EMD-DWT based transform domain feature reduction approach for quantitative multi-class classification of breast lesions

P.1.a.003 - PDE7B, NMBR and EPM2A variants and schizophrenia: a case-control and pharmacogenetics study

Magnetic resonance imaging (MRI) is playing an important role in the classification of breast tumors. MRI can be used to obtain multiparametric (mp) information, such as structural, hemodynamic, and physiological information. Quantitative analysis of mp-MRI data has shown potential in improving the accuracy of breast tumor classification. In general, a large set of quantitative and texture features can be generated depending upon the type of methodology used. A suitable combination of selected quantitative and texture features can further improve the accuracy of tumor classification. Machine learning (ML) classifiers based upon features derived from MRI data have shown potential in tumor classification. There is a need for further research studies on selecting an appropriate combination of features and evaluating the performance of different ML classifiers for accurate classification of breast tumors. The objective of the current study was to develop and optimize an ML framework based upon mp-MRI features for the characterization of breast tumors (malignant vs. benign and low- vs. high-grade). This study included the breast mp-MRI data of 60 female patients with histopathology results. A total of 128 features were extracted from the mp-MRI tumor data followed by features selection. Five ML classifiers were evaluated for tumor classification using 10-fold crossvalidation with 10 repetitions. The support vector machine (SVM) classifier based on optimum features selected using a wrapper method with an adaptive boosting (AdaBoost) technique provided the highest sensitivity (0.96 ± 0.03), specificity (0.92 ± 0.09), and accuracy (94% ± 2.91%) in the classification of malignant versus benign tumors. This method also provided the highest sensitivity (0.94 ± 0.07), specificity (0.80 ± 0.05), and accuracy (90% ± 5.48%) in the classification of low- versus high-grade tumors. These findings suggest that the SVM classifier outperformed other ML methods in the binary classification of breast tumors.

Automatic classification of benign and malignant breast ultrasound images is an important and challenging task to improve the efficiency and accuracy of clinical diagnosis of breast tumors and reduce the rate of missed and misdiagnosis. The task often requires a large amount of data to train. However, it is difficult to obtain medical images, which contradicts the large amount of data needed to obtain good diagnostic models for training. In this paper, a novel classification model for the classification of breast tumors is proposed to improve the performance of diagnosis models trained by small datasets. The method integrates three features from medical features extracted from segmented images, features selected from the pre-trained ResNet101 output by principal component analysis (PCA), and texture features. Among the medical features that are used to train the naive Bayes (NB) classifier, and the PCA-selected features are used to train the support vector machine (SVM) classifier. Subsequently, the final results of boosting are obtained by weighting the classifiers. A five-fold cross-validation experiment yields an average accuracy of 89.17%, an average precision of 90.00%, and an average AUC value of 0.95. According to the experimental results, the proposed method has better classification accuracy compared to the accuracy obtained by other models trained on only small datasets. This approach can serve as a reliable second opinion for radiologists, and it can also provide useful advice for junior radiologists who do not have sufficient clinical experience.

To retrospectively evaluate the accuracy of neural network analysis of tumor vascular features at three-dimensional (3D) power Doppler ultrasonography (US) for classification of breast tumors as benign or malignant, with histologic findings as the reference standard. This study was approved by the local ethics committee; informed consent was waived. Three-dimensional power Doppler US images of 221 solid breast masses (110 benign, 111 malignant) were obtained in 221 women (mean age, 46 years; range, 25-71 years). After narrowing down vessels to skeletons with a 3D thinning algorithm, six vascular feature values--vessel-to-volume ratio, number of vascular trees, total vessel length, longest path length, number of bifurcations, and vessel diameter-were computed. A neural network was used to classify tumors by using these features. Independent-samples t test and receiver operating characteristic (ROC) curve analysis were used. Mean values of vessel-to-volume ratio, number of vascular trees, total vessel length, longest path length, number of bifurcations, and vessel diameter were 0.0089 +/- 0.0073 (standard deviation), 26.41 +/- 14.73, 23.02 cm +/- 19.53, 8.44 cm +/- 10.38, 36.31 +/- 37.06, and 0.088 cm +/- 0.021 in malignant tumors, respectively, and 0.0028 +/- 0.0021, 9.69 +/- 6.75, 5.17 cm +/- 4.78, 1.68 cm +/- 1.79, 6.05 +/- 7.55, and 0.064 cm +/- 0.028 in benign tumors, respectively (P < .001 for all six features). Area under ROC curve (A(z)) values of the six features were 0.84, 0.87, 0.87, 0.82, 0.84, and 0.75, respectively. Accuracy, sensitivity, specificity, and positive and negative predictive values were 85% (187 of 221), 83% (96 of 115), 86% (91 of 106), 86% (96 of 111), and 83% (91 of 110), respectively, with A(z) of 0.92 based on all six feature values. Three-dimensional power Doppler US images and neural network analysis of features can aid in classification of breast tumors as benign or malignant.

Differentiating the dielectric properties of the breast between healthy and cancerous tissues is essential to contribute to the early detection studies of breast cancer. There are not many reports based on in vivo and ex vivo dielectric measurements on breasts with tumor. In addition, women show high variability of dielectric parameters between breast tissues (skin, fat, gland, and tumors). This study focuses on the measurement and analysis of breast and tumor tissues in vivo and ex vivo, over a frequency range of 1 to 8 GHz, based on dielectric protocol measurement, using an open coaxial probe. The result led to make a tumor classification (early, middle, and advanced), size and type. These measurements were made in mastectomy and lumpectomy surgeries, considering the same pressure for each group of measurements through a regulated pressure system, with a temperature varies from 22 to 27 degrees, all these measurement and parameters validated by a pathologist. As a result, some differences are observed in distinguishing breasts with tumors from healthy breasts in middle and advanced ages, under a controlled system protocol, that is, increases in real permittivity and conductivity in the breast over the areas where benign and malignant anomalies were located.

Early detection of breast tumor is critical in determining the best possible treatment approach. Due to its superiority compared with mammography in its possibility to detect lesions in dense breast tissue, ultrasound imaging has become an important modality in breast tumor detection and classification. This paper discusses the novel Fourier-based shape feature extraction techniques that provide enhanced classification accuracy for breast tumor in the computer-aided B-mode ultrasound diagnosis system. To demonstrate the effectiveness of the proposed method, experiments were performed using 4,107 ultrasound images with 2,508 malignancy cases. Experimental results show that the breast tumor classification accuracy of the proposed technique was 15.8%, 5.43%, 17.32%, and 13.86% higher than the previous shape features such as number of protuberances, number of depressions, lobulation index, and dissimilarity, respectively.

Primary breast carcinomas constitute a divergent group of neoplasms. The classification of breast tumors has been evolving. Recent advances in molecular genetic techniques have enhanced our understanding of these diseases. Integration of state-of-the-art knowledge from research and practice has resulted in the recognition of novel entities as primary carcinomas of the breast with therapeutic and prognostic significance. To provide an overview of current concepts in the classification and diagnosis of selective salivary-type carcinomas of the breast, focusing on their salient histologic and immunophenotypic characteristics and recent molecular genetic advancements. Data were obtained from review of pertinent English-language literature and firsthand experience of the authors as practicing breast pathologists. The cutting-edge knowledge has led us to further understand a growing number of uncommon types of breast carcinoma that demonstrate morphologic and immunophenotypic resemblance to those more frequently encountered in other organ systems, particularly salivary glands. Some of them also harbor identical molecular genetic alterations to those in their salivary gland counterparts. Yet they typically have diverse prognostic outcomes, thus warranting different clinical management. Accurate diagnosis of these tumors necessitates recognition of salient histologic features and judicious assessment of ancillary studies in the pursuit of precision medicine.

Lymphoepithelioma-like breast carcinoma (LELC) is a rare type of malignant breast tumor that is not included in the current edition of the World Health Organization (WHO) classification of breast tumors. Currently, there are no clearly defined therapeutic strategies, and the general information on breast LELC is based on sporadic clinical cases described in the medical literature. We present a clinical case that describes a 49-year-old woman with a tumor formation in the right breast, histologically verified as LELC, together with a non-palpable, synchronous high-grade invasive ductal carcinoma and ductal carcinoma in situ Grade 2 (DCIS G2) in a different quadrant of the same breast. To our knowledge, this is the first case described in the literature that combines a LELC with a synchronous carcinoma in the same breast.

Nowadays, inspired by the great success of Transformers in Natural Language Processing, many applications of Vision Transformers (ViTs) have been investigated in the field of medical image analysis including breast ultrasound (BUS) image segmentation and classification. In this paper, we propose an efficient multi-task framework to segment and classify tumors in BUS images using hybrid convolutional neural networks (CNNs)-ViTs architecture and Multi-Perceptron (MLP)-Mixer. The proposed method uses a two-encoder architecture with EfficientNetV2 backbone and an adapted ViT encoder to extract tumor regions in BUS images. The self-attention (SA) mechanism in the Transformer encoder allows capturing a wide range of high-level and complex features while the EfficientNetV2 encoder preserves local information in image. To fusion the extracted features, a Channel Attention Fusion (CAF) module is introduced. The CAF module selectively emphasizes important features from both encoders, improving the integration of high-level and local information. The resulting feature maps are reconstructed to obtain the segmentation maps using a decoder. Then, our method classifies the segmented tumor regions into benign and malignant using a simple and efficient classifier based on MLP-Mixer, that is applied for the first time, to the best of our knowledge, for the task of lesion classification in BUS images. Experimental results illustrate the outperformance of our framework compared to recent works for the task of segmentation by producing 83.42% in terms of Dice coefficient as well as for the classification with 86% in terms of accuracy.

HER2 (ERBB2) is an established prognostic and predictive marker for patients with invasive breast cancer and used in HER2-targeted therapy. The clinical and biological significance of expression of HER2 protein in patients with intraductal neoplasms of the breast is not well characterized. In this study, we used our HS-HER2 (High Sensitivity-HER2) which is a CAP/CLIA certified laboratory derived test (LDT) using quantitative immunofluorescence (QIF). We evaluated HER2 expression in DCIS (ductal carcinoma in situ), LCIS (lobular carcinoma in situ) and benign lesions to correlate with their clinicopathologic characteristics. Thirty- eight cases of DCIS, and fourteen cases of LCIS and thirty benign cases were selected from the Yale Pathology department archives and 252 ROIs (regions of interest) were annotated by a board-certified pathologist. DCIS is classified according to the three-tier nuclear grading system: low, intermediate, and high-nuclear grade. LCIS is classified as PLCIS (pleomorphic variant) and CLCIS (classic LCIS) according to the WHO classification of breast tumors. Serial sections of FFPE tumor specimens were used to accurately quantify the HER2 expression by the HS-HER2 assay and the acquisition by QuPath v.04 with Qymia extension. According to the LOD (limit of detection,3 amol/mm2), LOQ (limit of quantification, 9 amol/mm2) and LOL (limit of linearity, 23 amol/mm2) of the optimized HS-HER2 assay as performed in our CLIA lab shows a broad range of HER2 expression in both DCIS and LCIS lesions, ranging from 0.7 amol/mm2 to 111.4 amol/mm2 in DCIS and 3.6 amol/mm2 to 49.9 amol/mm2 in LCIS. High grade DCIS lesions express higher average HER2 levels (111.4 + 0.7 amol/mm2) than combined low and intermediate grade DCIS (77.9 +1.9 amol/mm2) with no significant difference between low/intermediate and high-grade DCIS. CLCIS lesions express higher average HER2 level (49.9 + 3.6 amol/mm2) than PLCIS (35.1 + 5.4amol/mm2) but again this is not statistically significant. HER protein expression is relatively low in benign lesions ranging from 0.8 to 8.4 amol/mm2. Using the HS-HER2 assay, our results show quantitative level HER2 levels in DCIS and LCIS breast lesions. These findings may be important as HER2 targeted therapies that work in gene-unamplified cancer (HER2 conjugated ADCs) gain broader usage in the clinic. Citation Format: Nay Nwe Nyein Chan, Haiying Zhan, Revekka Khaimova, Thazin Aung, Charles Robbins, Patricia Gaule, David Rimm. Quantitative Assessment of HER2 Expression in Intraductal Neoplasms of the Breast [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 PO5-13-11.

Solid papillary carcinoma (SPC) is rare, accounting for less than 1% of all breast cancers. According to the 5th edition of the WHO Classification of Breast Tumors (2019), SPC is divided into invasive (ISPC) and in situ (SPC in situ) subtypes. ISPC is characterized by an irregular or jigsaw pattern lacking myoepithelial cells, while SPC in situ is well-circumscribed, regardless of myoepithelial cell presence. Metastasis is infrequently reported, with most patients involving lymph nodes and arising from ISPC. Metastases from SPC in situ without myoepithelial cells have rarely been reported. We present a 66-year-old woman with a 28 × 13 mm hypoechoic breast mass detected on ultrasound. Microscopic examination revealed well-circumscribed nests of SPC, positive for ER, PR, and synaptophysin, with a Ki-67 proliferation index of 10%. HER2 was negative. Myoepithelial markers, including p63, smooth muscle actin, keratin 5/6, p40, and CD10, were all negative, confirming the absence of a myoepithelial layer. Axillary lymph node dissection revealed metastasis in three of eighteen lymph nodes, with the metastatic foci demonstrating the same well-circumscribed SPC architecture as the primary tumor in the breast. Additionally, a lung biopsy exhibited metastatic SPC with identical histomorphology and immunohistochemical profile to the primary tumor. Although rare, SPC presenting a well-circumscribed morphology, and classified as in situ according to the WHO, should be approached with caution when myoepithelial cells are absent, as the in situ histomorphology does not necessarily exclude the potential for metastasis.

Breast cancer is one of the most common cancers occurring in women and is the main cause of mortality among women aged 20 to 59. Detection of breast cancer in the early stage increases survival rate to 80%, and the rate decreases to around 40% if diagnosed in its later stage. There exists significant need to diagnose breast cancer in its early stage itself among women suspected with the disease. The hematoxylin and eosin (H&E)–stained histopathology tissue samples taken by pathologists from suspected breast areas are viewed with whole slide imaging (WSI) and can be stored as digital histopathology images. Manual analysis of histopathology images is time-consuming and leads to inter- and intraobserver variability. Hence, automated classification methods are in research focus. Breast tumors are classified as benign and malignant tumors, which are non-cancerous and cancerous tumors, respectively. Currently, deep learning through convolutional neural networks is an emerging technology and has proven to be successful for image segmentation and classification applications. These deep CNNs require large labeled datasets, which are lacking in the field of medicine, especially for breast cancer images. Also, high computational power, time, and memory are required if CNN is to be trained from scratch. To overcome this, transfer learning from natural images like ImageNet, comprising of 1.2 million natural images belonging to 1,000 classes, is used. Pretrained models like VGG16, ResNet50, Inception, AlexNet, and DenseNet are employed for transfer learning, and fine-tuning is to be performed for refinement. In the proposed method, four different network architectures, namely, AlexNet, VGG16, Inception v3, and DenseNet121, are used for classification of histopathological images using transfer learning. The top dense layers of these architectures are removed, and custom dense layers are added for classification. Also, the weights of the last two layers are fine-tuned with the histopathology input images dataset. The publicly available breast cancer hisopathology images dataset BreakHis is used for validation of the proposed method. The experimental results prove that transfer learning–based CNN gives better classification accuracy compared to networks learned from scratch. Also, it is inferred from the experimental results that the pretrained model of DenseNet121 gives the highest classification accuracy of 0.9520 compared to the other models and state-of-the-art methods.

The purpose of this study was to determine the impact of strain elastography (SE) on the Breast Imaging Reporting Data System (BI-RADS) classification depending on invasive lobular carcinoma (ILC) lesion size. We performed a retrospective analysis on a sample of 152 female subjects examined between January 2010 - January 2017. SE was performed on all patients and ILC was subsequently diagnosed by surgical or ultrasound-guided biopsy. BI-RADS 1, 2, 6 and Tsukuba BGR cases were omitted. BI-RADS scores were recorded before and after the use of SE. The differences between scores were compared to the ILC tumor size using nonparametric tests and logistic binary regression. We controlled for age, focality, clinical assessment, heredo-collateral antecedents, B-mode and Doppler ultrasound examination. An ROC curve was used to identify the optimal cut-off point for size in relationship to BI-RADS classificationdifference using Youden's index. The histological subtypes of ILC lesions (n=180) included in the sample were luminal A (70%, n=126), luminal B (27.78%, n=50), triple negative (1.67%, n=3) and HER2+ (0.56%, n=1). The BI-RADS classification was higher when SE was performed (Z=- 6.629, p<0.000). The ROC curve identified a cut-off point of 13 mm for size in relationship to BI-RADS classification difference (J=0.670, p<0.000). Small ILC tumors were 17.92% more likely to influence BI-RADS classification (p<0.000). SE offers enhanced BI-RADS classification in small ILC tumors (<13 mm). Sonoelastography brings added value to B-mode breast ultrasound as an adjacent to mammography in breast cancer screening.

HighlightsWhat are the main findings?Time-frequency features extracted using EMD, DWT, and combined EMD+DWT effectively separate earthquake and noise signals.Random Forest classifier with Lasso-selected EMD+DWT features achieved 100% accuracy, specificity, and sensitivity.What are the implication of the main findings?Time-frequency-based feature extraction and selection improve real-time earth-quake detection.The approach provides a robust foundation for operational monitoring and ear-ly-warning systems.Earthquakes are sudden and destructive natural events caused by tectonic movements in the Earth’s crust. Although they cannot be predicted with certainty, rapid and reliable detection is essential to reduce loss of life and property. This study aims to automatically distinguish earthquake and noise signals from real seismic data by analyzing time-frequency features. Signals were scaled using z-score normalization, and extracted with Empirical Mode Decomposition (EMD), Discrete Wavelet Transform (DWT), and combined EMD+DWT methods. Feature selection methods such as Lasso, ReliefF, and Student’s t-test were applied to identify the most discriminative features. Classification was performed with Ensemble Bagged Trees, Decision Trees, Random Forest, k-Nearest Neighbors (k-NN), and Support Vector Machines (SVM). The highest performance was achieved using the RF classifier with the Lasso-based EMD+DWT feature set, reaching 100% accuracy, specificity, and sensitivity. Overall, DWT and EMD+DWT features yielded higher performance than EMD alone. While k-NN and SVM were less effective, tree-based methods achieved superior results. Moreover, Lasso and ReliefF outperformed Student’s t-test. These findings show that time-frequency-based features are crucial for separating earthquake signals from noise and provide a basis for improving real-time detection. The study contributes to the academic literature and holds significant potential for integration into early warning and earthquake monitoring systems.

Ultrasound imaging is an effective screening tool for early diagnosis of breast tumor to decrease the mortality rate. However, differentiation of tumor type based on ultrasound images remains challenging in the field of medical imaging due to the inherent noise and speckles. Thus, obtaining additional information for lesion localization could better support the decision-making by clinicians and improve diagnosis fidelity. Recently, multi-task learning (MTL) methods have been proposed for joint tumor classification and localization, where promising results were demonstrated. However, most MTL methods trained independent network branches for the two different tasks, which might cause conflicts in optimizing features due to their different purposes. In addition, these methods usually require fully-segmented datasets for model training, which poses a heavy burden in data annotation. To overcome these limitations, we propose a novel MTL framework for joint breast tumor classification and localization, motivated by the idea of attention mechanism and weakly-supervised learning strategy. Our method has three major advantages. First, an auxiliary lesion-aware network (LA-Net) with multiple attention modules for lesion localization was designed on top of a pre-defined classification network. In this way, the extracted features for classification were directly augmented by the region of interest (ROI) predicted by the LA-Net, alleviating the potential conflicts between the two tasks. Second, a sequential training strategy with a weakly-supervised learning scheme was employed to train the LA-Net and the classification network iteratively, which allows the model to be trained on the partially-segmented datasets and reduces the burden on data annotation. Third, the LA-Net and classification network design are modularized so that both architectures can be flexibly adjusted for various applications. Results from experiments performed on two breast ultrasound image datasets demonstrated the effectiveness of the proposed method.