Dynamic Contrast-enhanced Breast MRI Research Articles

Automatic semantic segmentation of breast cancer in DCE-MRI using DeepLabV3+ with modified ResNet50

Research on breast cancer segmentation is essential due to its high prevalence as the most common cancer in women and its occurrence in men as well. Breast cancer involves abnormal cell growth in the breast, highlighting the importance of advanced imaging. Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) is an effective technique for this purpose. Deep learning has significantly influenced medical imaging in recent years, especially in accurately segmenting tumors from MRI images. Two techniques have been proposed for breast tumor segmentation: Dilated ResNet50 (RN50D) and Parallel Layers Added ResNet50 (PLA-RN50). RN50D involves altering the dilation factor of the convolution layer within the residual block of ResNet50. PLA-RN50 entails the integration of parallel layers following the final residual block of the ResNet50 architecture. The modified architectures serve as the backbone for the DeepLabV3+ network. The DeepLabV3+ with RN50D or PLA-RN50D is a powerful and effective architecture that integrates deep feature extraction, multiscale spatial information, and precise segmentation to achieve high accuracy in lesion segmentation for breast DCE-MRI images. The proposed technique is tested on a QIN Breast DCE-MRI dataset comprising 233 images sourced from The Cancer Image Archive. The proposed method achieves a dice score of 0.92. The superior segmentation performance of DeepLabV3+ with PLA-RN50, as compared to its counterparts using ResNet18 and ResNet50, highlights the impactful modifications incorporated in PLA-RN50 for optimizing breast tumor segmentation.

Presurgical Upgrade Prediction of DCIS to Invasive Ductal Carcinoma Using Time-dependent Deep Learning Models with DCE MRI.

Purpose To determine whether time-dependent deep learning models can outperform single time point models in predicting preoperative upgrade of ductal carcinoma in situ (DCIS) to invasive malignancy at dynamic contrast-enhanced (DCE) breast MRI without a lesion segmentation prerequisite. Materials and Methods In this exploratory study, 154 cases of biopsy-proven DCIS (25 upgraded at surgery and 129 not upgraded) were selected consecutively from a retrospective cohort of preoperative DCE MRI in women with a mean age of 59 years at time of diagnosis from 2012 to 2022. Binary classification was implemented with convolutional neural network (CNN)-long short-term memory (LSTM) architectures benchmarked against traditional CNNs without manual segmentation of the lesions. Combinatorial performance analysis of ResNet50 versus VGG16-based models was performed with each contrast phase. Binary classification area under the receiver operating characteristic curve (AUC) was reported. Results VGG16-based models consistently provided better holdout test AUCs than did ResNet50 in CNN and CNN-LSTM studies (multiphase test AUC, 0.67 vs 0.59, respectively, for CNN models [P = .04] and 0.73 vs 0.62 for CNN-LSTM models [P = .008]). The time-dependent model (CNN-LSTM) provided a better multiphase test AUC over single time point (CNN) models (0.73 vs 0.67; P = .04). Conclusion Compared with single time point architectures, sequential deep learning algorithms using preoperative DCE MRI improved prediction of DCIS lesions upgraded to invasive malignancy without the need for lesion segmentation. Keywords: MRI, Dynamic Contrast-enhanced, Breast, Convolutional Neural Network (CNN) Supplemental material is available for this article. © RSNA, 2024.

Ultrafast sequence-based prediction model and nomogram to differentiate additional suspicious lesions on preoperative breast MRI.

To investigate whether ultrafast sequence improves the diagnostic performance of conventional dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in differentiating additional suspicious lesions (ASLs) on preoperative breast MRI. A retrospective database search identified 668 consecutive patients who underwent preoperative breast DCE-MRI with ultrafast sequence between June 2020 and July 2021. Among these, 107 ASLs from 98 patients with breast cancer (36 multifocal, 42 multicentric, and 29 contralateral) were identified. Clinical, pathological, conventional MRI findings, and ultrafast sequence-derived parameters were collected. A prediction model that adds ultrafast sequence-derived parameters to clinical, pathological, and conventional MRI findings was developed and validated internally. Decision curve analysis and net reclassification index statistics were performed. A nomogram was constructed. The ultrafast model adding time to peak enhancement, time to enhancement, and maximum slope showed a significantly increased area under the receiver operating characteristic curve compared with the conventional model which includes age, human epidermal growth factor receptor 2 expression of index cancer, size of index cancer, lesion type of index cancer, location of ASL, and size of ASL (0.92 vs. 0.82; p = 0.002). The decision curve analysis showed that the ultrafast model had a higher overall net benefit than the conventional model. The net reclassification index of ultrafast model was 23.3% (p = 0.001). A combination of ultrafast sequence-derived parameters with clinical, pathological, and conventional MRI findings can aid in the differentiation of ASL on preoperative breast MRI. Our prediction model and nomogram that was based on ultrafast sequence-derived parameters could help radiologists differentiate ASLs on preoperative breast MRI. Ultrafast MRI can diminish background parenchymal enhancement and possibly improve diagnostic accuracy for additional suspicious lesions (ASLs). Locationof ASL, larger size of ASL, and highermaximum slope were associated with malignant ASL. The ultrafast model and nomogram can help preoperatively differentiate additional malignancies.

Quantitative analysis from ultrafast dynamic contrast-enhanced breast MRI using population-based versus individual arterial input functions, and comparison with semi-quantitative analysis

PurposeTo evaluate the value of inline quantitative analysis of ultrafast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) using a population-based arterial input function (P-AIF) compared with offline quantitative analysis with an individual AIF (I-AIF) and semi-quantitative analysis for diagnosing breast cancer. MethodsThis prospective study included 99 consecutive patients with 109 lesions (85 malignant and 24 benign). Model-based parameters (Ktrans, kep, and ve) and model-free parameters (washin and washout) were derived from CAIPIRINHA-Dixon-TWIST-VIBE (CDTV) DCE-MRI. Univariate analysis and multivariate logistic regression analysis with forward stepwise covariate selection were performed to identify significant variables. The AUC and F1 score were assessed for semi-quantitative and two quantitative analyses. Resultskep from inline quantitative analysis with P-AIF for diagnosing breast cancer provided an AUC similar to kep from offline quantitative analysis with I-AIF (0.782 vs 0.779, p = 0.954), higher compared to washin from semi-quantitative analysis (0.782 vs 0.630, p = 0.034). Furthermore, the inline quantitative analysis with P-AIF achieved the larger F1 score (0.920) compared with offline quantitative analysis with I-AIF (0.780) and semi-quantitative analysis (0.480). There were no statistically significant differences for kep values between the two quantitative analysis schemes (p = 0.944). ConclusionThe inline quantitative analysis with P-AIF from CDTV in characterizing breast lesions could offer similar diagnostic accuracy to offline quantitative analysis with I-AIF, and higher diagnostic accuracy to semi-quantitative analysis.

Abstract PO1-07-07: Radiomics features for distinguishing true recurrence versus new primary tumor following breast-conserving treatment

Abstract Background Despite modern surgical and irradiation techniques, ipsilateral breast tumor recurrence (IBTR) accounts for 5-15% of all cancer recurrences in women treated with breast conservative treatment. Historically, the methods to define true recurrence (TR) and new primary tumor (NP) for IBTR mainly rely on clinical and pathological criteria, limiting the accuracy of the discerption and causing misclassification. This study aimed to develop a preoperative, noninvasive model for distinguishing IBTR by integrating clinicopathological review with paired dynamic contrast-enhanced breast magnetic resonance imaging (DCE-MRI) at diagnosis and following in-breast recurrence. MethodsWe retrospectively extracted radiomics features from MRI to develop a radiomic cohort of IBTR (n =46), among which all patients underwent paired preoperative DCE-MRI. This radiomic cohort was divided into a training cohort (n =27) and a validation cohort (n =19) with stratified random sampling. Classification of IBTR as TR or NP on the basis of tumor location, histologic subtype, estrogen receptor, and HER2 status was set as the gold standard. The least absolute shrinkage and selection operator (LASSO) regression and logistic regression were utilized to perform radiomics feature selection and model training, respectively. The clinical utility of the model was determined via decision curve analysis (DCA). Results We selected three radiomics features (first-order feature Kurtosis from the first postcontrast phase on IBTR DCE-MRI, GLCM feature Imc2 from the first postcontrast phase on IBTR DCE-MRI, and the delta feature Correlation from the precontrast phase of the change between IBTR and primary tumor DCE-MRI) to develop an IBTR-classification predicting radiomic model, which performed well in the validation cohort (AUC 0.867, 95% confidence interval (CI) 0.694-1). Further investigation for sensitivity (73.3%) and specificity (100%) verified a favorable concordance between the radiomic classification and the conventional standard, with a diagnostic accuracy of 79%. Conclusions: Our study demonstrated the feasibility of the radiomics model in predicting IBTR classification and provided preoperative information about the nature of “recurrence”. This might have important implications in surgical approaches and multidisciplinary care for IBTR. Further efforts are needed to improve the reproducibility of radiomics features and models in multiple centers. Citation Format: Feilin Qu, Guan-Hua Su, Jin-Hui Li, Chao You, Zhi-Ming Shao. Radiomics features for distinguishing true recurrence versus new primary tumor following breast-conserving treatment [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-07-07.

Non-mass enhancement on breast MRI: Clues to a more confident diagnosis

BackgroundNon-mass enhancement (NME) seen on dynamic contrast enhanced breast MRI (DCE-MRI) may be caused by benign, high risk or malignant lesions. Making a clear distinction between these lesions is challenging due to the significant overlap in their imaging appearance. Our study aims to assess the various patterns of distribution, internal enhancement patterns (IEPs) and kinetics of NME using the BI-RADS lexicon fifth edition with histopathologic correlation to aid in making a more confident recommendation regarding clinical management.ResultsSixty-six female patients with NME on DCE-MRI were included. Thirty-four lesions (51.5%) were histopathologically proven to be benign and 32 (48.5%) were malignant. Segmental distribution was the most common pattern and was found in 22 cases (33.3%), 14 of them were malignant with p-value < 0.05. Linear distribution was reported in 14 cases, (21.2%), five of which were malignant, with p-value > 0.05. Thirteen cases (19.7%) had focal distribution, only two of them were malignant with p-value < 0.05. Twelve cases (18.2%) were of regional distribution, seven of which were malignant. Multiregional and diffuse distribution were the least common and were found in 3% and 4.5% of cases respectively. As for the enhancement pattern, 30 cases (45.5%) had heterogeneous enhancement. Nineteen of which were malignant with a p-value < 0.05. Clumped enhancement was found in 24 cases (36.4%); 12 cases were found to be malignant. Nine cases (13.6%) were of homogeneous enhancement, all of them were benign and three cases (4.5%) were of clustered ring enhancement with p-value > 0.05. Restricted diffusion value was detected in 75% of malignant cases with p-value < 0.05. In terms of kinetic curve, the most frequent curve was found to be type II plateau curve (26 cases, 39.4%), 15 cases were of benign pathology and the other 11 cases were proven to be malignant. Followed by type III washout curve which was detected in 25 cases (37.9%), 20 cases were malignant and five cases were benign. And type I persistent curve was found in 15 cases (22.7%); 14 cases were histopathologically proven to be benign, and only one case was of malignant pathology, with a total p-value < 0.05.ConclusionsOur study found that the most common distribution pattern was segmental distribution, being statistically significant with p-value < 0.05, being more common among malignant lesions. As for the enhancement pattern, heterogeneous enhancement was the most common pattern, mainly detected in malignant lesions, with p-value < 0.05. The most common type of kinetic curve was type II curve.

Abstract PO5-07-07: The biological basis of breast MRI background parenchymal enhancement in women with high breast cancer risk

Abstract Introduction: The amount and degree of the fibroglandular tissue (FGT) enhancement on breast MRI, known as background parenchymal enhancement (BPE), is an emerging marker of breast cancer risk. Higher qualitative and quantitative BPE levels have been shown to correlate with both future risk of initial breast cancer diagnosis and breast cancer recurrence after treatment. Furthermore, adjuvant and preventative therapies such as radiation and endocrine therapy have been shown to lower BPE in many women, potentially serving as a marker of treatment efficacy. However, the biological basis of BPE remains unclear and elucidation of this mechanism could identify novel targeted prevention strategies. The goal of this study was to assess the association of BPE levels with markers of inflammation, vascularity, proliferation, and estrogen sensitivity in a cohort of high-risk women. Materials and Methods: In this IRB-approved retrospective study, we identified asymptomatic high-risk women who underwent at least one dynamic contrast-enhanced (DCE) breast MRI within one year of prophylactic bilateral mastectomy. Qualitative BPE assessments were obtained from clinical reports. Quantitative BPE measures were obtained using in-house semi-automated software: whole breast FGT was first segmented on pre-contrast images, percent enhancement (PE) was calculated for each voxel as (S1-S0)/S0 × 100%, where S0 and S1 are pre- and early post-contrast signal intensity, respectively, and quantitative BPE was then defined as the mean PE of the FGT. Additionally, the BPE volume ratio was calculated as the fraction of FGT that had a PE &amp;gt;50%. Histopathology markers of inflammation (COX-2), vascularity (VEGF), proliferation (Ki-67), and estrogen receptor status (ER) were prospectively measured on archived mastectomy specimens of each breast. BPE metrics were correlated with histopathological markers using Spearman’s rank correlation, with and without adjustment by menopausal status of the patient. All tests were based on generalized estimating equations models to account for the non-independence of the two breasts per patient. A p-value &amp;lt; 0.05 was considered significant. Results: We identified 56 women (median age: 39; range: 26-63 years) for this study, of which 39% were post-menopausal (N=22). Distribution of qualitative BPE levels were 41% minimal (N=23), 29% mild (N=16), 18% moderate (N=10), and 12% marked (N=7). Qualitative BPE and quantitative BPE markers were moderately positively correlated (r=0.37-0.54, p&amp;lt; 0.01). Table 1 lists the correlations between BPE metrics and histopathology markers. After adjusting for menopausal status, qualitative BPE was positively associated with ER (r=0.27, p=0.02) and VEGF (r=0.24, p&amp;lt; 0.01; r=0.21, p=0.02 adjusted). Quantitative BPE and BPE volume ratio were both positively correlated with VEGF (p&amp;lt; 0.04) and negatively correlated with COX2 (p&amp;lt; 0.05). No significant associations were observed between BPE and Ki-67. Discussion: Our study suggests that increased BPE on MRI positively correlates with increased vascularity and higher estrogen receptor expression but not inflammation or proliferation within normal breast tissue in high-risk women. These findings, along with established associations between BPE and cancer risk, could be useful to select and assess efficacy of targeted preventative treatments. Table. Spearman rank correlations between BPE metrics and histopathological markers. Citation Format: Pegah Khoshpouri, Anum Kazerouni, Sana Parsian, Daniel Hippe, Olivia Walsh, Lisa Koch, Savannah Partridge, Habib Rahbar. The biological basis of breast MRI background parenchymal enhancement in women with high breast cancer risk [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO5-07-07.

Clinical role of abbreviated and ultrafast MRI in breast imaging.

Current breast cancer screening relies on mammography, digital breast tomosynthesis and breast ultrasound. In select populations, breast MRI is also of great utility. However, multiple factors limit widespread use of breast MRI for screening. Efforts have been made to increase the availability of breast MRI for screening, in large part due to the increased cancer detection rate of breast MRI compared to mammography. Techniques include shortening standard breast MRI protocols with the potential for accommodating MRI screening in a higher number of patients. This review will explain the role of abbreviated breast MRI and ultrafast breast MRI in breast imaging, and detail how these approaches differ from standard dynamic contrast-enhanced breast MRI. In addition, limitations and advantages of these techniques will also be discussed.

Radiomics Analysis of Contrast-Enhanced Breast MRI for Optimized Modelling of Virtual Prognostic Biomarkers in Breast Cancer.

Breast cancer clinical stage and nodal status are the most clinically significant drivers of patient management, in combination with other pathological biomarkers, such as estrogen receptor (ER), progesterone receptor or human epidermal growth factor receptor 2 (HER2) receptor status and tumor grade. Accurate prediction of such parameters can help avoid unnecessary intervention, including unnecessary surgery. The objective was to investigate the role of magnetic resonance imaging (MRI) radiomics for yielding virtual prognostic biomarkers (ER, HER2 expression, tumor grade, molecular subtype, and T-stage). Patients with primary invasive breast cancer who underwent dynamic contrast-enhanced (DCE) breast MRI between July 2013 and July 2016 in a single center were retrospectively reviewed. Age, N-stage, grade, ER and HER2 status, and Ki-67 (%) were recorded. DCE images were segmented and Haralick texture features were extracted. The Bootstrap Lasso feature selection method was used to select a small subset of optimal texture features. Classification of the performance of the final model was assessed with the area under the receiver operating characteristic curve (AUC). Median age of patients (n = 209) was 49 (21-79) years. Sensitivity, specificity, positive predictive value, negative predictive value and accuracy of the model for differentiating N0 vs N1-N3 was: 71%, 79%, 76%, 74%, 75% [AUC = 0.78 (95% confidence interval (CI) 0.72-0.85)], N0-N1 vs N2-N3 was 81%, 59%, 24%, 95%, 62% [AUC = 0.74 (95% CI 0.63-0.85)], distinguishing HER2(+) from HER2(-) was 79%, 48%, 34%, 87%, 56% [AUC = 0.64 (95% CI 0.54-0.73)], high nuclear grade (grade 2-3) vs low grade (grades 1) was 56%, 88%, 96%, 29%, 61% [AUC = 0.71 (95% CI 0.63-0.80)]; and for ER (+) vs ER(-) status the [AUC=0.67 (95% CI 0.59-0.76)]. Radiomics performance in distinguishing triple-negative vs other molecular subtypes was [0.60 (95% CI 0.49-0.71)], and Luminal A [0.66 (95% CI 0.56-0.76)]. Quantitative radiomics using MRI contrast texture shows promise in identifying aggressive high grade, node positive triple negative breast cancer, and correlated well with higher nuclear grades, higher T-stages, and N-positive stages.

Explainable 3D CNN based on baseline breast DCE-MRI to give an early prediction of pathological complete response to neoadjuvant chemotherapy

BackgroundSo far, baseline Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) has played a key role for the application of sophisticated artificial intelligence-based models using Convolutional Neural Networks (CNNs) to extract quantitative imaging information as earlier indicators of pathological Complete Response (pCR) achievement in breast cancer patients treated with neoadjuvant chemotherapy (NAC). However, these models did not exploit the DCE-MRI exams in their full geometry as 3D volume but analysed only few individual slices independently, thus neglecting the depth information. MethodThis study aimed to develop an explainable 3D CNN, which fulfilled the task of pCR prediction before the beginning of NAC, by leveraging the 3D information of post-contrast baseline breast DCE-MRI exams. Specifically, for each patient, the network took in input a 3D sequence containing the tumor region, which was previously automatically identified along the DCE-MRI exam. A visual explanation of the decision-making process of the network was also provided. ResultsTo the best of our knowledge, our proposal is competitive than other models in the field, which made use of imaging data alone, reaching a median AUC value of 81.8%, 95%CI [75.3%; 88.3%], a median accuracy value of 78.7%, 95%CI [74.8%; 82.5%], a median sensitivity value of 69.8%, 95%CI [59.6%; 79.9%] and a median specificity value of 83.3%, 95%CI [82.6%; 84.0%], respectively. The median and CIs were computed according to a 10-fold cross-validation scheme for 5 rounds. ConclusionFinally, this proposal holds high potential to support clinicians on non-invasively early pursuing or changing patient-centric NAC pathways.

MTFN: multi-temporal feature fusing network with co-attention for DCE-MRI synthesis.

Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) plays an important role in the diagnosis and treatment of breast cancer. However, obtaining complete eight temporal images of DCE-MRI requires a long scanning time, which causes patients' discomfort in the scanning process. Therefore, to reduce the time, the multi temporal feature fusing neural network with Co-attention (MTFN) is proposed to generate the eighth temporal images of DCE-MRI, which enables the acquisition of DCE-MRI images without scanning. In order to reduce the time, multi-temporal feature fusion cooperative attention mechanism neural network (MTFN) is proposed to generate the eighth temporal images of DCE-MRI, which enables DCE-MRI image acquisition without scanning. In this paper, we propose multi temporal feature fusing neural network with Co-attention (MTFN) for DCE-MRI Synthesis, in which the Co-attention module can fully fuse the features of the first and third temporal image to obtain the hybrid features. The Co-attention explore long-range dependencies, not just relationships between pixels. Therefore, the hybrid features are more helpful to generate the eighth temporal images. We conduct experiments on the private breast DCE-MRI dataset from hospitals and the multi modal Brain Tumor Segmentation Challenge2018 dataset (BraTs2018). Compared with existing methods, the experimental results of our method show the improvement and our method can generate more realistic images. In the meanwhile, we also use synthetic images to classify the molecular typing of breast cancer that the accuracy on the original eighth time-series images and the generated images are 89.53% and 92.46%, which have been improved by about 3%, and the classification results verify the practicability of the synthetic images. The results of subjective evaluation and objective image quality evaluation indicators show the effectiveness of our method, which can obtain comprehensive and useful information. The improvement of classification accuracy proves that the images generated by our method are practical.

A Physiological-Informed Generative Model for Improving Breast Lesion Classification in Small DCE-MRI Datasets.

In biomedical image processing, Deep Learning (DL) is increasingly exploited in various forms and for diverse purposes. Despite unprecedented results, the huge number of parameters to learn, which necessitates a substantial number of annotated samples, remains a significant challenge. In medical domains, obtaining high-quality labelled datasets is still a challenging task. In recent years, several works have leveraged data augmentation to face this issue, mostly thanks to the introduction of generative models able to produce artificial samples having the same characteristics as the acquired ones. However, we claim that biological principles must be considered in this process, as all medical imaging techniques exploit one or more physical laws or properties directly associated with the physiological characteristics of the tissues under analysis. A notable example is the Dynamic Contrast Enhanced-Magnetic Resonance Imaging (DCE-MRI), in which the kinetic of the contrast agent (CA) highlights both morphological and physiological aspects. In this paper, we introduce a novel generative approach explicitly relying on Physiologically Based Pharmacokinetic (PBPK) modelling and on an Intrinsic Deforming Autoencoder (DAE) to implement a physiologically-aware data augmentation strategy. As a case of study, we consider breast DCE-MRI. In particular, we tested our proposal on two private and one public datasets with different acquisition protocols, demonstrating that the proposed method significantly improves the performance of several DL-based lesion classifiers.

Dynamic contrast-enhanced mammography and breast MRI in the diagnosis of breast cancer and detection of tumor size.

The aim of this study is to evaluate the performance of contrast-enhanced mammography (CEM) and dynamic breast MRI techniques for diagnosing breast lesions, assess the diagnostic accuracy of CEM's using histopathological findings, and compare lesion size measurements obtained from both methods with pathological size. This prospective study included 120 lesions, of which 70 were malignant, in 104 patients who underwent CEM and MRI within a week. Two radiologists independently evaluated the MR and CEM images in separate sessions, using the BI-RADS classification system. Additionally, the maximum sizes of lesion were measured. Diagnostic accuracy parameters and the receiver operating characteristics (ROC) curves were constructed for the two modalities. The correlation between the maximum diameter of breast lesions observed in MRI, CEM, and pathology was analyzed. The overall diagnostic values for MRI were as follows: sensitivity 97.1%, specificity 60%, positive predictive value (PPV) 77.3%, negative predictive value (NPV) 93.8%, and accuracy 81.7%. Correspondingly, for CEM, the sensitivity, accuracy, specificity, PPV, and NPV were 97.14%, 81.67%, 60%, 77.27%, and 93.75%, respectively. The ROC analysis of CEM revealed an area under the curve (AUC) of 0.907 for observer 1 and 0.857 for observer 2, whereas MRI exhibited an AUC of 0.910 for observer 1 and 0.914 for observer 2. Notably, CEM showed the highest correlation with pathological lesion size (r = 0.660 for observer 1 and r = 0.693 for observer 2, p < 0.001 for both). CEM can be used with high sensitivity and similar diagnostic performance comparable to MRI for diagnosing breast cancer. CEM proves to be a successful diagnostic method for precisely determining tumor size.

Lesion detection in women breast’s dynamic contrast-enhanced magnetic resonance imaging using deep learning

Breast cancer is one of the most common cancers in women and the second foremost cause of cancer death in women after lung cancer. Recent technological advances in breast cancer treatment offer hope to millions of women in the world. Segmentation of the breast’s Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) is one of the necessary tasks in the diagnosis and detection of breast cancer. Currently, a popular deep learning model, U-Net is extensively used in biomedical image segmentation. This article aims to advance the state of the art and conduct a more in-depth analysis with a focus on the use of various U-Net models in lesion detection in women’s breast DCE-MRI. In this article, we perform an empirical study of the effectiveness and efficiency of U-Net and its derived deep learning models including ResUNet, Dense UNet, DUNet, Attention U-Net, UNet++, MultiResUNet, RAUNet, Inception U-Net and U-Net GAN for lesion detection in breast DCE-MRI. All the models are applied to the benchmarked 100 Sagittal T2-Weighted fat-suppressed DCE-MRI slices of 20 patients and their performance is compared. Also, a comparative study has been conducted with V-Net, W-Net, and DeepLabV3+. Non-parametric statistical test Wilcoxon Signed Rank Test is used to analyze the significance of the quantitative results. Furthermore, Multi-Criteria Decision Analysis (MCDA) is used to evaluate overall performance focused on accuracy, precision, sensitivity, F1\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$_1$$\\end{document}-score, specificity, Geometric-Mean, DSC, and false-positive rate. The RAUNet segmentation model achieved a high accuracy of 99.76%, sensitivity of 85.04%, precision of 90.21%, and Dice Similarity Coefficient (DSC) of 85.04% whereas ResNet achieved 99.62% accuracy, 62.26% sensitivity, 99.56% precision, and 72.86% DSC. ResUNet is found to be the most effective model based on MCDA. On the other hand, U-Net GAN takes the least computational time to perform the segmentation task. Both quantitative and qualitative results demonstrate that the ResNet model performs better than other models in segmenting the images and lesion detection, though computational time in achieving the objectives varies.

Efficient estimation of pharmacokinetic parameters from breast dynamic contrast-enhanced MRI based on a convolutional neural network for predicting molecular subtypes

Objective. Tracer kinetic models allow for estimating pharmacokinetic (PK) parameters, which are related to pathological characteristics, from breast dynamic contrast-enhanced magnetic resonance imaging. However, existing tracer kinetic models subject to inaccuracy are time-consuming for PK parameters estimation. This study aimed to accurately and efficiently estimate PK parameters for predicting molecular subtypes based on convolutional neural network (CNN). Approach. A CNN integrating global and local features (GL-CNN) was trained using synthetic data where known PK parameters map was used as the ground truth, and subsequently used to directly estimate PK parameters (volume transfer constant K trans and flux rate constant K ep) map. The accuracy assessed by the peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and concordance correlation coefficient (CCC) was compared between the GL-CNN and Tofts-based PK parameters in synthetic data. Radiomic features were calculated from the PK parameters map in 208 breast tumors. A random forest classifier was constructed to predict molecular subtypes using a discovery cohort (n = 144). The diagnostic performance evaluated on a validation cohort (n = 64) using the area under the receiver operating characteristic curve (AUC) was compared between the GL-CNN and Tofts-based PK parameters. Main results. The average PSNR (48.8884), SSIM (0.9995), and CCC (0.9995) between the GL-CNN-based K trans map and ground truth were significantly higher than those between the Tofts-based K trans map and ground truth. The GL-CNN-based K trans obtained significantly better diagnostic performance (AUCs = 0.7658 and 0.8528) than the Tofts-based K trans for luminal B and HER2 tumors. The GL-CNN method accelerated the computation by speed approximately 79 times compared to the Tofts method for the whole breast of all patients. Significance. Our results indicate that the GL-CNN method can be used to accurately and efficiently estimate PK parameters for predicting molecular subtypes.

Generalizable attention U-Net for segmentation of fibroglandular tissue and background parenchymal enhancement in breast DCE-MRI

ObjectivesDevelopment of automated segmentation models enabling standardized volumetric quantification of fibroglandular tissue (FGT) from native volumes and background parenchymal enhancement (BPE) from subtraction volumes of dynamic contrast-enhanced breast MRI. Subsequent assessment of the developed models in the context of FGT and BPE Breast Imaging Reporting and Data System (BI-RADS)-compliant classification.MethodsFor the training and validation of attention U-Net models, data coming from a single 3.0-T scanner was used. For testing, additional data from 1.5-T scanner and data acquired in a different institution with a 3.0-T scanner was utilized. The developed models were used to quantify the amount of FGT and BPE in 80 DCE-MRI examinations, and a correlation between these volumetric measures and the classes assigned by radiologists was performed.ResultsTo assess the model performance using application-relevant metrics, the correlation between the volumes of breast, FGT, and BPE calculated from ground truth masks and predicted masks was checked. Pearson correlation coefficients ranging from 0.963 ± 0.004 to 0.999 ± 0.001 were achieved. The Spearman correlation coefficient for the quantitative and qualitative assessment, i.e., classification by radiologist, of FGT amounted to 0.70 (p < 0.0001), whereas BPE amounted to 0.37 (p = 0.0006).ConclusionsGeneralizable algorithms for FGT and BPE segmentation were developed and tested. Our results suggest that when assessing FGT, it is sufficient to use volumetric measures alone. However, for the evaluation of BPE, additional models considering voxels’ intensity distribution and morphology are required.Critical relevance statementA standardized assessment of FGT density can rely on volumetric measures, whereas in the case of BPE, the volumetric measures constitute, along with voxels’ intensity distribution and morphology, an important factor.Key points• Our work contributes to the standardization of FGT and BPE assessment.• Attention U-Net can reliably segment intricately shaped FGT and BPE structures.• The developed models were robust to domain shift.Graphical

A weakly supervised NMF method to decipher molecular subtype-related dynamic patterns in breast DCE-MR images

Objective. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is an important imaging modality for breast cancer diagnosis. Intratumoral heterogeneity causes a major challenge in the interpretation of breast DCE-MRI. Previous studies have introduced decomposition methods on DCE-MRI to reveal intratumoral heterogeneity by analyzing distinct dynamic patterns within each tumor. However, these methods estimated the dynamic patterns and their corresponding component coefficients in an unsupervised manner, without considering any clinically relevant information. Approach. To decipher molecular subtype-related dynamic patterns, we propose a weakly supervised nonnegative matrix factorization method (WSNMF), which is able to decompose the pixel kinetics of DCE-MRI with image-level subtype labels. The WSNMF is developed based on a discriminant nonnegative matrix factorization (NMF) to utilize coarse-grained subtype information, in which between- and within-class scatters are defined on the mean vector of component coefficients over all pixels in each tumor, rather than directly on the vector of component coefficients of each pixel. Main results. Experiments demonstrated that the dynamic patterns identified by WSNMF had superior performance in distinguishing between luminal A and the other subtype tumors. The classification performance was evaluated using the area under the receiver operating characteristic curve (AUC). WSNMF yielded better classification performance (AUC = 0.822) than other heterogeneity analysis methods, including two partitioning-based methods (KPC with AUC = 0.697 and TTP with AUC = 0.760) and two unsupervised decomposition-based methods (PCA with AUC = 0.774 and NMF with AUC = 0.797). Significance. Our method adds a valuable new perspective into DCE-MRI decomposition-based heterogeneity analysis by taking advantage of intrinsic tumor characteristics to improve the diagnosis of breast cancer.

Molecular subtypes classification of breast cancer in DCE-MRI using deep features

Breast cancer is a major cause of concern on a global scale due to its high incidence rate. It is one of the leading causes of death for women, if left untreated. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is increasingly being used in the evaluation of breast cancer. Prior studies neglected to take into account breast cancer characteristics and features that might be helpful for distinguishing the four molecular subtypes of breast cancer. The use of breast DCE-MRI to identify the molecular subtypes is now the focus of research in breast cancer analysis. It offers breast cancer patients a better chance for an early and effective treatment plan. A manually annotated dataset of 1359 DCE-MRI images was used in this study, with 70% used for training and the remaining for testing. Twelve deep features were extracted from this dataset. The dataset was initially preprocessed through placing the ROIs by a radiologist experienced in breast MRI interpretation, then deep features are extracted using the proposed convolutional neural network (CNN). Finally, the deep features extracted are classified into molecular subtypes of breast cancer using the support vector machine (SVM). The effectiveness of the predictive model was assessed using accuracy and area under curve (AUC) measures. The test was performed on unseen held-out data. The maximum achieved accuracy and AUC were 99.78% and 100% respectively, with substantially a low complexity rate.

Standardization of Breast Dynamic Contrast-enhanced MRI Signal with Application to the Assessment of Background Parenchymal Enhancement Rate.

There is currently no clinically accepted method for quantifying background parenchymal enhancement (BPE), though a sensitive method might allow individualized risk management based on the response to cancer-preventative hormonal therapy. The objective of this pilot study is to demonstrate the utility of linear modeling of standardized dynamic contrast-enhanced MRI (DCEMRI) signal for quantifying changes in BPE rates. On a retrospective database search, 14 women with DCEMRI examinations pre- and post- treatment with tamoxifen were identified. DCEMRI signal was averaged over the parenchymal ROIs to obtain time-dependent signal curves S(t). The gradient echo signal equation was used to standardize scale S(t) to values of (FA) ̃ = 10° and (TR) ̃ = 5.5 ms, and obtain the standardized parameters of DCE-MRI signal S ̃_p (t). Relative signal enhancement (〖RSE〗_p ) ̃ was calculated from S ̃_p, and the reference tissue method for T1 calculation was used to standardize (〖RSE〗_p ) ̃ to gadodiamide as the contrast agent, obtaining (RSE) ̃. (RSE) ̃, in the first 6 minutes, post-contrast administration was fit to a linear model with the slope α ̃_RSE denoting the standardized rate relative BPE. Changes in α ̃_RSE were not found to be significantly correlated with the average duration of tamoxifen treatment, age at the initiation of preventative treatment, or pre-treatment BIRADS breast density category. The average change in α ̃_RSE showed a large effect size of -1.12, significantly higher than -0.86 observed without signal standardization (p < 0.01). Linear modeling of BPE in standardized DCEMRI can provide quantitative measurements of BPE rates, improving sensitivity to changes due to tamoxifen treatment.

BreastDM: A DCE-MRI dataset for breast tumor image segmentation and classification

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) has shown high sensitivity to diagnose breast cancer. However, few computer-aided algorithms focus on employing DCE-MR images for breast cancer diagnosis due to the lack of publicly available DCE-MRI datasets. To address this issue, our work releases a new DCE-MRI dataset called BreastDM for breast tumor segmentation and classification. In particular, a dataset of 232 patients selected with DCE-MR images for benign and malignant cases is established. Each case consists of three types of sequences: pre-contrast, post-contrast, and subtraction sequences. To show the difficulty of breast DCE-MRI tumor image segmentation and classification tasks, benchmarks are achieved by state-of-the-art image segmentation and classification algorithms, including conventional hand-crafted based methods and recently-emerged deep learning-based methods. More importantly, a local–global cross attention fusion network (LG-CAFN) is proposed to further improve the performance of breast tumor images classification. Specifically, LG-CAFN achieved the highest accuracy (88.20%, 83.93%) and AUC value (0.9154,0.8826) in both groups of experiments. Extensive experiments are conducted to present strong baselines based on various typical image segmentation and classification algorithms. Experiment results also demonstrate the superiority of the proposed LG-CAFN to other breast tumor images classification methods. The related dataset and evaluation codes are publicly available at smallboy-code/Breast-cancer-dataset.