Breast Magnetic Resonance Imaging Images Research Articles

Use of Superparamagnetic Iron Oxide Particles in Sentinel Node Biopsy—Evaluation of Its Effects on Breast MRI, Full‐Field Digital Mammography, and Contrast‐Enhanced Mammography

Upon surgical removal of solid tumors in patients with breast cancer, the sentinel lymph nodes are mostly concomitantly removed as this is the first site to which tumor cells metastasize. To identify the sentinel lymph nodes, superparamagnetic iron oxide (SPIO) particles can be employed, a method that has several advantages over the classical tracer with radioisotopes. However, one downside of SPIO particles is the formation of artifacts on magnetic resonance imaging (MRI) of the breast. In a previous study, we investigated if lowering the dose of these particles from 2 to 1 mL would prevent the occurrence of these artifacts. In a study of 14 patients, radiologists do not observe any artifacts on the full‐field digital mammography and contrast‐enhanced mammography images, contrary to breast MRI images, in which case artifacts are observed on all images. Hence, lowering the dose from 2 to 1 mL SPIO particles did not change the outcome of the breast MRI images. In this article, based on these results, we discuss the origins of these artifacts from a mathematical and physical point of view.

Memory-efficient transformer network with feature fusion for breast tumor segmentation and classification task

The analysis of breast cancer using Ultrasounds, Magnetic resonance imaging (MRI), and Mammogram images plays a crucial role in the early detection of breast tumors in women. Nevertheless, accurately segmenting breast tumors remains a challenging task due to the presence of random variations, irregular shapes, and blurred boundaries within tumor regions. Blurred boundaries between breast tumor regions and healthy tissues may also cause higher rates of false-positives or incorrect segmentation results. Traditional convolutional neural networks (CNNs) are constrained by their inability to effectively capture global context information. As a result, they often yield unsatisfactory results in many scenarios. To address this, we have proposed a double encoder block to capture local and global context information related to breast tumor segmentation. In this model, we used a memory-efficient decoder block and a feature fusion block (FFB) to improve feature adaptation and classification capability among adjacent-level with the help of activating channels. Pyramid-dilated fusion (PDF) block based on depth-wise separable convolution is integrated within the transformer to achieve a comprehensive multiscale context. We enabled our model to suppress the irrelevant background noises of breast ultrasound and MRI images. Following this, we conducted comprehensive experiments to evaluate the effectiveness of our proposed Memory-efficient transformer network (MET-Net) for breast tumor segmentation and classification. To ensure reliable results, we employed ultrasounds and MRI datasets within a simulation environment. During the experiment analysis, the MET-Net model demonstrated extraordinary results in the presence of existing state-of-the-art CNNs models in terms of comparative measures. Source code of this work is available at https://github.com/ahmedeqbal/MET-Net.

Improvement of semantic segmentation through transfer learning of multi-class regions with convolutional neural networks on supine and prone breast MRI images

Semantic segmentation of breast and surrounding tissues in supine and prone breast magnetic resonance imaging (MRI) is required for various kinds of computer-assisted diagnoses for surgical applications. Variability of breast shape in supine and prone poses along with various MRI artifacts makes it difficult to determine robust breast and surrounding tissue segmentation. Therefore, we evaluated semantic segmentation with transfer learning of convolutional neural networks to create robust breast segmentation in supine breast MRI without considering supine or prone positions. Total 29 patients with T1-weighted contrast-enhanced images were collected at Asan Medical Center and two types of breast MRI were performed in the prone position and the supine position. The four classes, including lungs and heart, muscles and bones, parenchyma with cancer, and skin and fat, were manually drawn by an expert. Semantic segmentation on breast MRI scans with supine, prone, transferred from prone to supine, and pooled supine and prone MRI were trained and compared using 2D U-Net, 3D U-Net, 2D nnU-Net and 3D nnU-Net. The best performance was 2D models with transfer learning. Our results showed excellent performance and could be used for clinical purposes such as breast registration and computer-aided diagnosis.

Using Breast Tissue Information and Subject-Specific Finite-Element Models to Optimize Breast Compression Parameters for Digital Mammography

Digital mammography has become a first-line diagnostic tool for clinical breast cancer screening due to its high sensitivity and specificity. Mammographic compression force is closely associated with image quality and patient comfort. Therefore, optimizing breast compression parameters is essential. Subjects were recruited for digital mammography and breast magnetic resonance imaging (MRI) within a month. Breast MRI images were used to calculate breast volume and volumetric breast density (VBD) and construct finite element models. Finite element analysis was performed to simulate breast compression. Simulated compressed breast thickness (CBT) was compared with clinical CBT and the relationships between compression force, CBT, breast volume, and VBD were established. Simulated CBT had a good linear correlation with the clinical CBT (R2 = 0.9433) at the clinical compression force. At 10, 12, 14, and 16 daN, the mean simulated CBT of the breast models was 5.67, 5.13, 4.66, and 4.26 cm, respectively. Simulated CBT was positively correlated with breast volume (r > 0.868) and negatively correlated with VBD (r < –0.338). The results of this study provides a subject-specific and evidence-based suggestion of mammographic compression force for radiographers considering image quality and patient comfort.

Breast cancer diagnosis in an early stage using novel deep learning with hybrid optimization technique.

Breast cancer is one of the primary causes of death that is occurred in females around the world. So, the recognition and categorization of initial phase breast cancer are necessary to help the patients to have suitable action. However, mammography images provide very low sensitivity and efficiency while detecting breast cancer. Moreover, Magnetic Resonance Imaging (MRI) provides high sensitivity than mammography for predicting breast cancer. In this research, a novel Back Propagation Boosting Recurrent Wienmed model (BPBRW) with Hybrid Krill Herd African Buffalo Optimization (HKH-ABO) mechanism is developed for detecting breast cancer in an earlier stage using breast MRI images. Initially, the MRI breast images are trained to the system, and an innovative Wienmed filter is established for preprocessing the MRI noisy image content. Moreover, the projected BPBRW with HKH-ABO mechanism categorizes the breast cancer tumor as benign and malignant. Additionally, this model is simulated using Python, and the performance of the current research work is evaluated with prevailing works. Hence, the comparative graph shows that the current research model produces improved accuracy of 99.6% with a 0.12% lower error rate.

Recasted nonlinear complex diffusion method for removal of Rician noise from breast MRI images

The evolution of magnetic resonance imaging (MRI) leads to the study of the internal anatomy of the breast. It maps the physical features along with functional characteristics of selected regions. However, its mapping accuracy is affected by the presence of Rician noise. This noise limits the qualitative and quantitative measures of breast image. This paper proposes recasted nonlinear complex diffusion filter for sharpening the details and removal of Rician noise. It follows maximum likelihood estimation along with optimal parameter selection of complex diffusion where the overall functionality is balanced by regularization parameters. To make recasted nonlinear complex diffusion, the edge threshold constraint “k” of diffusion coefficient is reformed. It is replaced by the standard deviation of the image. It offers a wide range of threshold as variability present in the image with respect to edge. It also provides an automatic selection of “k” instead of user-based value. A series of evaluation has been conducted with respect to different noise ratios further quality improvement of MRI. The qualitative and quantitative assessments of evaluations are tested for the Reference Image Database to Evaluate Therapy Response (RIDER) Breast database. The proposed method is also compared with other existing methods. The quantitative assessment includes the parameters of the full-reference image, human visual system, and no-reference image. It is observed that the proposed method is capable of preserving edges, sharpening the details, and removal of Rician noise.

Weakly Supervised Deep Learning Approach to Breast MRI Assessment

To evaluate a weakly supervised deep learning approach to breast Magnetic Resonance Imaging (MRI) assessment without pixel level segmentation in order to improve the specificity of breast MRI lesion classification. In this IRB approved study, the dataset consisted of 278,685 image slices from 438 patients. The weakly supervised network was based on the Resnet-101 architecture. Training was implemented using the Adam optimizer and a final SoftMax score threshold of 0.5 was used for two class classification (malignant or benign). 278,685 image slices were combined into 92,895 3-channel images. 79,871 (85%) images were used for training and validation while 13,024 (15%) images were separated for testing. Of the testing dataset, 11,498 (88%) were benign and 1531 (12%) were malignant. Model performance was assessed. The weakly supervised network achieved an AUC of 0.92 (SD ± 0.03) in distinguishing malignant from benign images. The model had an accuracy of 94.2% (SD ± 3.4) with a sensitivity and specificity of 74.4% (SD ± 8.5) and 95.3% (SD ± 3.3) respectively. It is feasible to use a weakly supervised deep learning approach to assess breast MRI images without the need for pixel-by-pixel segmentation yielding a high degree of specificity in lesion classification.

Adversarial Data Augmentation on Breast MRI Segmentation

The scarcity of balanced and annotated datasets has been a recurring problem in medical image analysis. Several researchers have tried to fill this gap employing dataset synthesis with adversarial networks (GANs). Breast magnetic resonance imaging (MRI) provides complex, texture-rich medical images, with the same annotation shortage issues, for which, to the best of our knowledge, no previous work tried synthesizing data. Within this context, our work addresses the problem of synthesizing breast MRI images from corresponding annotations and evaluate the impact of this data augmentation strategy on a semantic segmentation task. We explored variations of image-to-image translation using conditional GANs, namely fitting the generator’s architecture with residual blocks and experimenting with cycle consistency approaches. We studied the impact of these changes on visual verisimilarity and how an U-Net segmentation model is affected by the usage of synthetic data. We achieved sufficiently realistic-looking breast MRI images and maintained a stable segmentation score even when completely replacing the dataset with the synthetic set. Our results were promising, especially when concerning to Pix2PixHD and Residual CycleGAN architectures.

Clinical Assessment of Breast Volume: Can 3D Imaging Be the Gold Standard?

Background:Three-dimensional (3D) camera systems are increasingly used for computerized volume calculations. In this study we investigate whether the Vectra XT 3D imaging system is a reliable tool for determination of breast volume in clinical practice. It is compared with the current gold standard in literature, magnetic resonance imaging (MRI), and current clinical practice (plastic surgeon’s clinical estimation).Methods:Breast volumes of 29 patients (53 breasts) were evaluated. 3D images were acquired by Vectra XT 3D imaging system. Pre-existing breast MRI images were collected. Both imaging techniques were used for volume analyses, calculated by two independent investigators. Breast volume estimations were done by plastic surgeons during outpatient consultations. All volume measurements were compared using paired samples t-test, intra-class correlation coefficient, Pearson’s correlation, and Bland–Altman analysis.Results:Two 3D breast volume measurements showed an excellent reliability (intra-class correlation coefficient: 0.991), which was comparable to the reliability of MRI measurements (intra-class correlation coefficient: 0.990). Mean (SD) breast volume measured with 3D breast volume was 454 cm3 (157) and with MRI was 687 cm3 (312). These volumes were significantly different, but a linear association could be found: y(MRI) = 1.58 × (3D) – 40. Three-dimensional breast volume was not significantly different from volume estimation made by plastic surgeons (472 cm3 (69), P = 0.323).Conclusions:The 3D imaging system measures lower volumes for breasts than MRI. However, 3D measurements show a linear association with MRI and have excellent reliability, making them an objective and reproducible measuring method suitable for clinical practice.

Utility and Diagnostic Performance of Automated Breast Ultrasound System in Evaluating Pure Non-Mass Enhancement on Breast Magnetic Resonance Imaging.

ObjectiveTo compare the utility and diagnostic performance of automated breast ultrasound system (ABUS) with that of hand-held ultrasound (HHUS) in evaluating pure non-mass enhancement (NME) lesions on breast magnetic resonance imaging (MRI).Materials and MethodsOne hundred twenty-six consecutive MRI-visible pure NME lesions of 122 patients with breast cancer were assessed from April 2016 to March 2017. Two radiologists reviewed the preoperative breast MRI, ABUS, and HHUS images along with mammography (MG) findings. The NME correlation rate and diagnostic performance of ABUS were compared with that of HHUS, and the imaging features associated with ABUS visibility were analyzed.ResultsAmong 126 pure NME lesions, 100 (79.4%) were malignant and 26 (20.6%) were benign. The overall correlation rate was 87.3% (110/126) in ABUS and 92.9% (117/126) in HHUS. The sensitivity and specificity were 87% and 50% for ABUS and 92% and 42.3% for HHUS, respectively, with no significant differences (p = 0.180 and 0.727, respectively). Malignant NME was more frequently visualized than benign NME lesions on ABUS (93% vs. 65.4%, p = 0.001). Significant factors associated with the visibility of ABUS were the size of NME lesions on MRI (p < 0.001), their distribution pattern (p < 0.001), and microcalcifications on MG (p = 0.027).ConclusionABUS evaluation of pure NME lesions on MRI in patients with breast cancer is a useful technique with high visibility, especially in malignant lesions. The diagnostic performance of ABUS was comparable with that of conventional HHUS in evaluating NME lesions.

Breast MRI and X-ray mammography registration using gradient values.

Breast magnetic resonance imaging (MRI) and X-ray mammography are two image modalities widely used for early detection and diagnosis of breast diseases in women. The combination of these modalities, traditionally done using intensity-based registration algorithms, leads to a more accurate diagnosis and treatment, due to the capability of co-localizing lesions and susceptibles areas between the two image modalities. In this work, we present the first attempt to register breast MRI and X-ray mammographic images using intensity gradients as the similarity measure. Specifically, a patient-specific biomechanical model of the breast, extracted from the MRI image, is used to mimic the mammographic acquisition. The intensity gradients of the glandular tissue are directly projected from the 3D MRI volume to the 2D mammographic space, and two different gradient-based metrics are tested to lead the registration, the normalized cross-correlation of the scalar gradient values and the gradient correlation of the vectoral gradients. We compare these two approaches to an intensity-based algorithm, where the MRI volume is transformed to a synthetic computed tomography (pseudo-CT) image using the partial volume effect obtained by the glandular tissue segmentation performed by means of an Expectation-Maximization algorithm. This allows us to obtain the digitally reconstructed radiographies by a direct intensity projection. The best results are obtained using the scalar gradient approach along with a transversal isotropic material model, obtaining a target registration error (TRE), in millimeters, of 5.65 ± 2.76 for CC- and of 7.83 ± 3.04 for MLO-mammograms, while the TRE is 7.33 ± 3.62 in the 3D MRI. We also evaluate the effect of the glandularity of the breast as well as the landmark position on the TRE, obtaining moderated correlation values (0.65 and 0.77 respectively), concluding that these aspects need to be considered to increase the accuracy in further approaches.

A step-by-step review on patient-specific biomechanical finite element models for breast MRI to x-ray mammography registration.

Breast magnetic resonance imaging (MRI) and x-ray mammography are two image modalities widely used for the early detection and diagnosis of breast diseases in women. The combination of these modalities leads to a more accurate diagnosis and treatment of breast diseases. The aim of this paper is to review the registration between breast MRI and x-ray mammographic images using patient-specific finite element-based biomechanical models. Specifically, a biomechanical model is obtained from the patient's MRI volume and is subsequently used to mimic the mammographic acquisition. Due to the different patient positioning and movement restrictions applied in each image modality, the finite element analysis provides a realistic physics-based approach to perform the breast deformation. In contrast with other reviews, we do not only expose the overall process of compression and registration but we also include main ideas, describe challenges, and provide an overview of the used software in each step of the process. Extracting an accurate description from the MR images and preserving the stability during the finite element analysis require an accurate knowledge about the algorithms used, as well as the software and underlying physics. The wide perspective offered makes the paper suitable not only for expert researchers but also for graduate students and clinicians. We also include several medical applications in the paper, with the aim to fill the gap between the engineering and clinical performance.

Breast-region segmentation in MRI using chest region atlas and SVM

An important step for computerized analysis of breast magnetic resonance imaging (MRI) is segmentation of the breast region. Due to the similar signal intensity of fibroglandular tissue and the chest wall, the segmentation process is difficult for breasts with fibroglandular tissue connected to the chest wall. In order to overcome this challenge, a new framework is presented that relies on a chest region atlas. The proposed method first detects the approximated breast-chest wall boundary using an intensity-based operation. A support vector machine (SVM) then determines the connectivity of fibroglandular tissue to the chest wall by the extracted features from the obtained breast-chest wall boundary. Finally, the obtained breast-chest wall boundary is accurately refined using the geometric shape of the chest region, which is obtained by an atlas-based segmentation method. The proposed method is validated using a dataset of 5964 breast MRI images from 126 women. The Dice similarity coefficient (DSC), total overlap (TO), false negative (FN), and false positive (FP) values are calculated to measure the similarity between automatic and manual segmentation results. Our method achieves DSC, TO, FN, and FP values of 96.46%, 96.41%, 3.59%, and 3.51%, respectively. The results prove the effectiveness of the presented algorithm for breasts with different sizes, shapes, and density patterns.

Un-supervised segmentation and quantisation of malignancy from breast MRI images

In this paper, a magnetic resonance imaging (MRI) based image segmentation technique has been proposed, which uses a magnetic resonance parametric information model for breast tumor segmentation. The methodology has been developed on two dimensional MRI datasets. With the help of the proposed technique, breast tumor tissues can be segmented in 6 - 8 minutes with more precision and reproducibility than manual (supervised) segmentation, which takes more than two hours to segment breast tumor tissues. Thus, the proposed semi-automatic (un-supervised) technique can be applied to analyse MRI images, which improves the procedure for diagnosing breast cancer, and it can also be used to generate two-dimensional view of tumor in case of surgical operations.

A modified Seeded Region Growing algorithm for vessel segmentation in breast MRI images for investigating the nature of potential lesions

The role of Magnetic Resonance Imaging (MRI) as an alternative protocol for screening of breast cancer has been intensively investigated during the past decade. Preliminary research results have indicated that gadolinium-agent administrative MRI scans may reveal the nature of breast lesions by analyzing the contrast-agent's uptake time. In this study, we attempt to deduce the same conclusion, however, from a different perspective by investigating, using image processing, the vascular network of the breast at two different time intervals following the administration of gadolinium. Twenty cases obtained from a 3.0-T MRI system (SIGNA HDx; GE Healthcare) were included in the study. A new modification of the Seeded Region Growing (SRG) algorithm was used to segment vessels from surrounding background. Delineated vessels were investigated by means of their topology, morphology and texture. Results have shown that it is possible to estimate the nature of the lesions with approximately 94.4% accuracy, thus, it may be claimed that the breast vascular network does encodes useful, patterned, information, which can be used for characterizing breast lesions.

Similarity of fibroglandular breast tissue content measured from magnetic resonance and mammographic images and by a mathematical algorithm.

Women with high breast density (BD) have a 4- to 6-fold greater risk for breast cancer than women with low BD. We found that BD can be easily computed from a mathematical algorithm using routine mammographic imaging data or by a curve-fitting algorithm using fat and nonfat suppression magnetic resonance imaging (MRI) data. These BD measures in a strictly defined group of premenopausal women providing both mammographic and breast MRI images were predicted as well by the same set of strong predictor variables as were measures from a published laborious histogram segmentation method and a full field digital mammographic unit in multivariate regression models. We also found that the number of completed pregnancies, C-reactive protein, aspartate aminotransferase, and progesterone were more strongly associated with amounts of glandular tissue than adipose tissue, while fat body mass, alanine aminotransferase, and insulin like growth factor-II appear to be more associated with the amount of breast adipose tissue. Our results show that methods of breast imaging and modalities for estimating the amount of glandular tissue have no effects on the strength of these predictors of BD. Thus, the more convenient mathematical algorithm and the safer MRI protocols may facilitate prospective measurements of BD.

Breast density quantification using magnetic resonance imaging (MRI) with bias field correction: A postmortem study

Quantification of breast density based on three-dimensional breast MRI may provide useful information for the early detection of breast cancer. However, the field inhomogeneity can severely challenge the computerized image segmentation process. In this work, the effect of the bias field in breast density quantification has been investigated with a postmortem study. T1-weighted images of 20 pairs of postmortem breasts were acquired on a 1.5 T breast MRI scanner. Two computer-assisted algorithms were used to quantify the volumetric breast density. First, standard fuzzy c-means (FCM) clustering was used on raw images with the bias field present. Then, the coherent local intensity clustering (CLIC) method estimated and corrected the bias field during the iterative tissue segmentation process. Finally, FCM clustering was performed on the bias-field-corrected images produced by CLIC method. The left-right correlation for breasts in the same pair was studied for both segmentation algorithms to evaluate the precision of the tissue classification. Finally, the breast densities measured with the three methods were compared to the gold standard tissue compositions obtained from chemical analysis. The linear correlation coefficient, Pearson's r, was used to evaluate the two image segmentation algorithms and the effect of bias field. The CLIC method successfully corrected the intensity inhomogeneity induced by the bias field. In left-right comparisons, the CLIC method significantly improved the slope and the correlation coefficient of the linear fitting for the glandular volume estimation. The left-right breast density correlation was also increased from 0.93 to 0.98. When compared with the percent fibroglandular volume (%FGV) from chemical analysis, results after bias field correction from both the CLIC the FCM algorithms showed improved linear correlation. As a result, the Pearson's r increased from 0.86 to 0.92 with the bias field correction. The investigated CLIC method significantly increased the precision and accuracy of breast density quantification using breast MRI images by effectively correcting the bias field. It is expected that a fully automated computerized algorithm for breast density quantification may have great potential in clinical MRI applications.

Projection Based Region of Interest Segmentation in Breast MRI Images

In this paper, a computer aided design auto breast region segmentation system is presented to identify the region of interest (ROI) in magnetic resonance imaging (MRI) images. The system is proposed due to the necessary for performing useful postprocessing on the image for breast cancer research and treatment. Besides, while the ROI is segmented, the image post-processing efficiency of the system is greatly improved. The vertical and horizontal projections algorithms are employed to refine the breast ROI. The methodology has been applied on 55 sets of Digital Image and Communications in Medicine (DICOM) breast MRI datasets images. The experimental results show that the system is able to segment the breast ROI accurately.