Automatic Pectoral Muscle Region Segmentation in Mammograms Using Genetic Algorithm and Morphological Selection.

In computer-aided diagnosis systems for breast mammography, the pectoral muscle region can easily cause a high false positive rate and misdiagnosis due to its similar texture and low contrast with breast parenchyma. Pectoral muscle region segmentation is a crucial pre-processing step to identify lesions, and accurate segmentation in poor-contrast mammograms is still a challenging task. In order to tackle this problem, a novel method is proposed to automatically segment pectoral muscle region in this paper. The proposed method combines genetic algorithm and morphological selection algorithm, incorporating four steps: pre-processing, genetic algorithm, morphological selection, and polynomial curve fitting. For the evaluation results on different databases, the proposed method achieves average FP rate and FN rate of 2.03 and 6.90% (mini MIAS), 1.60 and 4.03% (DDSM), and 2.42 and 13.61% (INBreast), respectively. The results can be comparable performance in various metrics over the state-of-the-art methods.

Relationship between the changes of positivity rate of HER2 expression and the diameter of invasive lesions in early breast cancer and its clinical significance.

BACKGROUND: The expression status of HER2 is an important factor in evaluating the prognosis of breast cancer. We found that the positivity rate of HER2 varied with the diameter of invasive lesions in early breast cancer.We aimed to explore the relationship between the change of HER2 positivity rate of early breast cancer and the diameter of the invasive foci and its clinical significance. METHODS: A total of 217 patients with microinvasive breast cancer and T1a stage breast cancer were enrolled in this study. Machine learning algorithm was used to extract morphological features of invasive lesions in early breast cancer. Using Spearman to analysis the clinicopathological and morphological features related to HER2 expression McNemar test was used to analyze the consistency of HER2 expression between the carcinoma in situ area and the invasive cancer area. RESULTS: In early breast cancer, the diameter of the invasive foci was strongly negatively correlated with the expression status of HER2 (rho=-0.468, p < 0.001). As the diameter of the invasive foci increases, the HER2 positivity rate gradually decreases. When the diameter of the invasive foci> 2 mm, the positivity rate of HER2 was significantly reduced (from 52.6% to 16.1%, p < 0.001), which was close to the positivity rate of HER2 in ordinary invasive breast cancer. Moreover, most of the clinicopathological characteristics of breast cancer with an invasive lesion diameter of 1-2 mm and DCIS-MI were not significantly different (p > 0.05). Among 217 patients, the consistency rate of HER2 expression in carcinoma in situ and invasive foci areas was 97.7% (212/217), and there was no significant difference in HER2 expression status (p = 0.25 and p = 0.50, respectively). CONCLUSIONS: We recommend that breast cancer with an invasive lesion diameter of 1-2 mm should be classified as microinvasive breast cancer. In early breast cancer,the expression status of HER2 in the invasive foci area can refer to the HER2 expression status in carcinoma in situ area. However, it needs further support from a large amount of data and follow-up results.

Spatial-ID: a cell typing method for spatially resolved transcriptomics via transfer learning and spatial embedding.

Spatially resolved transcriptomics provides the opportunity to investigate the gene expression profiles and the spatial context of cells in naive state, but at low transcript detection sensitivity or with limited gene throughput. Comprehensive annotating of cell types in spatially resolved transcriptomics to understand biological processes at the single cell level remains challenging. Here we propose Spatial-ID, a supervision-based cell typing method, that combines the existing knowledge of reference single-cell RNA-seq data and the spatial information of spatially resolved transcriptomics data. We present a series of benchmarking analyses on publicly available spatially resolved transcriptomics datasets, that demonstrate the superiority of Spatial-ID compared with state-of-the-art methods. Besides, we apply Spatial-ID on a self-collected mouse brain hemisphere dataset measured by Stereo-seq, that shows the scalability of Spatial-ID to three-dimensional large field tissues with subcellular spatial resolution.

Multicontext multitask learning networks for mass detection in mammogram.

PURPOSE: In this paper, for the purpose of accurate and efficient mass detection, we propose a new deep learning framework, including two major stages: Suspicious region localization (SRL) and Multicontext Multitask Learning (MCMTL). METHODS: In the first stage, SRL focuses on finding suspicious regions [regions of interest (ROIs)] and extracting multisize patches of these suspicious regions. A set of bounding boxes with different size is used to extract multisize patches, which aim to capture diverse context information. In the second stage, MCMTL networks integrate features from multisize patches of suspicious regions for classification and segmentation simultaneously, where the purpose of this stage is to keep the true positive suspicious regions and to reduce the false positive suspicious regions. RESULTS: According to the experimental results on two public datasets (i.e., CBIS-DDSM and INBreast), our method achieves the overall performance of 0.812 TPR@2.53 FPI and 0.919 TPR@0.12 FPI on test sets, respectively. CONCLUSIONS: Our proposed method suggests comparable performance to the state-of-the-art methods.