Integrated segmentation of brain tumor images for radiotherapy and neurosurgery

Segmentation of brain tumor images is an important task in diagnosis and treatment planning for cancer patients. To achieve this goal with standard clinical acquisition protocols, conventionally, either classification algorithms are applied on multimodal MR images or atlas‐based segmentation is used on a high‐resolution monomodal MR image. These two approaches have been commonly regarded separately. We propose to integrate all the available imaging information into one framework to be able to use the information gained from the tissue classification of the multimodal images to perform a more precise segmentation on the high‐resolution monomodal image by atlas‐based segmentation. For this, we combine a state of the art regularized classification method with an enhanced version of an atlas‐registration approach including multiscale tumor‐growth modeling. This contribution offers the possibility to simultaneously segment subcortical structures in the patient by warping the respective atlas labels, which is important for neurosurgical planning and radiotherapy planning. © 2013 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 23, 59–63, 2013     

Dual‐pathway DenseNets with fully lateral connections for multimodal brain tumor segmentation

Multimodal medical image data provide different structured and functional information, which helps segment brain tumor and gets a reliable and accurate diagnosis. Segmenting brain tumors in magnetic resonance imaging (MRI) is a challenging task because brain tumors can be at any location with changeable shape and size. Existing deep neural networks for brain tumor segmentation use few connections to fuse multilevel information. To make use of multilevel information from multimodal MRIs, we propose dual‐pathway DenseNets with fully lateral connections (DP‐DenseNets), a three‐dimensional (3D) fully convolutional neural network that uses dense connectivity to construct dual‐pathway architecture to multimodal brain tumor segmentation problem. Each two similar imaging modalities have a pathway, for one thing, the bottom‐up pathway with dense connectivity is developed for extracting features; another, the top‐down pathway concatenates the features of the bottom‐up pathway in all layers. Dual pathways with different loss functions and fully lateral connectivity from the bottom‐up pathway to the top‐down pathway provide an abundant combination of different levels of features. Comparing to these fusion schemes such as input‐level fusion and later‐level fusion, this architecture leverages semantics from low to high levels, which is provided by fully lateral connectivity. Our model is evaluated on the dataset from Brain Tumor Segmentation Challenge 2017 (BRATS 2017), and the experiments show that our method achieves better performance than other 3D networks.     

Brain tumor segmentation using a deep Shuffled-YOLO network

Brain tumor is an anomalous proliferation of cells in the brain that can evolve to malignant and benign tumors. Currently, segmentation of brain tumor is the most important surgical and pharmaceutical procedures. However, manually segmenting brain tumors is hard because it is hard to find erratically shaped tumors with only one modality; the MRI modalities are integrated to provide multi-modal images with data that can be utilized to segment tumors. The recent developments in machine learning and the accessibility of medical diagnostic imaging have made it possible to tackle the challenges of segmenting brain tumors with deep neural networks. In this work, a novel Shuffled-YOLO network has been proposed for segmenting brain tumors from multimodal MRI images. Initially, the scalable range-based adaptive bilateral filer (SCRAB) pre-processing technique was used to eliminate the noise artifacts from MRI while preserving the edges. In the segmentation phase, we propose a novel deep Shuffled-YOLO architecture for segmenting the internal tumor structures that include non-enhancing, edema, necrosis, and enhancing tumors from the multi-modality MRI sequences. The experimental fallouts reveal that the proposed Shuffled-YOLO network achieves a better accuracy range of 98.07% for BraTS 2020 and 97.04% for BraTS 2019 with very minimal computational complexity compared to the state-of-the-art models.     

Multiview cluster ensembles for multimodal MRI segmentation

It has been shown that the combination of multimodal magnetic resonance imaging (MRI) images can improve the discrimination of diseased tissue. The fusion of dissimilar imaging data for classification and segmentation purposes, however, is not a trivial task, as there is an inherent difference in information domains, dimensionality, and scales. This work proposed a multiview consensus clustering methodology for the integration of multimodal MR images into a unified segmentation aiming at heterogeneity assessment in tumoral lesions. Using a variety of metrics and distance functions this multiview imaging approach calculated multiple vectorial dissimilarity‐spaces for each MRI modality and it maked use of cluster ensembles to combine a set of unsupervised base segmentations into an unified partition of the voxel‐based data. The methodology was demonstrated with simulated data in application to dynamic contrast enhanced MRI and diffusion tensor imaging MR, for which a manifold learning step was implemented in order to account for the geometric constrains of the high dimensional diffusion information. © 2015 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 25, 56–67, 2015     

Automated brain tumour segmentation techniques— A review

Automatic segmentation of brain tumour is the process of separating abnormal tissues from normal tissues, such as white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). The process of segmentation is still challenging due to the diversity of shape, location, and size of the tumour segmentation. The metabolic process, psychological process, and detailed information of the images, are obtained using positron emission tomography (PET) image, Computer Tomography (CT) image and Magnetic Resonance Image (MRI). Multimodal imaging techniques (such as PET/CT and PET/MRI) that combine the information from many imaging techniques contribute more for accurate brain tumour segmentation. In this article, a comprehensive overview of recent automatic brain tumour segmentation techniques of MRI, PET, CT, and multimodal imaging techniques has been provided. The methods, techniques, their working principle, advantages, their limitations, and their future challenges are discussed in this article. © 2017 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 27, 66–77, 2017     

Fully automatic method for segmentation of brain tumor from multimodal magnetic resonance images using wavelet transformation and clustering technique

Fully automatic brain tumor segmentation is one of the critical tasks in magnetic resonance imaging (MRI) images. This proposed work is aimed to develop an automatic method for brain tumor segmentation process by wavelet transformation and clustering technique. The proposed method using discrete wavelet transform (DWT) for pre‐ and post‐processing, fuzzy c‐means (FCM) for brain tissues segmentation. Initially, MRI images are preprocessed by DWT to sharpen the images and enhance the tumor region. It assists to quicken the FCM clustering technique and classified into four major classes: gray matter (GM), white matter (WM), cerebrospinal fluid (CSF), and background (BG). Then check the abnormality detection using Fuzzy symmetric measure for GM, WM, and CSF classes. Finally, DWT method is applied in segmented abnormal region of images respectively and extracts the tumor portion. The proposed method used 30 multimodal MRI training datasets from BraTS2012 database. Several quantitative measures were calculated and compared with the existing. The proposed method yielded the mean value of similarity index as 0.73 for complete tumor, 0.53 for core tumor, and 0.35 for enhancing tumor. The proposed method gives better results than the existing challenging methods over the publicly available training dataset from MICCAI multimodal brain tumor segmentation challenge and a minimum processing time for tumor segmentation. © 2016 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 26, 305–314, 2016     

In vivo correlation between semi‐quantitative hemodynamic parameters and Ktrans derived from DCE‐MRI of brain tumors

Dynamic contrast‐enhanced (DCE) magnetic resonance imaging (MRI) has become more and more widely applied in cancer diagnosis and treatment follow‐up. Without complicated calculation, a semiquantitative parameter – modified initial area under the curve (mIAUC_c) – was proposed for better correlation with volume transfer constant (K^trans) by computer simulation. In this study, we aim to further investigate the correlation between mIAUC_c and K^trans in clinical. A total of 10 patients with brain tumors participated in this study and images were acquired by using a 3‐Tesla clinical MR scanner. The results showed that mIAUC_c was highly correlated with K^trans with the correlation coefficient of 0.913. Although the ideals of K^trans and mIAUC_c are different, mIAUC_c does the trick for brain tumors evaluations in DCE‐MRI. It reveals that mIAUC_c could be an alternative for physiological condition evaluation in DCE‐MRI. © 2012 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 22, 132–136, 2012     

Multiatlas‐based segmentation of female pelvic organs: Application for computer‐aided diagnosis of cervical cancer

Atlas‐based segmentation is a high level segmentation technique which has become a standard paradigm for exploiting prior knowledge in image segmentation. Recent multiatlas‐based methods have provided greatly accurate segmentations of different parts of the human body by propagating manual delineations from multiple atlases in a data set to a query subject and fusing them. The female pelvic region is known to be of high variability which makes the segmentation task difficult. We propose, here, an approach for the segmentation of magnetic resonance imaging (MRI) called multiatlas‐based segmentation using online machine learning (OML). The proposed approach allows separating regions which may be affected by cervical cancer in a female pelvic MRI. The suggested approach is based on an online learning method for the construction of the dataset of atlases. The experiments demonstrate the higher accuracy of the suggested approach compared to a segmentation technique based on a fixed dataset of atlases and single‐atlas‐based segmentation technique.     

Pediatric brain extraction from T2-weighted MR images using 3D dual frame U-net and human connectome database

Accurate extraction of brain tissues from magnetic resonance (MR) images is important in neuroradiology. However, brain extraction is more difficult for pediatric brains than for adult brains due to several factors including smaller brain sizes and lower tissue contrasts. In this work, we propose a brain extraction technique that utilizes dual frame (DF) 3D U-net deep learning architecture and the human connectome project (HCP) database for multislice 2D pediatric T2-weighted MR images with diseases. To improve segmentation accuracy in small pediatric brains with detailed boundary regions, DF 3D U-net architecture was used. We pretrained networks with the HCP database to compensate for the limited amount of MR images and manual segmentation masks of pediatric patients. For quantitative analysis, we compared the brain extraction results of brain extraction tool, DF, and conventional 3D U-net using the dice similarity coefficient (DSC), intersection of union (IoU), and boundary F1 (BF) scores; each deep learning architecture was evaluated with and without pretraining using the HCP. This study included 10 patients with diseases and all images were acquired using a PROPELLER MR sequence. Pretraining using the HCP database enhanced segmentation performance of the network, and the skip connections in the DF 3D U-net could enhance the contour similarity of segmentation results. Experimental results showed that the proposed method increased the DSC, IoU, and BF scores by 0.8%, 1.6%, and 1.5%, respectively, compared with those of the conventional 3D U-net without pretraining.     

3.0T磁共振弥散加权成像对宫颈癌术前分期评估的临床价值分析

目的:分析3.0T磁共振(MRI)弥散加权成像(DWI)对宫颈癌术前分期评估的临床价值。方法:选取2017年12月-2020年2月间本院收治的50例宫颈癌患者作为研究对象,所有研究对象均经接受手术治疗并切除病灶后送病理科检测,经手术病理检测确诊为宫颈癌。对所有患者术前的3.0T MRI检查图像及结果进行回顾性分析,探究3.0T磁共振(MRI)弥散加权成像(DWI)序列检查对宫颈癌分期等指标的评估意义。结果:以手术病理检测结果为宫颈癌诊断分期的金标准,术前3.0T MRI平扫联合DWI检测诊断宫颈癌的准确率为90.00%;对诊断结果进行分析,3.0T MRI平扫联合DWI检测宫颈癌Ⅰ期、宫颈癌Ⅱ期、宫颈癌Ⅲ期、宫颈癌Ⅳ期的灵敏度为72.73%、93.33%、93.75%、100.00%。结论:3.0T MRI平扫联合DWI检测对宫颈癌术前分期评估意义较大,分期越高,意义越大,对于分期高的患者具有较高的诊断准确率,对患者后续的治疗有显著的积极意义,值得临床应用及推广。     

FPGA implementation of particle swarm optimization based on new fitness function for MRI images segmentation

Magnetic resonance imaging (MRI) is considered as a key part in therapeutic procedures because it clearly defines the aim. It also avoids sensitive organs and it determines the desired paths. This phenomenon requires image processing operations such as segmentation to locate the tumor. Medical image segmentation is still an important topic in the field of brain tumor. In the present article, we propose a Hardware Architecture of segmentation based on a Modified Particle Swarm Optimization (HAMPSO) algorithm for MRI images segmentation. To achieve this, we use the Xilinx System Generator (XSG) to be implemented on a Field Programmable Gate Array (FPGA). This architecture is based on a new variant of objective function. These performances of the proposed method are proved using a set of MRI images and were compared to the Hardware Architecture of segmentation based on Particle Swarm Optimization (HAPSO) in terms of either device utilization, execution time, qualitatively or quantitatively results.     

3D brain magnetic resonance imaging segmentation by using bitplane and adaptive fast marching

Diagnosis using medical images helps doctors detect diseases and treat patients effectively. A system that segments objects automatically from magnetic resonance imaging (MRI) plays an important role when doctors diagnose injuries and brain diseases. This article presents a method for automatic brain, scalp, and skull segmentation from MRI that uses Bitplane and the Adaptive Fast Marching method (FMM). We focus on the segmentation of these tissues, especially the brain, because they are the essential objects, and their segmentation is the first step in the segmentation of other tissues. First, the type of each slice is set based on the shape of the brain, and the head region is segmented by removing its background. Second, the sure region and the unsure region are segmented based on the Bitplane method. Finally, this work proposes an approach for classification that is based on the Adaptive FMM. This approach is evaluated with the BrainWeb and Neurodevelopmental MRI databases and compared with other methods. The Dice Averages for brain, scalp, and skull segmentation are 96%, 80%, and 93%, respectively, on the BrainWeb database and 91%, 67%, and 80%, respectively, on the Neurodevelopmental MRI database.     

Core–Satellite Polydopamine–Gadolinium‐Metallofullerene Nanotheranostics for Multimodal Imaging Guided Combination Cancer Therapy

Integration of magnetic resonance imaging (MRI) and other imaging modalities is promising to furnish complementary information for accurate cancer diagnosis and imaging‐guided therapy. However, most gadolinium (Gd)–chelator MR contrast agents are limited by their relatively low relaxivity and high risk of released‐Gd‐ions‐associated toxicity. Herein, a radionuclide‐⁶⁴Cu‐labeled doxorubicin‐loaded polydopamine (PDA)–gadolinium‐metallofullerene core–satellite nanotheranostic agent (denoted as CDPGM) is developed for MR/photoacoustic (PA)/positron emission tomography (PET) multimodal imaging‐guided combination cancer therapy. In this system, the near‐infrared (NIR)‐absorbing PDA acts as a platform for the assembly of different moieties; Gd₃N@C₈₀, a kind of gadolinium metallofullerene with three Gd ions in one carbon cage, acts as a satellite anchoring on the surface of PDA. The as‐prepared CDPGM NPs show good biocompatibility, strong NIR absorption, high relaxivity (r ₁ = 14.06 mM^?1 s^?1), low risk of release of Gd ions, and NIR‐triggered drug release. In vivo MR/PA/PET multimodal imaging confirms effective tumor accumulation of the CDPGM NPs. Moreover, upon NIR laser irradiation, the tumor is completely eliminated with combined chemo‐photothermal therapy. These results suggest that the CDPGM NPs hold great promise for cancer theranostics.     

Fully automatic brain tumor extraction and tissue segmentation from multimodal MRI brain images

Segmentation of Brain tumor from the magnetic resonance imaging (MRI) of head scans is an essential requirement for clinical diagnosis since manual segmentation is a fatigue and time‐consuming process. Recent computer‐aided diagnosis systems depend on the development of fully automatic methods to overcome these problems. In the present work, a fully automated algorithm is proposed to extract and segment tumor regions from multimodal magnetic resonance imaging (MMMRI) sequences. The algorithm has three phases: (a) tumor portion extraction, (b) tumor substructure segmentation, and (c) 3D postprocessing. First, the algorithm extracts tumor portion using a set of image processing operations from T2, fluid‐attenuated inversion recovery (FLAIR), and T1C images. Here, the proposed modified fuzzy c means clustering algorithm is used for enhancing the tumor portion extraction process. Then, the substructures of tumor such as edema, enhancing tumor, and necrotic regions are segmented from MMMRI sequences, T2, FLAIR, and T1C using region‐wise set operations in Phase II. Finally, 3D visualization of the segmented tumor and volume estimation is performed as postprocessing in Phase III. The proposed work was experimented on BraTS 2013 dataset. The quantitative analysis is performed using William's Index, Dice, sensitivity, specificity, and accuracy and is compared with 19 state‐of‐the‐art methods. The proposed method yields comparable results as 77%, 53%, and 59% of Dice for complete, core, and enhancing tumor regions, respectively.     

Brain Tumor Auto-Segmentation on Multimodal Imaging Modalities Using Deep Neural Network

Due to the difficulties of brain tumor segmentation, this paper proposes a strategy for extracting brain tumors from three-dimensional Magnetic Resonance Image (MRI) and Computed Tomography (CT) scans utilizing 3D U-Net Design and ResNet50, taken after by conventional classification strategies. In this inquire, the ResNet50 picked up accuracy with 98.96%, and the 3D U-Net scored 97.99% among the different methods of deep learning. It is to be mentioned that traditional Convolutional Neural Network (CNN) gives 97.90% accuracy on top of the 3D MRI. In expansion, the image fusion approach combines the multimodal images and makes a fused image to extricate more highlights from the medical images. Other than that, we have identified the loss function by utilizing several dice measurements approach and received Dice Result on top of a specific test case. The average mean score of dice coefficient and soft dice loss for three test cases was 0.0980. At the same time, for two test cases, the sensitivity and specification were recorded to be 0.0211 and 0.5867 using patch level predictions. On the other hand, a software integration pipeline was integrated to deploy the concentrated model into the webserver for accessing it from the software system using the Representational state transfer (REST) API. Eventually, the suggested models were validated through the Area Under the Curve–Receiver Characteristic Operator (AUC–ROC) curve and Confusion Matrix and compared with the existing research articles to understand the underlying problem. Through Comparative Analysis, we have extracted meaningful insights regarding brain tumour segmentation and figured out potential gaps. Nevertheless, the proposed model can be adjustable in daily life and the healthcare domain to identify the infected regions and cancer of the brain through various imaging modalities.     

Enhancing breast pectoral muscle segmentation performance by using skip connections in fully convolutional network

The precise detection and segmentation of pectoral muscle areas in mediolateral oblique (MLO) views is an essential step in the development of a computer-aided diagnosis system to access breast malignant lesions or parenchyma. The goal of this article is to develop a robust and fully automatic algorithm for pectoral muscle segmentation from mammography images. This paper presents an image enhancement approach that improves the quality of mammogram scans and a convolutional neural network-based fully convolutional network architecture enhanced with residual connections for automatic segmentation of the pectoral muscle from the MLO views of a digital mammogram. For this purpose, the model is tested and trained on three different mammogram datasets named MIAS, INBREAST, and DDSM. The ground truth labels of the pectoral muscle were identified under the supervision of experienced radiologists. For training and testing, 10-fold cross-validation was used. The proposed model was compared with baseline U-Net-based architecture. Finally, we used a postprocessing step to find the actual boundary of the pectoral muscle. Our presented architecture generated a mean Intersection over Union (IoU) of 97%, dice similarity coefficient (DSC) of 96% and 98% accuracy on testing data. The proposed architecture for pectoral muscle segmentation from the MLO views of mammogram images with high accuracy and dice score can be quickly merged with the breast tumor segmentation problem.     

Multimodal magnetic resonance imaging for Alzheimer's disease diagnosis using hybrid features extraction and ensemble support vector machines

Magnetic resonance imaging (MRI) is increasingly used in the diagnosis of Alzheimer's disease (AD) in order to identify abnormalities in the brain. Indeed, cortical atrophy, a powerful biomarker for AD, can be detected using structural MRI (sMRI), but it cannot detect impairment in the integrity of the white matter (WM) preceding cortical atrophy. The early detection of these changes is made possible by the novel MRI modality known as diffusion tensor imaging (DTI). In this study, we integrate DTI and sMRI as complementary imaging modalities for the early detection of AD in order to create an effective computer-assisted diagnosis tool. The fused Bag-of-Features (BoF) with Speeded-Up Robust Features (SURF) and modified AlexNet convolutional neural network (CNN) are utilized to extract local and deep features. This is applied to DTI scalar metrics (fractional anisotropy and diffusivity metric) and segmented gray matter images from T1-weighted MRI images. Then, the classification of local unimodal and deep multimodal features is first performed using support vector machine (SVM) classifiers. Then, the majority voting technique is adopted to predict the final decision from the ensemble SVMs. The study is directed toward the classification of AD versus mild cognitive impairment (MCI) versus cognitively normal (CN) subjects. Our proposed method achieved an accuracy of 98.42% and demonstrated the robustness of multimodality imaging fusion.     

Intracellular Construction of Cathepsin B-Guided Gadolinium Nanoparticles for Enhanced T2-Weighted MR Tumor Imaging

Magnetic resonance imaging (MRI) is a superior and noninvasive imaging technique with unlimited tissue penetration depth and superb spatiotemporal resolution, however, using intracellular self-assembly of Gd-containing nanoparticles to enhance the T₂-weighted MR contrast of cancer cells in vivo for precise tumor MRI is rarely reported. The lysosomal cysteine protease cathepsin B (CTSB) is regarded as an attractive biomarker for the early diagnosis of cancers and metastasis. Herein, taking advantage of a biocompatible condensation reaction, a “smart” Gd-based CTSB-responsive small molecular contrast agent VC-Gd-CBT is developed, which can self-assemble into large intracellular Gd-containing nanoparticles by glutathione reduction and CTSB cleavage to enhance the T₂-weighted MR contrast of CTSB-overexpressing MDA-MB-231 cells at 9.4 T. In vivo T₂-weighted MRI studies using MDA-MB-231 murine xenografts show that the T₂-weighted MR contrast change of tumors in VC-Gd-CBT-injected mice is distinctly larger than the mice injected with the commercial agent gadopentetate dimeglumine, or co-injected with CTSB inhibitor and VC-Gd-CBT, indicating that the accumulation of self-assembled Gd-containing nanoparticles at tumor sites effectively enhances the T₂-weighted MR tumor imaging. Hence, this CTSB-targeted small molecule VC-Gd-CBT has the potential to be employed as a T₂ contrast agent for the clinical diagnosis of cancers at an early stage.     

Nebulized Gadolinium‐Based Nanoparticles: A Theranostic Approach for Lung Tumor Imaging and Radiosensitization

Lung cancer is the most common and most fatal cancer worldwide. Thus, improving early diagnosis and therapy is necessary. Previously, gadolinium‐based ultra‐small rigid platforms (USRPs) were developed to serve as multimodal imaging probes and as radiosensitizing agents. In addition, it was demonstrated that USRPs can be detected in the lungs using ultrashort echo‐time magnetic resonance imaging (UTE‐MRI) and fluorescence imaging after intrapulmonary administration in healthy animals. The goal of the present study is to evaluate their theranostic properties in mice with bioluminescent orthotopic lung cancer, after intrapulmonary nebulization or conventional intravenous administration. It is found that lung tumors can be detected non‐invasively using fluorescence tomography or UTE‐MRI after nebulization of USRPs, and this is confirmed by histological analysis of the lung sections. The deposition of USRPs around the tumor nodules is sufficient to generate a radiosensitizing effect when the mice are subjected to a single dose of 10 Gy conventional radiation one day after inhalation (mean survival time of 112 days versus 77 days for irradiated mice without USRPs treatment). No apparent systemic toxicity or induction of inflammation is observed. These results demonstrate the theranostic properties of USRPs for the multimodal detection of lung tumors and improved radiotherapy after nebulization.     

Multiscale ICA for fundus image analysis

Independent component analysis is a technique used for separation of statistically independent sources. It can estimate unknown sources from a mixture of sources without any prior knowledge about them. The sources should be non‐Gaussian and independent with each other. In this work, multiscale ICA is proposed for medical images (fundus images, MRI Images). The data matrix is formed by considering the higher sub‐bands of multiscale decompositions. Performance of multiscale ICA is evaluated and compared with the ICA algorithms using simulated signals and different medical images using Amari performance index and Comon test values. Results show that API and Comon test values are less for multiscale ICA for simulated signals. In case of pathological images, the features are separated correctly by multiscale ICA. Multiscale ICA performs better than simple ICA for separation and detection of independent components from medical images (fundus images), such as blood vessels and artifacts. © 2013 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 23, 327–337, 2013