Accurate Tumor Segmentation Research Articles

A Bayesian approach to tissue-fraction estimation for oncological PET segmentation

Tumor segmentation in oncological PET is challenging, a major reason being the partial-volume effects (PVEs) that arise due to low system resolution and finite voxel size. The latter results in tissue-fraction effects (TFEs), i.e. voxels contain a mixture of tissue classes. Conventional segmentation methods are typically designed to assign each image voxel as belonging to a certain tissue class. Thus, these methods are inherently limited in modeling TFEs. To address the challenge of accounting for PVEs, and in particular, TFEs, we propose a Bayesian approach to tissue-fraction estimation for oncological PET segmentation. Specifically, this Bayesian approach estimates the posterior mean of the fractional volume that the tumor occupies within each image voxel. The proposed method, implemented using a deep-learning-based technique, was first evaluated using clinically realistic 2D simulation studies with known ground truth, in the context of segmenting the primary tumor in PET images of patients with lung cancer. The evaluation studies demonstrated that the method accurately estimated the tumor-fraction areas and significantly outperformed widely used conventional PET segmentation methods, including a U-net-based method, on the task of segmenting the tumor. In addition, the proposed method was relatively insensitive to PVEs and yielded reliable tumor segmentation for different clinical-scanner configurations. The method was then evaluated using clinical images of patients with stage IIB/III non-small cell lung cancer from ACRIN 6668/RTOG 0235 multi-center clinical trial. Here, the results showed that the proposed method significantly outperformed all other considered methods and yielded accurate tumor segmentation on patient images with Dice similarity coefficient (DSC) of 0.82 (95% CI: 0.78, 0.86). In particular, the method accurately segmented relatively small tumors, yielding a high DSC of 0.77 for the smallest segmented cross-section of 1.30 cm2. Overall, this study demonstrates the efficacy of the proposed method to accurately segment tumors in PET images.

PA-ResSeg: A phase attention residual network for liver tumor segmentation from multiphase CT images.

Liver tumor segmentation is a crucial prerequisite for computer-aided diagnosis of liver tumors. In the clinical diagnosis of liver tumors, radiologists usually examine multiphase CT images as these images provide abundant and complementary information of tumors. However, most known automatic segmentation methods extract tumor features from CT images merely of a single phase, in which valuable multiphase information is ignored. Therefore, it is highly demanded to develop a method effectively incorporating multiphase information for automatic and accurate liver tumor segmentation. In this paper, we propose a phase attention residual network (PA-ResSeg) to model multiphase features for accurate liver tumor segmentation. A phase attention (PA) is newly proposed to additionally exploit the images of arterial (ART) phase to facilitate the segmentation of portal venous (PV) phase. The PA block consists of an intraphase attention (intra-PA) module and an interphase attention (inter-PA) module to capture channel-wise self-dependencies and cross-phase interdependencies, respectively. Thus, it enables the network to learn more representative multiphase features by refining the PV features according to the channel dependencies and recalibrating the ART features based on the learned interdependencies between phases. We propose a PA-based multiscale fusion (MSF) architecture to embed the PA blocks in the network at multiple levels along the encoding path to fuse multiscale features from multiphase images. Moreover, a 3D boundary-enhanced loss (BE-loss) is proposed for training to make the network more sensitive to boundaries. To evaluate the performance of our proposed PA-ResSeg, we conducted experiments on a multiphase CT dataset of focal liver lesions (MPCT-FLLs). Experimental results show the effectiveness of the proposed method by achieving a dice per case (DPC) of 0.7787, a dice global (DG) of 0.8682, a volumetric overlap error (VOE) of 0.3328, and a relative volume difference (RVD) of 0.0443 on the MPCT-FLLs. Furthermore, to validate the effectiveness and robustness of PA-ResSeg, we conducted extra experiments on another multiphase liver tumor dataset and obtained a DPC of 0.8290, a DG of 0.9132, a VOE of 0.2637, and a RVD of 0.0163. The proposed method shows its robustness and generalization capability in different datasets and different backbones. The study demonstrates that our method can effectively model information from multiphase CT images to segment liver tumors and outperforms other state-of-the-art methods. The PA-based MSF method can learn more representative multiphase features at multiple scales and thereby improve the segmentation performance. Besides, the proposed 3D BE-loss is conducive to tumor boundary segmentation by enforcing the network focus on boundary regions and marginal slices. Experimental results evaluated by quantitative metrics demonstrate the superiority of our PA-ResSeg over the best-known methods.

Unsupervised brain tumor segmentation using a symmetric-driven adversarial network

The aim of this study was to computationally model, in an unsupervised manner, a manifold of symmetry variations in normal brains, such that the learned manifold can be used to segment brain tumors from magnetic resonance (MR) images that fail to exhibit symmetry. An unsupervised brain tumor segmentation method, named as symmetric driven generative adversarial network (SD-GAN), was proposed. SD-GAN model was trained to learn a non-linear mapping between the left and right brain images, and thus being able to present the variability of the (symmetry) normal brains. The trained SD-GAN was then used to reconstruct normal brains and to segment brain tumors based on higher reconstruction errors arising from their being unsymmetrical. SD-GAN was evaluated on two public benchmark datasets (Multi-modal Brain Tumor Image Segmentation (BRATS) 2012 and 2018). SD-GAN provided best performance with tumor segmentation accuracy superior to the state-of-the-art unsupervised segmentation methods and performed comparably (<3% lower in Dice score) to the supervised U-Net (the most widely used supervised method for medical images). This study demonstrated that symmetric features presenting variations (i.e., inherent anatomical variations) can be modelled using unannotated normal MR images and thus be used in segmenting tumors.

Brain tumor segmentation based on deep learning and an attention mechanism using MRI multi-modalities brain images

Brain tumor localization and segmentation from magnetic resonance imaging (MRI) are hard and important tasks for several applications in the field of medical analysis. As each brain imaging modality gives unique and key details related to each part of the tumor, many recent approaches used four modalities T1, T1c, T2, and FLAIR. Although many of them obtained a promising segmentation result on the BRATS 2018 dataset, they suffer from a complex structure that needs more time to train and test. So, in this paper, to obtain a flexible and effective brain tumor segmentation system, first, we propose a preprocessing approach to work only on a small part of the image rather than the whole part of the image. This method leads to a decrease in computing time and overcomes the overfitting problems in a Cascade Deep Learning model. In the second step, as we are dealing with a smaller part of brain images in each slice, a simple and efficient Cascade Convolutional Neural Network (C-ConvNet/C-CNN) is proposed. This C-CNN model mines both local and global features in two different routes. Also, to improve the brain tumor segmentation accuracy compared with the state-of-the-art models, a novel Distance-Wise Attention (DWA) mechanism is introduced. The DWA mechanism considers the effect of the center location of the tumor and the brain inside the model. Comprehensive experiments are conducted on the BRATS 2018 dataset and show that the proposed model obtains competitive results: the proposed method achieves a mean whole tumor, enhancing tumor, and tumor core dice scores of 0.9203, 0.9113 and 0.8726 respectively. Other quantitative and qualitative assessments are presented and discussed.

Accurate Tumor Segmentation via Octave Convolution Neural Network.

Three-dimensional (3D) liver tumor segmentation from Computed Tomography (CT) images is a prerequisite for computer-aided diagnosis, treatment planning, and monitoring of liver cancer. Despite many years of research, 3D liver tumor segmentation remains a challenging task. In this paper, we propose an effective and efficient method for tumor segmentation in liver CT images using encoder-decoder based octave convolution networks. Compared with other convolution networks utilizing standard convolution for feature extraction, the proposed method utilizes octave convolutions for learning multiple-spatial-frequency features, thus can better capture tumors with varying sizes and shapes. The proposed network takes advantage of a fully convolutional architecture which performs efficient end-to-end learning and inference. More importantly, we introduce a deep supervision mechanism during the learning process to combat potential optimization difficulties, and thus the model can acquire a much faster convergence rate and more powerful discrimination capability. Finally, we integrate octave convolutions into the encoder-decoder architecture of UNet, which can generate high resolution tumor segmentation in one single forward feeding without post-processing steps. Both architectures are trained on a subset of the LiTS (Liver Tumor Segmentation) Challenge. The proposed approach is shown to significantly outperform other networks in terms of various accuracy measures and processing speed.

Three-dimensional multipath DenseNet for improving automatic segmentation of glioblastoma on pre-operative multimodal MR images.

Convolutional neural networks have achieved excellent results in automatic medical image segmentation. In this study, we proposed a novel three-dimensional (3D) multipath DenseNet for generating the accurate glioblastoma (GBM) tumor contour from four multimodal pre-operative MR images. We hypothesized that the multipath architecture could achieve more accurate segmentation than a singlepath architecture. Two hundred and fifty-eight GBM patients were included in this study. Each patient had four MR images (T1-weighted, contrast-enhanced T1-weighted, T2-weighted, and FLAIR) and the manually segmented tumor contour. We built a 3D multipath DenseNet that could be trained to achieve an end-to-end mapping from four MR images to the corresponding GBM tumor contour. A 3D singlepath DenseNet was also built for comparison. Both DenseNets were based on the encoder-decoder architecture. All four images were concatenated and fed into a single encoder path in the singlepath DenseNet, while each input image had its own encoder path in the multipath DenseNet. The patient cohort was randomly split into a training set of 180 patients, a validation set of 39 patients, and a testing set of 39 patients. Model performance was evaluated using the Dice similarity coefficient (DSC), average surface distance (ASD), and 95% Hausdorff distance (HD95% ). Wilcoxon signed-rank tests were conducted to assess statistical significances. The singlepath DenseNet achieved the DSC of 0.911±0.060, ASD of 1.3±0.7mm, and HD95% of 5.2±7.1mm, while the multipath DenseNet achieved the DSC of 0.922±0.041, ASD of 1.1±0.5mm, and HD95% of 3.9±3.3mm. The P-values of all Wilcoxon signed-rank tests were less than 0.05. Both DenseNets generated GBM tumor contours in good agreement with the manually segmented contours from multimodal MR images. The multipath DenseNet achieved more accurate tumor segmentation than the singlepath DenseNet. Here presented the 3D multipath DenseNet that demonstrated an improved accuracy over comparable algorithms in the clinical task of GBM tumor segmentation.

A Comparison of Deep Learning and Pharmacokinetic Model Selection Methods in Segmentation of High-Grade Glioma

Purpose: Glioma tumor segmentation is an essential step in clinical decision making. Recently, computer-aided methods have been widely used for rapid and accurate delineation of the tumor regions. Methods based on image feature extraction can be used as fast methods, while segmentation based on the physiology and pharmacokinetic of the tissues is more accurate. This study aims to compare the performance of tumor segmentation based on these two different methods. Materials and Methods: Nested Model Selection (NMS) based on Extended-Toft’s model was applied to 190 Dynamic Contrast-Enhanced MRI (DCE-MRI) slices acquired from 25 Glioblastoma Multiforme (GBM) patients in 70 time-points. A model with three pharmacokinetic parameters, Model 3, is usually assigned to tumor voxel based on the time-contrast concentration signal. We utilized Deep-Net as a CNN network, based on Deeplabv3+ and layers of pre-trained resnet18, which has been trained with 17288 T1-Contrast MRI slices with HGG brain tumor to predict the tumor region in our 190 DCE MRI T1 images. The NMS-based physiological tumor segmentation was considered as a reference to compare the results of tumor segmentation by Deep-Net. Dice, Jaccard, and overlay similarity coefficients were used to evaluate the tumor segmentation accuracy and reliability of the Deep tumor segmentation method. Results: The results showed a relatively high similarity coefficient (Dice coefficient: 0.73±0.15, Jaccard coefficient: 0.66±0.17, and overlay coefficient: 0.71±0.15) between deep learning tumor segmentation and the tumor region identified by the NMS method. The results indicate that the deep learning methods may be used as accurate and robust tumor segmentation. Conclusion: Deep learning-based segmentation can play a significant role to increase the segmentation accuracy in clinical application, if their training process is completely automatic and independent from human error.

LBTS-Net: A fast and accurate CNN model for brain tumour segmentation.

An accurate tumour segmentation in brain images is a complicated task due to the complext structure and irregular shape of the tumour. In this letter, our contribution is twofold: (1) a lightweight brain tumour segmentation network (LBTS‐Net) is proposed for a fast yet accurate brain tumour segmentation; (2) transfer learning is integrated within the LBTS‐Net to fine‐tune the network and achieve a robust tumour segmentation. To the best of knowledge, this work is amongst the first in the literature which proposes a lightweight and tailored convolution neural network for brain tumour segmentation. The proposed model is based on the VGG architecture in which the number of convolution filters is cut to half in the first layer and the depth‐wise convolution is employed to lighten the VGG‐16 and VGG‐19 networks. Also, the original pixel‐labels in the LBTS‐Net are replaced by the new tumour labels in order to form the classification layer. Experimental results on the BRATS2015 database and comparisons with the state‐of‐the‐art methods confirmed the robustness of the proposed method achieving a global accuracy and a Dice score of 98.11% and 91%, respectively, while being much more computationally efficient due to containing almost half the number of parameters as in the standard VGG network.

CBSN: Comparative measures of normalization techniques for brain tumor segmentation using SRCNet

The segmentation of tumors in the brain MRI scans is a difficult job for doctors and radiologists. The segmentation done by different medical experts may also have differences in their opinion for the segmented region, which is popularly known as regions of interest (ROIs). To date, researchers and academicians have proposed several approaches and frameworks for semi- and full-automatic segmentation techniques to identify ROIs accurately. It is prevalent that automatic segmentation gives comparable or even better results compared to human experts for several publicly known and privately collected datasets. Additionally, these are beneficial in those areas where doctors and radiologists’ availability is either uneven or scarce because of geographical dispersion. The convolutional neural networks (CNN) are considered for segmentation for ROIs due to their wide popularity. They have outperformed humans over tasks like object identification and image classification. The publicly available datasets or those collected from different medical institutions may have different statistics, resolution, and properties. Therefore, pre-processing has an essential role in achieving better and accurate delineation and segmentation of tumors. In the proposed work, CBSN, we consider well-known normalization techniques such as Gaussian Mixture Models (GMM), Fuzzy C-Means (FCM), and Z-score normalization for pre-processing the BraTS (Brain Tumor Segmentation Challenge) 2018 dataset. We utilized three variants of U-Net architecture, convolutional block attention module (CBAM), squeeze and excitation module (SEM), and refinement module (RM) for the segmentation of the ROIs. Utilizing Z-score performs better than other normalization techniques for tumor core (TC) and whole tumor (WT) segmentation. In contrast, FCM performs superior to the other two normalization techniques on enhancement tumor (ET) segmentation.

Weakly-Supervised teacher-Student network for liver tumor segmentation from non-enhanced images.

Accurate liver tumor segmentation without contrast agents (non-enhanced images) avoids the contrast-agent-associated time-consuming and high risk, which offers radiologists quick and safe assistance to diagnose and treat the liver tumor. However, without contrast agents enhancing, the tumor in liver images presents low contrast and even invisible to naked eyes. Thus the liver tumor segmentation from non-enhanced images is quite challenging. We propose a Weakly-Supervised Teacher-Student network (WSTS) to address the liver tumor segmentation in non-enhanced images by leveraging additional box-level-labeled data (labeled with a tumor bounding-box). WSTS deploys a weakly-supervised teacher-student framework (TCH-ST), namely, a Teacher Module learns to detect and segment the tumor in enhanced images during training, which facilitates a Student Module to detect and segment the tumor in non-enhanced images independently during testing. To detect the tumor accurately, the WSTS proposes a Dual-strategy DRL (DDRL), which develops two tumor detection strategies by creatively introducing a relative-entropy bias in the DRL. To accurately predict a tumor mask for the box-level-labeled enhanced image and thus improve tumor segmentation in non-enhanced images, the WSTS proposes an Uncertainty-Sifting Self-Ensembling (USSE). The USSE exploits the weakly-labeled data with self-ensembling and evaluates the prediction reliability with a newly-designed Multi-scale Uncertainty-estimation. WSTS is validated with a 2D MRI dataset, where the experiment achieves 83.11% of Dice and 85.12% of Recall in 50 patient testing data after training by 200 patient data (half amount data is box-level-labeled). Such a great result illustrates the competence of WSTS to segment the liver tumor from non-enhanced images. Thus, WSTS has excellent potential to assist radiologists by liver tumor segmentation without contrast-agents.

Closing the Gap Between Deep Neural Network Modeling and Biomedical Decision-Making Metrics in Segmentation via Adaptive Loss Functions.

Deep learning is becoming an indispensable tool for various tasks in science and engineering. A critical step in constructing a reliable deep learning model is the selection of a loss function, which measures the discrepancy between the network prediction and the ground truth. While a variety of loss functions have been proposed in the literature, a truly optimal loss function that maximally utilizes the capacity of neural networks for deep learning-based decision-making has yet to be established. Here, we devise a generalized loss function with functional parameters determined adaptively during model training to provide a versatile framework for optimal neural network-based decision-making in small target segmentation. The method is showcased by more accurate detection and segmentation of lung and liver cancer tumors as compared with the current state-of-the-art. The proposed formalism opens new opportunities for numerous practical applications such as disease diagnosis, treatment planning, and prognosis.

Study on the Segmentation Method of Multi-Phase CT Liver Tumor Based on Dual Channel U-Nets

As a common malignant tumor disease, hepatocellular carcinoma is the most common cancers in the world. The incidence of hepatocellular carcinoma in China is higher than that in the world. Therefore, it is very important for doctors to separate liver and tumor from CT images by means of computer-aided diagnosis and treatment. In this paper, a multiscale DC-CUNets network liver tumor segmentation method is proposed to enhance the fusion of multi-phase image features in CT, the scale of liver tumors, and the optimization of network training process. (1) A multistage CT liver tumor segmentation method based on two-channel cascaded U-Nets (DC-CUNets) is proposed. The liver was segmented using the first-order U-Net, and then the segmented area of interest of the liver was input into the second-order U-Net network to segment liver tumors. We designed two-channel U-Nets to learn the image characteristics of CT images in arterial and venous phases respectively, and to achieve two-channel feature fusion through feature cascade to improve the overall accuracy of liver tumor segmentation.(2) A multistage CT liver tumor segmentation method based on multiscale DC-CUNets was proposed. For the scale problem of liver tumors, we designed a two-layer multiscale void convolution module to obtain image features at different scales for large, medium and small tumors, and fuse the multiscale features at the output of the module. We have replaced the convolution layer of the fourth module in the second-order two-channel liver tumor segmentation U-Nets by the two-layer multiscale cavity convolution module to implement multiscale DC-CUNets.

Repeatability of two semi-automatic artificial intelligence approaches for tumor segmentation in PET

BackgroundPositron emission tomography (PET) is routinely used for cancer staging and treatment follow-up. Metabolic active tumor volume (MATV) as well as total MATV (TMATV—including primary tumor, lymph nodes and metastasis) and/or total lesion glycolysis derived from PET images have been identified as prognostic factor or for the evaluation of treatment efficacy in cancer patients. To this end, a segmentation approach with high precision and repeatability is important. However, the implementation of a repeatable and accurate segmentation algorithm remains an ongoing challenge.MethodsIn this study, we compare two semi-automatic artificial intelligence (AI)-based segmentation methods with conventional semi-automatic segmentation approaches in terms of repeatability. One segmentation approach is based on a textural feature (TF) segmentation approach designed for accurate and repeatable segmentation of primary tumors and metastasis. Moreover, a convolutional neural network (CNN) is trained. The algorithms are trained, validated and tested using a lung cancer PET dataset. The segmentation accuracy of both segmentation approaches is compared using the Jaccard coefficient (JC). Additionally, the approaches are externally tested on a fully independent test–retest dataset. The repeatability of the methods is compared with those of two majority vote (MV2, MV3) approaches, 41%SUVMAX, and a SUV > 4 segmentation (SUV4). Repeatability is assessed with test–retest coefficients (TRT%) and intraclass correlation coefficient (ICC). An ICC > 0.9 was regarded as representing excellent repeatability.ResultsThe accuracy of the segmentations with the reference segmentation was good (JC median TF: 0.7, CNN: 0.73). Both segmentation approaches outperformed most other conventional segmentation methods in terms of test–retest coefficient (TRT% mean: TF: 13.0%, CNN: 13.9%, MV2: 14.1%, MV3: 28.1%, 41%SUVMAX: 28.1%, SUV4: 18.1%) and ICC (TF: 0.98, MV2: 0.97, CNN: 0.99, MV3: 0.73, SUV4: 0.81, and 41%SUVMAX: 0.68).ConclusionThe semi-automatic AI-based segmentation approaches used in this study provided better repeatability than conventional segmentation approaches. Moreover, both algorithms lead to accurate segmentations for both primary tumors as well as metastasis and are therefore good candidates for PET tumor segmentation.

Residual U-Network for Breast Tumor Segmentation from Magnetic Resonance Images

Breast cancer positions as the most well-known threat and the main source of malignant growth-related morbidity and mortality throughout the world. It is apical of all new cancer incidences analyzed among females. Two features substantially influence the classification accuracy of malignancy and benignity in automated cancer diagnostics. These are the precision of tumor segmentation and appropriateness of extracted attributes required for the diagnosis. In this research, the authors have proposed a ResU-Net (Residual U-Network) model for breast tumor segmentation. The proposed methodology renders augmented, and precise identification of tumor regions and produces accurate breast tumor segmentation in contrast-enhanced MR images. Furthermore, the proposed framework also encompasses the residual network technique, which subsequently enhances the performance and displays the improved training process. Over and above, the performance of ResU-Net has experimentally been analyzed with conventional U-Net, FCN8, FCN32. Algorithm performance is evaluated in the form of dice coefficient and MIoU (Mean Intersection of Union), accuracy, loss, sensitivity, specificity, F1score. Experimental results show that ResU-Net achieved validation accuracy & dice coefficient value of 73.22% & 85.32% respectively on the Rider Breast MRI dataset and outperformed as compared to the other algorithms used in experimentation.

Brain Tumor Segmentation Using Graph Coloring Approach in Magnetic Resonance Images.

It is important to have an accurate and reliable brain tumor segmentation for cancer diagnosis and treatment planning. There are few unsupervised approaches for brain tumor segmentation. In this paper, a new unsupervised approach based on graph coloring for brain tumor segmentation is introduced. In this study, a graph coloring approach is used for brain tumor segmentation. For this aim, each pixel of brain image assumed as a node of graph and difference between brightness of a couple of pixels considered as edge. This method was applied on T1-enhanced magnetic resonance images of low-grade and high-grade patients. Since a rigid graph was needed for graph coloring, edges must be divided into existing or nonexisting edge using a threshold. The value of this threshold has affected the accuracy of image segmentation, so the choice of the optimal threshold was important. The optimal value for this threshold was 0.42 of maximum value of difference of brightness between pixels that caused the 83.62% of correlation accuracy. The results showed that graph coloring approach can be a reliable unsupervised approach for brain tumor segmentation. This approach, as an unsupervised approach, shows better accuracy in comparison with neural networks and neuro-fuzzy networks. However, as a limitation, the accuracy of this approach is dependent on the threshold of edges.

A novel fully automated MRI-based deep-learning method for classification of 1p/19q co-deletion status in brain gliomas.

BackgroundOne of the most important recent discoveries in brain glioma biology has been the identification of the isocitrate dehydrogenase (IDH) mutation and 1p/19q co-deletion status as markers for therapy and prognosis. 1p/19q co-deletion is the defining genomic marker for oligodendrogliomas and confers a better prognosis and treatment response than gliomas without it. Our group has previously developed a highly accurate deep-learning network for determining IDH mutation status using T2-weighted (T2w) MRI only. The purpose of this study was to develop a similar 1p/19q deep-learning classification network.MethodsMultiparametric brain MRI and corresponding genomic information were obtained for 368 subjects from The Cancer Imaging Archive and The Cancer Genome Atlas. 1p/19 co-deletions were present in 130 subjects. Two-hundred and thirty-eight subjects were non-co-deleted. A T2w image-only network (1p/19q-net) was developed to perform 1p/19q co-deletion status classification and simultaneous single-label tumor segmentation using 3D-Dense-UNets. Three-fold cross-validation was performed to generalize the network performance. Receiver operating characteristic analysis was also performed. Dice scores were computed to determine tumor segmentation accuracy.Results1p/19q-net demonstrated a mean cross-validation accuracy of 93.46% across the 3 folds (93.4%, 94.35%, and 92.62%, SD = 0.8) in predicting 1p/19q co-deletion status with a sensitivity and specificity of 0.90 ± 0.003 and 0.95 ± 0.01, respectively and a mean area under the curve of 0.95 ± 0.01. The whole tumor segmentation mean Dice score was 0.80 ± 0.007.ConclusionWe demonstrate high 1p/19q co-deletion classification accuracy using only T2w MR images. This represents an important milestone toward using MRI to predict glioma histology, prognosis, and response to treatment.

MDCC-Net: Multiscale double-channel convolution U-Net framework for colorectal tumor segmentation

PurposeMultiscale feature fusion is a feasible method to improve tumor segmentation accuracy. However, current multiscale networks have two common problems: 1. Some networks only allow feature fusion between encoders and decoders of the same scale. It is obvious that such feature fusion is not sufficient. 2. Some networks have too many dense skip connections and too much nesting between the coding layer and the decoding layer, which causes some features to be lost and means that not enough information will be learned from multiple scales. To overcome these two problems, we propose a multiscale double-channel convolution U-Net (MDCC-Net) framework for colorectal tumor segmentation. MethodsIn the coding layer, we designed a dual-channel separation and convolution module and then added residual connections to perform multiscale feature fusion on the input image and the feature map after dual-channel separation and convolution. By fusing features at different scales in the same coding layer, the network can fully extract the detailed information of the original image and learn more tumor boundary information. ResultsThe segmentation results show that our proposed method has a high accuracy, with a Dice similarity coefficient (DSC) of 83.57%, which is an improvement of 9.59%, 6.42%, and 1.57% compared with nnU-Net, U-Net, and U-Net++, respectively. ConclusionThe experimental results show that our proposed method has good performance in the segmentation of colorectal tumors and is close to the expert level. The proposed method has potential clinical applicability.

Spatial feature fusion convolutional network for liver and liver tumor segmentation from CT images.

The accurate segmentation of liver and liver tumors from CT images can assist radiologists in decision-making and treatment planning. The contours of liver and liver tumors are currently obtained by manual labeling, which is time-consuming and subjective. Computer-aided segmentation methods have been widely used in the segmentation of liver and liver tumors. However, due to the diversity of shape, volume, and image intensity, the segmentation is still a difficult task. In this study, we present a Spatial Feature Fusion Convolutional Network (SFF-Net) to automatically segment liver and liver tumors from CT images. First, we extract side-outputs at each convolutional block in SFF-Net to make full use of multiscale features. Second, skip-connections are added in the down-sampling phase, therefore, the spatial information can be efficiently transferred to later layers. Third, we present feature fusion blocks (FFBs) to merge spatial features and high-level semantic features from early layers and later layers, respectively. Finally, a fully connected 3D conditional random fields (CRFs) is applied to refine the liver and liver tumor segmentation results. We test our method on the MICCAI 2017 Liver Tumor Segmentation (LiTS) challenge dataset. The Dice Global (DG) score, Dice per case (DC) score, Volume Overlap Error (VOE), Average Symmetric Surface Distance (ASSD), and tumor precision score are calculated to evaluate the liver and liver tumor segmentation accuracies. For the liver segmentation, DG is 0.955; DC is 0.937; VOE is 0.106; and ASSD is 3.678. For the tumor segmentation, DG is 0.746; DC is 0.592; VOE is 0.416; ASSD is 1.585 and the tumor precision score is 0.369. The SFF-Net learns more spatial information by adding skip-connections and feature fusion blocks. The experiments validate that our method can accurately segment liver and liver tumors from CT images.

Context aware deep learning for brain tumor segmentation, subtype classification, and survival prediction using radiology images

A brain tumor is an uncontrolled growth of cancerous cells in the brain. Accurate segmentation and classification of tumors are critical for subsequent prognosis and treatment planning. This work proposes context aware deep learning for brain tumor segmentation, subtype classification, and overall survival prediction using structural multimodal magnetic resonance images (mMRI). We first propose a 3D context aware deep learning, that considers uncertainty of tumor location in the radiology mMRI image sub-regions, to obtain tumor segmentation. We then apply a regular 3D convolutional neural network (CNN) on the tumor segments to achieve tumor subtype classification. Finally, we perform survival prediction using a hybrid method of deep learning and machine learning. To evaluate the performance, we apply the proposed methods to the Multimodal Brain Tumor Segmentation Challenge 2019 (BraTS 2019) dataset for tumor segmentation and overall survival prediction, and to the dataset of the Computational Precision Medicine Radiology-Pathology (CPM-RadPath) Challenge on Brain Tumor Classification 2019 for tumor classification. We also perform an extensive performance evaluation based on popular evaluation metrics, such as Dice score coefficient, Hausdorff distance at percentile 95 (HD95), classification accuracy, and mean square error. The results suggest that the proposed method offers robust tumor segmentation and survival prediction, respectively. Furthermore, the tumor classification results in this work is ranked at second place in the testing phase of the 2019 CPM-RadPath global challenge.

Brain Tumor Segmentation Using 3D Generative Adversarial Networks

Brain tumors have high morbidity and may lead to highly lethal cancer. In clinics, accurate segmentation of tumors is the means for diagnosis and determination of subsequent treatment options. Due to the irregularity and blurring of tumor boundaries, accurately segmenting the tumor lesions has received extensive attention in medical image analysis. In view of this situation, this paper proposed a brain tumor segmentation method based on generative adversarial networks (GANs). The GAN architecture consists of a densely connected three-dimensional (3D) U-Net used for segmentation and a classification network for discrimination, both of which use 3D convolutions to fuse multi-dimensional context information. The densely connected 3D U-Net model introduces a dense connection to accelerate network convergence, extracting more detailed information. The adversarial training makes the distribution of segmentation results closer to that of labeled data, which enables the network to segment some unexpected small tumor subregions. Alternately, train two networks and finally achieve a highly accurate classification of each voxel. The experiments conducted on BraTS2017 brain tumor MRI dataset show that the proposed method has higher accuracy in brain tumor segmentation.