Attention guided neural ODE network for breast tumor segmentation in medical images

Breast tumor segmentation in ultrasound using distance-adapted fuzzy connectedness, convolutional neural network, and active contour

This study addresses computer-aided breast cancer diagnosis through a hybrid framework for breast tumor segmentation in ultrasound images. The core of the three-stage method is based on the autoencoder convolutional neural network. In the first stage, we prepare a hybrid pseudo-color image through multiple instances of fuzzy connectedness analysis with a novel distance-adapted fuzzy affinity. We produce different weight combinations to determine connectivity maps driven by particular image specifics. After the hybrid image is processed by the deep network, we adjust the segmentation outcome with the Chan-Vese active contour model. We find the idea of incorporating fuzzy connectedness into the input data preparation for deep-learning image analysis our main contribution to the study. The method is trained and validated using a combined dataset of 993 breast ultrasound images from three public collections frequently used in recent studies on breast tumor segmentation. The experiments address essential settings and hyperparameters of the method, e.g., the network architecture, input image size, and active contour setup. The tumor segmentation reaches a median Dice index of 0.86 (mean at 0.79) over the combined database. We refer our results to the most recent state-of-the-art from 2022-2023 using the same datasets, finding our model comparable in segmentation performance.

EfficientUNetViT: Efficient Breast Tumor Segmentation Utilizing UNet Architecture and Pretrained Vision Transformer

This study introduces a sophisticated neural network structure for segmenting breast tumors. It achieves this by combining a pretrained Vision Transformer (ViT) model with a UNet framework. The UNet architecture, commonly employed for biomedical image segmentation, is further enhanced with depthwise separable convolutional blocks to decrease computational complexity and parameter count, resulting in better efficiency and less overfitting. The ViT, renowned for its robust feature extraction capabilities utilizing self-attention processes, efficiently captures the overall context within images, surpassing the performance of conventional convolutional networks. By using a pretrained ViT as the encoder in our UNet model, we take advantage of its extensive feature representations acquired from extensive datasets, resulting in a major enhancement in the model's ability to generalize and train efficiently. The suggested model has exceptional performance in segmenting breast cancers from medical images, highlighting the advantages of integrating transformer-based encoders with efficient UNet topologies. This hybrid methodology emphasizes the capabilities of transformers in the field of medical image processing and establishes a new standard for accuracy and efficiency in activities related to tumor segmentation.

MACCoM: A multiple attention and convolutional cross-mixer framework for detailed 2D biomedical image segmentation

The UNet architecture, which is widely used for biomedical image segmentation, has limitations like blurred feature maps and over- or under-segmented regions. To overcome these limitations, we propose a novel network architecture called MACCoM (Multiple Attention and Convolutional Cross-Mixer) - an end-to-end depthwise encoder-decoder fully convolutional network designed for binary and multi-class biomedical image segmentation built upon deeperUNet. We proposed a multi-scope attention module (MSAM) that allows the model to attend to diverse scale features, preserving fine details and high-level semantic information thus useful at the encoder-decoder connection. As the depth increases, our proposed spatial multi-head attention (SMA) is added to facilitate inter-layer communication and information exchange, enabling the network to effectively capture long-range dependencies and global context. MACCoM is also equipped with a convolutional cross-mixer we proposed to enhance the feature extraction capability of the model. By incorporating these modules, we effectively combine semantically similar features and reduce artifacts during the early stages of training. Experimental results on 4 biomedical datasets crafted from 3 datasets of varying modalities consistently demonstrate that MACCoM outperforms or matches state-of-the-art baselines in the segmentation tasks. With Breast Ultrasound Image (BUSI), MACCoM recorded 99.06 % Jaccard, 77.58 % Dice, and 93.92 % Accuracy, while recording 99.50 %, 98.44 %, and 99.29 % respectively for Jaccard, Dice, and Accuracy on the Chest X-ray (CXR) images used. The Jaccard, Dice, and Accuracy for the High-Resolution Fundus (HRF) images are 95.77 %, 74.35 %, and 95.95 % respectively. The findings here highlight MACCoM's effectiveness in improving segmentation performance and its valuable potential in image analysis.

A Retrospective Analysis of Diagnostic Breast Imaging Outcomes in Young Women at a Tertiary Care Center

(1) Purpose: The purpose of this study was to describe the outcomes of diagnostic breast imaging and the incidence of delayed breast cancer diagnosis in the study population. (2) Methods: We collected the outcome data from diagnostic mammograms and/or breast ultrasounds (USs) performed on women between the ages of 30 and 50 with symptomatic breast clinical presentations between 2018 and 2019. (3) Results: Out of 171 eligible patients, 10 patients (5.8%) had BIRADS 0, 90 patients (52.6%) had benign findings (BIRADS 1 and 2), 41 (24.0%) patients had probable benign findings requiring short-term follow-up (BIRADS 3), while 30 (17.5%) patients had findings suspicious of malignancy (BIRADS 4 and 5). In the BIRADS 3 group, 92.7% had recommended follow-up, while in BIRADS 4 and 5, only 83.3% underwent recommended biopsy at a mean time of 1.7 weeks (range 0-22 wks) from their follow-up scan. Ten (6%) patients were diagnosed with breast cancer, all of whom had BIRADS 4 or 5, with a mean time of breast cancer diagnosis from initial diagnostic imaging of 2.2 weeks (range 1-22 wks). No patients had delayed breast cancer diagnosis in our cohort. (4) Conclusions: We conclude that diagnostic mammograms and breast US are appropriate investigations for clinical breast concerns in women aged 30-50 years.

Spatial and geometric learning for classification of breast tumors from multi-center ultrasound images: a hybrid learning approach

BACKGROUND

Breast cancer is the most common cancer among women, and ultrasound is a usual tool for early screening. Nowadays, deep learning technique is applied as an auxiliary tool to provide the predictive results for doctors to decide whether to make further examinations or treatments. This study aimed to develop a hybrid learning approach for breast ultrasound classification by extracting more potential features from local and multi-center ultrasound data.

METHODS

We proposed a hybrid learning approach to classify the breast tumors into benign and malignant. Three multi-center datasets (BUSI, BUS, OASBUD) were used to pretrain a model by federated learning, then every dataset was fine-tuned at local. The proposed model consisted of a convolutional neural network (CNN) and a graph neural network (GNN), aiming to extract features from images at a spatial level and from graphs at a geometric level. The input images are small-sized and free from pixel-level labels, and the input graphs are generated automatically in an unsupervised manner, which saves the costs of labor and memory space.

RESULTS

The classification AUCROC of our proposed method is 0.911, 0.871 and 0.767 for BUSI, BUS and OASBUD. The balanced accuracy is 87.6%, 85.2% and 61.4% respectively. The results show that our method outperforms conventional methods.

CONCLUSIONS

Our hybrid approach can learn the inter-feature among multi-center data and the intra-feature of local data. It shows potential in aiding doctors for breast tumor classification in ultrasound at an early stage.

Advancing musculoskeletal tumor diagnosis: Automated segmentation and predictive classification using deep learning and radiomics

OBJECTIVES

Musculoskeletal (MSK) tumors, given their high mortality rate and heterogeneity, necessitate precise examination and diagnosis to guide clinical treatment effectively. Magnetic resonance imaging (MRI) is pivotal in detecting MSK tumors, as it offers exceptional image contrast between bone and soft tissue. This study aims to enhance the speed of detection and the diagnostic accuracy of MSK tumors through automated segmentation and grading utilizing MRI.

MATERIALS AND METHODS

The research included 170 patients (mean age, 58 years ±12 (standard deviation), 84 men) with MSK lesions, who underwent MRI scans from April 2021 to May 2023. We proposed a deep learning (DL) segmentation model MSAPN based on multi-scale attention and pixel-level reconstruction, and compared it with existing algorithms. Using MSAPN-segmented lesions to extract their radiomic features for the benign and malignant classification of tumors.

RESULTS

Compared to the most advanced segmentation algorithms, MSAPN demonstrates better performance. The Dice similarity coefficients (DSC) are 0.871 and 0.815 in the testing set and independent validation set, respectively. The radiomics model for classifying benign and malignant lesions achieves an accuracy of 0.890. Moreover, there is no statistically significant difference between the radiomics model based on manual segmentation and MSAPN segmentation.

CONCLUSION

This research contributes to the advancement of MSK tumor diagnosis through automated segmentation and predictive classification. The integration of DL algorithms and radiomics shows promising results, and the visualization analysis of feature maps enhances clinical interpretability.

DMAeEDNet: Dense Multiplicative Attention Enhanced Encoder Decoder Network for Ultrasound-Based Automated Breast Lesion Segmentation

Automated and precise segmentation of breast lesions can facilitate early diagnosis of breast cancer. Recent research studies employ deep learning for automatic segmentation of breast lesions using ultrasound imaging. Numerous studies introduce somewhat complex modifications to the well adapted segmentation network, U-Net for improved segmentation, however, at the expense of increased computational time. Towards this aspect, this study presents a low complex deep learning network, i.e., dense multiplicative attention enhanced encoder decoder network, for effective breast lesion segmentation in the ultrasound images. For the first time in this context, two dense multiplicative attention components are utilized in the encoding layer and the output layer of an encoder-decoder network with depthwise separable convolutions, to selectively enhance the relevant features. A rigorous performance evaluation using two public datasets demonstrates that the proposed network achieves dice coefficients of 0.83 and 0.86 respectively with an average segmentation latency of 19ms. Further, a noise robustness study using an in-clinic recorded dataset without pre-processing indicates that the proposed network achieves dice coefficient of 0.72. Exhaustive comparison with some commonly used networks indicate its adeptness with low time and computational complexity demonstrating feasibility in real time.

A review of mechanistic learning in mathematical oncology

Mechanistic learning refers to the synergistic combination of mechanistic mathematical modeling and data-driven machine or deep learning. This emerging field finds increasing applications in (mathematical) oncology. This review aims to capture the current state of the field and provides a perspective on how mechanistic learning may progress in the oncology domain. We highlight the synergistic potential of mechanistic learning and point out similarities and differences between purely data-driven and mechanistic approaches concerning model complexity, data requirements, outputs generated, and interpretability of the algorithms and their results. Four categories of mechanistic learning (sequential, parallel, extrinsic, intrinsic) of mechanistic learning are presented with specific examples. We discuss a range of techniques including physics-informed neural networks, surrogate model learning, and digital twins. Example applications address complex problems predominantly from the domain of oncology research such as longitudinal tumor response predictions or time-to-event modeling. As the field of mechanistic learning advances, we aim for this review and proposed categorization framework to foster additional collaboration between the data- and knowledge-driven modeling fields. Further collaboration will help address difficult issues in oncology such as limited data availability, requirements of model transparency, and complex input data which are embraced in a mechanistic learning framework.

The Convergence of Radiology and Genomics: Advancing Breast Cancer Diagnosis with Radiogenomics

Despite significant progress in the prevention, screening, diagnosis, prognosis, and therapy of breast cancer (BC), it remains a highly prevalent and life-threatening disease affecting millions worldwide. Molecular subtyping of BC is crucial for predictive and prognostic purposes due to the diverse clinical behaviors observed across various types. The molecular heterogeneity of BC poses uncertainties in its impact on diagnosis, prognosis, and treatment. Numerous studies have highlighted genetic and environmental differences between patients from different geographic regions, emphasizing the need for localized research. International studies have revealed that patients with African heritage are often diagnosed at a more advanced stage and exhibit poorer responses to treatment and lower survival rates. Despite these global findings, there is a dearth of in-depth studies focusing on communities in the African region. Early diagnosis and timely treatment are paramount to improving survival rates. In this context, radiogenomics emerges as a promising field within precision medicine. By associating genetic patterns with image attributes or features, radiogenomics has the potential to significantly improve early detection, prognosis, and diagnosis. It can provide valuable insights into potential treatment options and predict the likelihood of survival, progression, and relapse. Radiogenomics allows for visual features and genetic marker linkage that promises to eliminate the need for biopsy and sequencing. The application of radiogenomics not only contributes to advancing precision oncology and individualized patient treatment but also streamlines clinical workflows. This review aims to delve into the theoretical underpinnings of radiogenomics and explore its practical applications in the diagnosis, management, and treatment of BC and to put radiogenomics on a path towards fully integrated diagnostics.

Benign and malignant classification of breast tumor ultrasound images using conventional radiomics and transfer learning features: A multicenter retrospective study

This study aims to establish an effective benign and malignant classification model for breast tumor ultrasound images by using conventional radiomics and transfer learning features. We collaborated with a local hospital and collected a base dataset (Dataset A) consisting of 1050 cases of single lesion 2D ultrasound images from patients, with a total of 593 benign and 357 malignant tumor cases. The experimental approach comprises three main parts: conventional radiomics, transfer learning, and feature fusion. Furthermore, we assessed the model's generalizability by utilizing multicenter data obtained from Datasets B and C. The results from conventional radiomics indicated that the SVM classifier achieved the highest balanced accuracy of 0.791, while XGBoost obtained the highest AUC of 0.854. For transfer learning, we extracted deep features from ResNet50, Inception-v3, DenseNet121, MNASNet, and MobileNet. Among these models, MNASNet, with 640-dimensional deep features, yielded the optimal performance, with a balanced accuracy of 0.866, AUC of 0.937, sensitivity of 0.819, and specificity of 0.913. In the feature fusion phase, we trained SVM, ExtraTrees, XGBoost, and LightGBM with early fusion features and evaluated them with weighted voting. This approach achieved the highest balanced accuracy of 0.964 and AUC of 0.981. Combining conventional radiomics and transfer learning features demonstrated clear advantages over using individual features for breast tumor ultrasound image classification. This automated diagnostic model can ease patient burden and provide additional diagnostic support to radiologists. The performance of this model encourages future prospective research in this domain.