segmentation accuracy. In addition, batch normalization is added to the deep dilated convolutional U-network for fast and stable convergence. In the present study, the training and validation loss tended to be stable after 40 training epochs. This deep dilated convolutional U-network model was able to segment the clinical target volume with an overall mean Dice similarity coefficient of 86.7% and a respective 95% Hausdorff distance of 37.4 mm, indicating reasonable volume overlap of the auto-segmented and manual contours. The mean Cohen kappa coefficient was 0.863, indicating that the deep dilated convolutional U-network was robust. Comparisons with the U-network and attention U-network showed that the overall performance of the deep dilated convolutional U-network was best for the Dice similarity coefficient, 95% Hausdorff distance, and Cohen kappa coefficient. The test time for segmentation of the clinical target volume was approximately 25 seconds per patient. This deep dilated convolutional U-network could be applied in the clinical setting to save time in delineation and improve the consistency of contouring.

FCN Based Label Correction for Multi-Atlas Guided Organ Segmentation

Segmentation of medical images using multiple atlases has recently gained immense attention due to their augmented robustness against variabilities across different subjects. These atlas-based methods typically comprise of three steps: atlas selection, image registration, and finally label fusion. Image registration is one of the core steps in this process, accuracy of which directly affects the final labeling performance. However, due to inter-subject anatomical variations, registration errors are inevitable. The aim of this paper is to develop a deep learning-based confidence estimation method to alleviate the potential effects of registration errors. We first propose a fully convolutional network (FCN) with residual connections to learn the relationship between the image patch pair (i.e., patches from the target subject and the atlas) and the related label confidence patch. With the obtained label confidence patch, we can identify the potential errors in the warped atlas labels and correct them. Then, we use two label fusion methods to fuse the corrected atlas labels. The proposed methods are validated on a publicly available dataset for hippocampus segmentation. Experimental results demonstrate that our proposed methods outperform the state-of-the-art segmentation methods.

Application of Support Vector Machine (SVM) in the Sentiment Analysis of Twitter DataSet

At present, in the mainstream sentiment analysis methods represented by the Support Vector Machine, the vocabulary and the latent semantic information involved in the text are not well considered, and sentiment analysis of text is dependent overly on the statistics of sentiment words. Thus, a Fisher kernel function based on Probabilistic Latent Semantic Analysis is proposed in this paper for sentiment analysis by Support Vector Machine. The Fisher kernel function based on the model is derived from the Probabilistic Latent Semantic Analysis model. By means of this method, latent semantic information involving the probability characteristics can be used as the classification characteristics, along with the improvement of the effect of classification for support vector machine, and the problem of ignoring the latent semantic characteristics in text sentiment analysis can be addressed. The results show that the effect of the method proposed in this paper, compared with the comparison method, is obviously improved.

An Evaluation of Atlas Selection Methods for Atlas-Based Automatic Segmentation in Radiotherapy Treatment Planning

Atlas-based automatic segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed as a way to improve the accuracy and execution time of segmentation, assuming that, the more similar the atlas is to the patient, the better the results will be. This paper presents an analysis of atlas selection methods in the context of radiotherapy treatment planning. For a range of commonly contoured OARs, a thorough comparison of a large class of typical atlas selection methods has been performed. For this evaluation, clinically contoured CT images of the head and neck ( N=316 ) and thorax ( N=280 ) were used. The state-of-the-art intensity and deformation similarity-based atlas selection methods were found to compare poorly to perfect atlas selection. Counter-intuitively, atlas selection methods based on a fixed set of representative atlases outperformed atlas selection methods based on the patient image. This study suggests that atlas-based segmentation with currently available selection methods compares poorly to the potential best performance, hampering the clinical utility of atlas-based segmentation. Effective atlas selection remains an open challenge in atlas-based segmentation for radiotherapy planning.

3D whole brain segmentation using spatially localized atlas network tiles

Detailed whole brain segmentation is an essential quantitative technique in medical image analysis, which provides a non-invasive way of measuring brain regions from a clinical acquired structural magnetic resonance imaging (MRI). Recently, deep convolution neural network (CNN) has been applied to whole brain segmentation. However, restricted by current GPU memory, 2D based methods, downsampling based 3D CNN methods, and patch-based high-resolution 3D CNN methods have been the de facto standard solutions. 3D patch-based high resolution methods typically yield superior performance among CNN approaches on detailed whole brain segmentation (>100 labels), however, whose performance are still commonly inferior compared with state-of-the-art multi-atlas segmentation methods (MAS) due to the following challenges: (1) a single network is typically used to learn both spatial and contextual information for the patches, (2) limited manually traced whole brain volumes are available (typically less than 50) for training a network. In this work, we propose the spatially localized atlas network tiles (SLANT) method to distribute multiple independent 3D fully convolutional networks (FCN) for high-resolution whole brain segmentation. To address the first challenge, multiple spatially distributed networks were used in the SLANT method, in which each network learned contextual information for a fixed spatial location. To address the second challenge, auxiliary labels on 5111 initially unlabeled scans were created by multi-atlas segmentation for training. Since the method integrated multiple traditional medical image processing methods with deep learning, we developed a containerized pipeline to deploy the end-to-end solution. From the results, the proposed method achieved superior performance compared with multi-atlas segmentation methods, while reducing the computational time from >30 h to 15 min. The method has been made available in open source (https://github.com/MASILab/SLANTbrainSeg).

Feasibility of multi-atlas cardiac segmentation from thoracic planning CT in a probabilistic framework

Toxicity to cardiac and coronary structures is an important late morbidity for patients undergoing left-sided breast radiotherapy. Many current studies have relied on estimates of cardiac doses assuming standardised anatomy, with a calculated increase in relative risk of 7.4% per Gy (mean heart dose). To provide individualised estimates for dose, delineation of various cardiac structures on patient images is required. Automatic multi-atlas based segmentation can provide a consistent, robust solution, however there are challenges to this method. We are aiming to develop and validate a cardiac atlas and segmentation framework, with a focus on the limitations and uncertainties in the process. We present a probabilistic approach to segmentation, which provides a simple method to incorporate inter-observer variation, as well as a useful tool for evaluating the accuracy and sources of error in segmentation. A dataset consisting of 20 planning computed tomography (CT) images of Australian breast cancer patients with delineations of 17 structures (including whole heart, four chambers, coronary arteries and valves) was manually contoured by three independent observers, following a protocol based on a published reference atlas, with verification by a cardiologist. To develop and validate the segmentation framework a leave-one-out cross-validation strategy was implemented. Performance of the automatic segmentations was evaluated relative to inter-observer variability in manually-derived contours; measures of volume and surface accuracy (Dice similarity coefficient (DSC) and mean absolute surface distance (MASD), respectively) were used to compare automatic segmentation to the consensus segmentation from manual contours. For the whole heart, the resulting segmentation achieved a DSC of [Formula: see text], with a MASD of [Formula: see text] mm. Quantitative results, together with the analysis of probabilistic labelling, indicate the feasibility of accurate and consistent segmentation of larger structures, whereas this is not the case for many smaller structures, where a major limitation in segmentation accuracy is the inter-observer variability in manual contouring.

Fully automated segmentation of the left atrium, pulmonary veins, and left atrial appendage from magnetic resonance angiography by joint-atlas-optimization

PURPOSE

Atrial fibrillation (AF) originating from the left atrium (LA) and pulmonary veins (PVs) is the most prevalent cardiac electrophysiological disorder. Accurate segmentation and quantification of the LA chamber, PVs, and left atrial appendage (LAA) provides clinically important references for treatment of AF patients. The purpose of this work is to realize objective segmentation of the LA chamber, PVs, and LAA in an accurate and fully automated manner.

METHODS

In this work, we proposed a new approach, named joint-atlas-optimization, to segment the LA chamber, PVs, and LAA from magnetic resonance angiography (MRA) images. We formulated the segmentation as a single registration problem between the given image and all N atlas images, instead of N separate registration between the given image and an individual atlas image. Level sets was applied to refine the atlas-based segmentation. Using the publically available LA benchmark database, we compared the proposed joint-atlas-optimization approach to the conventional pairwise atlas approach and evaluated the segmentation performance in terms of Dice index and surface-to-surface (S2S) distance to the manual ground truth.

RESULTS

The proposed joint-atlas-optimization method showed systemically improved accuracy and robustness over the pairwise atlas approach. The Dice of LA segmentation using joint-atlas-optimization was 0.93 ± 0.04, compared to 0.91 ± 0.04 by the pairwise approach (P < 0.05). The mean S2S distance was 1.52 ± 0.58 mm, compared to 1.83 ± 0.75 mm (P < 0.05). In particular, it produced significantly improved segmentation accuracy of the LAA and PVs, the small distant part in LA geometry that is intrinsically difficult to segment using the conventional pairwise approach. The Dice of PVs segmentation was 0.69 ± 0.16, compared to 0.49 ± 0.15 (P < 0.001). The Dice of LAA segmentation was 0.91 ± 0.03, compared to 0.88 ± 0.05 (P < 0.01).

CONCLUSION

The proposed joint-atlas optimization method can segment the complex LA geometry in a fully automated manner. Compared to the conventional atlas approach in a pairwise manner, our method improves the performance on small distal parts of LA, for example, PVs and LAA, the geometrical and quantitative assessment of which is clinically interesting.

Dynamic Atlas-Based Segmentation and Quantification of Neuromelanin-Rich Brainstem Structures in Parkinson Disease

We present a dynamic atlas composed of neuromelanin-enhanced magnetic resonance brain images of 40 healthy subjects. The performance of this atlas is evaluated on the fully automated segmentation of two paired neuromelanin-rich brainstem healthy structures: the substantia nigra pars compacta and the locus coeruleus. We show that our dynamic atlas requires in average 60% less images and, therefore, 60% less computation time than a static multi-image atlas while achieving a similar segmentation performance. Then, we show that by applying our dynamic atlas, composed of healthy subjects, to the segmentation and neuromelanin quantification of a set of brain images of 39 Parkinson disease patients, we are able to find significant quantitative differences in the level of neuromelanin between healthy subjects and Parkinson disease patients, thus opening the door to the use of these structures as image biomarkers in future computer aided diagnosis systems for the diagnosis of Parkinson disease.

GAS: A genetic atlas selection strategy in multi-atlas segmentation framework

Multi-Atlas based Segmentation (MAS) algorithms have been successfully applied to many medical image segmentation tasks, but their success relies on a large number of atlases and good image registration performance. Choosing well-registered atlases for label fusion is vital for an accurate segmentation. This choice becomes even more crucial when the segmentation involves organs characterized by a high anatomical and pathological variability. In this paper, we propose a new genetic atlas selection strategy (GAS) that automatically chooses the best subset of atlases to be used for segmenting the target image, on the basis of both image similarity and segmentation overlap. More precisely, the key idea of GAS is that if two images are similar, the performances of an atlas for segmenting each image are similar. Since the ground truth of each atlas is known, GAS first selects a predefined number of similar images to the target, then, for each one of them, finds a near-optimal subset of atlases by means of a genetic algorithm. All these near-optimal subsets are then combined and used to segment the target image. GAS was tested on single-label and multi-label segmentation problems. In the first case, we considered the segmentation of both the whole prostate and of the left ventricle of the heart from magnetic resonance images. Regarding multi-label problems, the zonal segmentation of the prostate into peripheral and transition zone was considered. The results showed that the performance of MAS algorithms statistically improved when GAS is used.

Can Atlas-Based Auto-Segmentation Ever Be Perfect? Insights From Extreme Value Theory

Atlas-based segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed to improve the performance of segmentation, assuming that the more similar the atlas is to the patient, the better the result. It follows that the larger the database of atlases from which to select, the better the results should be. This paper seeks to estimate a clinically achievable expected performance under this assumption. Assuming a perfect atlas selection, an extreme value theory has been applied to estimate the accuracy of single-atlas and multi-atlas segmentation given a large database of atlases. For this purpose, clinical contours of most common OARs on computed tomography of the head and neck ( N=316 ) and thoracic ( N=280 ) cases were used. This paper found that while for most organs, perfect segmentation cannot be reasonably expected, auto-contouring performance of a level corresponding to clinical quality could be consistently expected given a database of 5000 atlases under the assumption of perfect atlas selection.

Multi atlas based segmentation: should we prefer the best atlas group over the group of best atlases?

Multi atlas based segmentation (MABS) uses a database of atlas images, and an atlas selection process is used to choose an atlas subset for registration and voting. In the current state of the art, atlases are chosen according to a similarity criterion between the target subject and each atlas in the database. In this paper, we propose a new concept for atlas selection that relies on selecting the best performing group of atlases rather than the group of highest scoring individual atlases. Experiments were performed using CT images of 50 patients, with contours of brainstem and parotid glands. The dataset was randomly split into two groups: 20 volumes were used as an atlas database and 30 served as target subjects for testing. Classic oracle selection, where atlases are chosen by the highest dice similarity coefficient (DSC) with the target, was performed. This was compared to oracle group selection, where all the combinations of atlas subgroups were considered and scored by computing DSC with the target subject. Subsequently, convolutional neural networks were designed to predict the best group of atlases. The results were also compared with the selection strategy based on normalized mutual information (NMI). Oracle group was proven to be significantly better than classic oracle selection (p  <  10^-5). Atlas group selection led to a median  ±  interquartile DSC of 0.740  ±  0.084, 0.718  ±  0.086 and 0.670  ±  0.097 for brainstem and left/right parotid glands respectively, outperforming NMI selection 0.676  ±  0.113, 0.632  ±  0.104 and 0.606  ±  0.118 (p  <  0.001) as well as classic oracle selection. The implemented methodology is a proof of principle that selecting the atlases by considering the performance of the entire group of atlases instead of each single atlas leads to higher segmentation accuracy, being even better then current oracle strategy. This finding opens a new discussion about the most appropriate atlas selection criterion for MABS.

Deep Deconvolutional Neural Network for Target Segmentation of Nasopharyngeal Cancer in Planning Computed Tomography Images

BACKGROUND

Radiotherapy is one of the main treatment methods for nasopharyngeal carcinoma (NPC). It requires exact delineation of the nasopharynx gross tumor volume (GTVnx), the metastatic lymph node gross tumor volume (GTVnd), the clinical target volume (CTV), and organs at risk in the planning computed tomography images. However, this task is time-consuming and operator dependent. In the present study, we developed an end-to-end deep deconvolutional neural network (DDNN) for segmentation of these targets.

METHODS

The proposed DDNN is an end-to-end architecture enabling fast training and testing. It consists of two important components: an encoder network and a decoder network. The encoder network was used to extract the visual features of a medical image and the decoder network was used to recover the original resolution by deploying deconvolution. A total of 230 patients diagnosed with NPC stage I or stage II were included in this study. Data from 184 patients were chosen randomly as a training set to adjust the parameters of DDNN, and the remaining 46 patients were the test set to assess the performance of the model. The Dice similarity coefficient (DSC) was used to quantify the segmentation results of the GTVnx, GTVnd, and CTV. In addition, the performance of DDNN was compared with the VGG-16 model.

RESULTS

The proposed DDNN method outperformed the VGG-16 in all the segmentation. The mean DSC values of DDNN were 80.9% for GTVnx, 62.3% for the GTVnd, and 82.6% for CTV, whereas VGG-16 obtained 72.3, 33.7, and 73.7% for the DSC values, respectively.

CONCLUSION

DDNN can be used to segment the GTVnx and CTV accurately. The accuracy for the GTVnd segmentation was relatively low due to the considerable differences in its shape, volume, and location among patients. The accuracy is expected to increase with more training data and combination of MR images. In conclusion, DDNN has the potential to improve the consistency of contouring and streamline radiotherapy workflows, but careful human review and a considerable amount of editing will be required.

Automatic labeling of MR brain images through extensible learning and atlas forests

PURPOSE

Multiatlas-based method is extensively used in MR brain images segmentation because of its simplicity and robustness. This method provides excellent accuracy although it is time consuming and limited in terms of obtaining information about new atlases. In this study, an automatic labeling of MR brain images through extensible learning and atlas forest is presented to address these limitations.

METHODS

We propose an extensible learning model which allows the multiatlas-based framework capable of managing the datasets with numerous atlases or dynamic atlas datasets and simultaneously ensure the accuracy of automatic labeling. Two new strategies are used to reduce the time and space complexity and improve the efficiency of the automatic labeling of brain MR images. First, atlases are encoded to atlas forests through random forest technology to reduce the time consumed for cross-registration between atlases and target image, and a scatter spatial vector is designed to eliminate errors caused by inaccurate registration. Second, an atlas selection method based on the extensible learning model is used to select atlases for target image without traversing the entire dataset and then obtain the accurate labeling.

RESULTS

The labeling results of the proposed method were evaluated in three public datasets, namely, IBSR, LONI LPBA40, and ADNI. With the proposed method, the dice coefficient metric values on the three datasets were 84.17 ± 4.61%, 83.25 ± 4.29%, and 81.88 ± 4.53% which were 5% higher than those of the conventional method, respectively. The efficiency of the extensible learning model was evaluated by state-of-the-art methods for labeling of MR brain images. Experimental results showed that the proposed method could achieve accurate labeling for MR brain images without traversing the entire datasets.

CONCLUSION

In the proposed multiatlas-based method, extensible learning and atlas forests were applied to control the automatic labeling of brain anatomies on large atlas datasets or dynamic atlas datasets and obtain accurate results.

BrainSegNet: a convolutional neural network architecture for automated segmentation of human brain structures

Automated segmentation of cortical and noncortical human brain structures has been hitherto approached using nonrigid registration followed by label fusion. We propose an alternative approach for this using a convolutional neural network (CNN) which classifies a voxel into one of many structures. Four different kinds of two-dimensional and three-dimensional intensity patches are extracted for each voxel, providing local and global (context) information to the CNN. The proposed approach is evaluated on five different publicly available datasets which differ in the number of labels per volume. The obtained mean Dice coefficient varied according to the number of labels, for example, it is [Formula: see text] and [Formula: see text] for datasets with the least (32) and the most (134) number of labels, respectively. These figures are marginally better or on par with those obtained with the current state-of-the-art methods on nearly all datasets, at a reduced computational time. The consistently good performance of the proposed method across datasets and no requirement for registration make it attractive for many applications where reduced computational time is necessary.

Fast automated segmentation of multiple objects via spatially weighted shape learning

Active shape models (ASMs) have proved successful in automatic segmentation by using shape and appearance priors in a number of areas such as prostate segmentation, where accurate contouring is important in treatment planning for prostate cancer. The ASM approach however, is heavily reliant on a good initialisation for achieving high segmentation quality. This initialisation often requires algorithms with high computational complexity, such as three dimensional (3D) image registration. In this work, we present a fast, self-initialised ASM approach that simultaneously fits multiple objects hierarchically controlled by spatially weighted shape learning. Prominent objects are targeted initially and spatial weights are progressively adjusted so that the next (more difficult, less visible) object is simultaneously initialised using a series of weighted shape models. The scheme was validated and compared to a multi-atlas approach on 3D magnetic resonance (MR) images of 38 cancer patients and had the same (mean, median, inter-rater) Dice's similarity coefficients of (0.79, 0.81, 0.85), while having no registration error and a computational time of 12-15 min, nearly an order of magnitude faster than the multi-atlas approach.

The use of atlas registration and graph cuts for prostate segmentation in magnetic resonance images

PURPOSE

An automatic method for 3D prostate segmentation in magnetic resonance (MR) images is presented for planning image-guided radiotherapy treatment of prostate cancer.

METHODS

A spatial prior based on intersubject atlas registration is combined with organ-specific intensity information in a graph cut segmentation framework. The segmentation is tested on 67 axial T2-weighted MR images in a leave-one-out cross validation experiment and compared with both manual reference segmentations and with multiatlas-based segmentations using majority voting atlas fusion. The impact of atlas selection is investigated in both the traditional atlas-based segmentation and the new graph cut method that combines atlas and intensity information in order to improve the segmentation accuracy. Best results were achieved using the method that combines intensity information, shape information, and atlas selection in the graph cut framework.

RESULTS

A mean Dice similarity coefficient (DSC) of 0.88 and a mean surface distance (MSD) of 1.45 mm with respect to the manual delineation were achieved.

CONCLUSIONS

This approaches the interobserver DSC of 0.90 and interobserver MSD 0f 1.15 mm and is comparable to other studies performing prostate segmentation in MR.

Brain extraction in pediatric ADC maps, toward characterizing neuro-development in multi-platform and multi-institution clinical images

Apparent Diffusion Coefficient (ADC) maps can be used to characterize myelination and to detect abnormalities in the developing brain. However, given the normal variation in regional ADC with myelination, detection of abnormalities is difficult when based on visual assessment. Quantitative and automated analysis of pediatric ADC maps is thus desired but requires accurate brain extraction as the first step. Currently, most existing brain extraction methods are optimized for structural T1-weighted MR images of fully myelinated brains. Due to differences in age and image contrast, these approaches do not translate well to pediatric ADC maps. To address this problem, we present a multi-atlas brain extraction framework that has 1) specificity: designed and optimized specifically for pediatric ADC maps; 2) generality: applicable to multi-platform and multi-institution data, and to subjects at various neuro-developmental stages across the first 6 years of life; 3) accuracy: highly accurate compared to expert annotations; and 4) consistency: consistently accurate regardless of sources of data and ages of subjects. We show how we achieve these goals, via optimizing major components in a multi-atlas brain extraction framework, and via developing and evaluating new criteria for its atlas ranking component. Moreover, we demonstrate that these goals can be achieved with a fixed set of atlases and a fixed set of parameters, which opens doors for our optimized framework to be used in large-scale and multi-institution neuro-developmental and clinical studies. In a pilot study, we use this framework in a dataset containing scanner-generated ADC maps from 308 pediatric patients collected during the course of routine clinical care. Our framework leads to successful quantifications of the changes in whole-brain volumes and mean ADC values across the first 6 years of life.

Atlas-guided prostate intensity modulated radiation therapy (IMRT) planning

An atlas-based IMRT planning technique for prostate cancer was developed and evaluated. A multi-dose atlas was built based on the anatomy patterns of the patients, more specifically, the percent distance to the prostate and the concaveness angle formed by the seminal vesicles relative to the anterior-posterior axis. A 70-case dataset was classified using a k-medoids clustering analysis to recognize anatomy pattern variations in the dataset. The best classification, defined by the number of classes or medoids, was determined by the largest value of the average silhouette width. Reference plans from each class formed a multi-dose atlas. The atlas-guided planning (AGP) technique started with matching the new case anatomy pattern to one of the reference cases in the atlas; then a deformable registration between the atlas and new case anatomies transferred the dose from the atlas to the new case to guide inverse planning with full automation. 20 additional clinical cases were re-planned to evaluate the AGP technique. Dosimetric properties between AGP and clinical plans were evaluated. The classification analysis determined that the 5-case atlas would best represent anatomy patterns for the patient cohort. AGP took approximately 1 min on average (corresponding to 70 iterations of optimization) for all cases. When dosimetric parameters were compared, the differences between AGP and clinical plans were less than 3.5%, albeit some statistical significances observed: homogeneity index (p  >  0.05), conformity index (p  <  0.01), bladder gEUD (p  <  0.01), and rectum gEUD (p  =  0.02). Atlas-guided treatment planning is feasible and efficient. Atlas predicted dose can effectively guide the optimizer to achieve plan quality comparable to that of clinical plans.

Multi-atlas learner fusion: An efficient segmentation approach for large-scale data

We propose multi-atlas learner fusion (MLF), a framework for rapidly and accurately replicating the highly accurate, yet computationally expensive, multi-atlas segmentation framework based on fusing local learners. In the largest whole-brain multi-atlas study yet reported, multi-atlas segmentations are estimated for a training set of 3464 MR brain images. Using these multi-atlas estimates we (1) estimate a low-dimensional representation for selecting locally appropriate example images, and (2) build AdaBoost learners that map a weak initial segmentation to the multi-atlas segmentation result. Thus, to segment a new target image we project the image into the low-dimensional space, construct a weak initial segmentation, and fuse the trained, locally selected, learners. The MLF framework cuts the runtime on a modern computer from 36 h down to 3-8 min - a 270× speedup - by completely bypassing the need for deformable atlas-target registrations. Additionally, we (1) describe a technique for optimizing the weak initial segmentation and the AdaBoost learning parameters, (2) quantify the ability to replicate the multi-atlas result with mean accuracies approaching the multi-atlas intra-subject reproducibility on a testing set of 380 images, (3) demonstrate significant increases in the reproducibility of intra-subject segmentations when compared to a state-of-the-art multi-atlas framework on a separate reproducibility dataset, (4) show that under the MLF framework the large-scale data model significantly improve the segmentation over the small-scale model under the MLF framework, and (5) indicate that the MLF framework has comparable performance as state-of-the-art multi-atlas segmentation algorithms without using non-local information.

Multi-atlas segmentation of biomedical images: A survey

Multi-atlas segmentation (MAS), first introduced and popularized by the pioneering work of Rohlfing, et al. (2004), Klein, et al. (2005), and Heckemann, et al. (2006), is becoming one of the most widely-used and successful image segmentation techniques in biomedical applications. By manipulating and utilizing the entire dataset of "atlases" (training images that have been previously labeled, e.g., manually by an expert), rather than some model-based average representation, MAS has the flexibility to better capture anatomical variation, thus offering superior segmentation accuracy. This benefit, however, typically comes at a high computational cost. Recent advancements in computer hardware and image processing software have been instrumental in addressing this challenge and facilitated the wide adoption of MAS. Today, MAS has come a long way and the approach includes a wide array of sophisticated algorithms that employ ideas from machine learning, probabilistic modeling, optimization, and computer vision, among other fields. This paper presents a survey of published MAS algorithms and studies that have applied these methods to various biomedical problems. In writing this survey, we have three distinct aims. Our primary goal is to document how MAS was originally conceived, later evolved, and now relates to alternative methods. Second, this paper is intended to be a detailed reference of past research activity in MAS, which now spans over a decade (2003-2014) and entails novel methodological developments and application-specific solutions. Finally, our goal is to also present a perspective on the future of MAS, which, we believe, will be one of the dominant approaches in biomedical image segmentation.

Morphometric Atlas Selection for Automatic Brachial Plexus Segmentation

PURPOSE

The purpose of this study was to determine the effects of atlas selection based on different morphometric parameters, on the accuracy of automatic brachial plexus (BP) segmentation for radiation therapy planning. The segmentation accuracy was measured by comparing all of the generated automatic segmentations with anatomically validated gold standard atlases developed using cadavers.

METHODS AND MATERIALS

Twelve cadaver computed tomography (CT) atlases (3 males, 9 females; mean age: 73 years) were included in the study. One atlas was selected to serve as a patient, and the other 11 atlases were registered separately onto this "patient" using deformable image registration. This procedure was repeated for every atlas as a patient. Next, the Dice and Jaccard similarity indices and inclusion index were calculated for every registered BP with the original gold standard BP. In parallel, differences in several morphometric parameters that may influence the BP segmentation accuracy were measured for the different atlases. Specific brachial plexus-related CT-visible bony points were used to define the morphometric parameters. Subsequently, correlations between the similarity indices and morphometric parameters were calculated.

RESULTS

A clear negative correlation between difference in protraction-retraction distance and the similarity indices was observed (mean Pearson correlation coefficient = -0.546). All of the other investigated Pearson correlation coefficients were weak.

CONCLUSIONS

Differences in the shoulder protraction-retraction position between the atlas and the patient during planning CT influence the BP autosegmentation accuracy. A greater difference in the protraction-retraction distance between the atlas and the patient reduces the accuracy of the BP automatic segmentation result.

Robust inverse-consistent affine CT-MR registration in MRI-assisted and MRI-alone prostate radiation therapy

BACKGROUND

CT-MR registration is a critical component of many radiation oncology protocols. In prostate external beam radiation therapy, it allows the propagation of MR-derived contours to reference CT images at the planning stage, and it enables dose mapping during dosimetry studies. The use of carefully registered CT-MR atlases allows the estimation of patient specific electron density maps from MRI scans, enabling MRI-alone radiation therapy planning and treatment adaptation. In all cases, the precision and accuracy achieved by registration influences the quality of the entire process.

PROBLEM

Most current registration algorithms do not robustly generalize and lack inverse-consistency, increasing the risk of human error and acting as a source of bias in studies where information is propagated in a particular direction, e.g. CT to MR or vice versa. In MRI-based treatment planning where both CT and MR scans serve as spatial references, inverse-consistency is critical, if under-acknowledged.

PURPOSE

A robust, inverse-consistent, rigid/affine registration algorithm that is well suited to CT-MR alignment in prostate radiation therapy is presented.

METHOD

The presented method is based on a robust block-matching optimization process that utilises a half-way space definition to maintain inverse-consistency. Inverse-consistency substantially reduces the influence of the order of input images, simplifying analysis, and increasing robustness. An open source implementation is available online at http://aehrc.github.io/Mirorr/.

RESULTS

Experimental results on a challenging 35 CT-MR pelvis dataset demonstrate that the proposed method is more accurate than other popular registration packages and is at least as accurate as the state of the art, while being more robust and having an order of magnitude higher inverse-consistency than competing approaches.

CONCLUSION

The presented results demonstrate that the proposed registration algorithm is readily applicable to prostate radiation therapy planning.

A multi-atlas-based segmentation framework for prostate brachytherapy

Low-dose-rate brachytherapy is a radiation treatment method for localized prostate cancer. The standard of care for this treatment procedure is to acquire transrectal ultrasound images of the prostate in order to devise a plan to deliver sufficient radiation dose to the cancerous tissue. Brachytherapy planning involves delineation of contours in these images, which closely follow the prostate boundary, i.e., clinical target volume. This process is currently performed either manually or semi-automatically, which requires user interaction for landmark initialization. In this paper, we propose a multi-atlas fusion framework to automatically delineate the clinical target volume in ultrasound images. A dataset of a priori segmented ultrasound images, i.e., atlases, is registered to a target image. We introduce a pairwise atlas agreement factor that combines an image-similarity metric and similarity between a priori segmented contours. This factor is used in an atlas selection algorithm to prune the dataset before combining the atlas contours to produce a consensus segmentation. We evaluate the proposed segmentation approach on a set of 280 transrectal prostate volume studies. The proposed method produces segmentation results that are within the range of observer variability when compared to a semi-automatic segmentation technique that is routinely used in our cancer clinic.

A fully automatic multi-atlas based segmentation method for prostate MR images

Most of multi-atlas segmentation methods focus on the registration between the full-size volumes of the data set. Although the transformations obtained from these registrations may be accurate for the global field of view of the images, they may not be accurate for the local prostate region. This is because different magnetic resonance (MR) images have different fields of view and may have large anatomical variability around the prostate. To overcome this limitation, we proposed a two-stage prostate segmentation method based on a fully automatic multi-atlas framework, which includes the detection stage i.e. locating the prostate, and the segmentation stage i.e. extracting the prostate. The purpose of the first stage is to find a cuboid that contains the whole prostate as small cubage as possible. In this paper, the cuboid including the prostate is detected by registering atlas edge volumes to the target volume while an edge detection algorithm is applied to every slice in the volumes. At the second stage, the proposed method focuses on the registration in the region of the prostate vicinity, which can improve the accuracy of the prostate segmentation. We evaluated the proposed method on 12 patient MR volumes by performing a leave-one-out study. Dice similarity coefficient (DSC) and Hausdorff distance (HD) are used to quantify the difference between our method and the manual ground truth. The proposed method yielded a DSC of 83.4%±4.3%, and a HD of 9.3 mm±2.6 mm. The fully automated segmentation method can provide a useful tool in many prostate imaging applications.

Kernel-based atlas image selection for brain tissue segmentation

We propose a new Kernel-based Atlas Image Selection computed in the Embedding Representation space (termed KAISER) aiming to support labeling of brain tissue on 3D magnetic resonance (MR) images. KAISER approach provides efficient feature extraction from MR volumes based on an introduced inter-slice kernel (ISK). Thus, using the ISK matrix eigendecomposition, the inherent structure of data distribution is accentuated through estimation of low dimensional compact space where every pair-wise image similarity can be better measured. We compare our proposal against the whole-population atlas, randomly and demographically selected multiatlas approaches in a four-tissue image labeling task. Obtained results show that the KAISER approach outperforms other alternative techniques (98% Dice index similarity against 94%), while exhibiting better repeatability.

Automated segmentation of breast in 3-D MR images using a robust atlas

This paper presents a robust atlas-based segmentation (ABS) algorithm for segmentation of the breast boundary in 3-D MR images. The proposed algorithm combines the well-known methodologies of ABS namely probabilistic atlas and atlas selection approaches into a single framework where two configurations are realized. The algorithm uses phase congruency maps to create an atlas which is robust to intensity variations. This allows an atlas derived from images acquired with one MR imaging sequence to be used to segment images acquired with a different MR imaging sequence and eliminates the need for intensity-based registration. Images acquired using a Dixon sequence were used to create an atlas which was used to segment both Dixon images (intra-sequence) and T1-weighted images (inter-sequence). In both cases, highly accurate results were achieved with the median Dice similarity coefficient values of 94% ±4% and 87 ±6.5%, respectively.

Multi-atlas segmentation of subcortical brain structures via the AutoSeg software pipeline

Automated segmenting and labeling of individual brain anatomical regions, in MRI are challenging, due to the issue of individual structural variability. Although atlas-based segmentation has shown its potential for both tissue and structure segmentation, due to the inherent natural variability as well as disease-related changes in MR appearance, a single atlas image is often inappropriate to represent the full population of datasets processed in a given neuroimaging study. As an alternative for the case of single atlas segmentation, the use of multiple atlases alongside label fusion techniques has been introduced using a set of individual “atlases” that encompasses the expected variability in the studied population. In our study, we proposed a multi-atlas segmentation scheme with a novel graph-based atlas selection technique. We first paired and co-registered all atlases and the subject MR scans. A directed graph with edge weights based on intensity and shape similarity between all MR scans is then computed. The set of neighboring templates is selected via clustering of the graph. Finally, weighted majority voting is employed to create the final segmentation over the selected atlases. This multi-atlas segmentation scheme is used to extend a single-atlas-based segmentation toolkit entitled AutoSeg, which is an open-source, extensible C++ based software pipeline employing BatchMake for its pipeline scripting, developed at the Neuro Image Research and Analysis Laboratories of the University of North Carolina at Chapel Hill. AutoSeg performs N4 intensity inhomogeneity correction, rigid registration to a common template space, automated brain tissue classification based skull-stripping, and the multi-atlas segmentation. The multi-atlas-based AutoSeg has been evaluated on subcortical structure segmentation with a testing dataset of 20 adult brain MRI scans and 15 atlas MRI scans. The AutoSeg achieved mean Dice coefficients of 81.73% for the subcortical structures.