Improved PAA algorithm for breast mass detection in mammograms.

Mammography screening is instrumental in the early detection and diagnosis of breast cancer by identifying masses in mammograms. With the rapid development of deep learning, numerous deep learning-based object detection algorithms have been explored for mass detection studies. However, these methods often yield a high false positive rate per image (FPPI) while achieving a high true positive rate (TPR). To maintain a higher TPR while also ensuring lower FPPI, we improved the Probability Anchor Assignment (PAA) algorithm to enhance the detection capability for mammographic characteristics with our previous work. We considered three dimensions: the backbone network, feature fusion module, and dense detection heads. The final experiment showed the effectiveness of the proposed method, and the TPR/FPPI values of the final improved PAA algorithm were 0.96/0.56 on the INbreast datasets. Compared to other methods, our method stands distinguished with its effectiveness in addressing the imbalance between positive and negative classes in cases of single lesion detection.

Breast cancer detection and classification in mammogram using a three-stage deep learning framework based on PAA algorithm.

In recent years, deep learning has been used to develop an automatic breast cancer detection and classification tool to assist doctors. In this paper, we proposed a three-stage deep learning framework based on an anchor-free object detection algorithm, named the Probabilistic Anchor Assignment (PAA) to improve diagnosis performance by automatically detecting breast lesions (i.e., mass and calcification) and further classifying mammograms into benign or malignant. Firstly, a single-stage PAA-based detector roundly finds suspicious breast lesions in mammogram. Secondly, we designed a two-branch ROI detector to further classify and regress these lesions that aim to reduce the number of false positives. Besides, in this stage, we introduced a threshold-adaptive post-processing algorithm with dense breast information. Finally, the benign or malignant lesions would be classified by an ROI classifier which combines local-ROI features and global-image features. In addition, considering the strong correlation between the task of detection head of PAA and the task of whole mammogram classification, we added an image classifier that utilizes the same global-image features to perform image classification. The image classifier and the ROI classifier jointly guide to enhance the feature extraction ability and further improve the performance of classification. We integrated three public datasets of mammograms (CBIS-DDSM, INbreast, MIAS) to train and test our model and compared our framework with recent state-of-the-art methods. The results show that our proposed method can improve the diagnostic efficiency of radiologists by automatically detecting and classifying breast lesions and classifying benign and malignant mammograms.

Improved profiling of low molecular weight serum proteome for gastric carcinoma by data-independent acquisition.

Low molecular weight proteins (LMWPs) in the bloodstream participate in various biological processes and are closely associated with disease status, whereas identification of serous LMWPs remains a great technical challenge due to the wide dynamic range of protein components. In this study, we constructed an integrated LMWP library by combining the LMWPs obtained by three enrichment methods (50% ACN, 20% ACN + 20 mM ABC, and 30 kDa) and their fractions identified by the data-dependent acquisition method. With this newly constructed library, we comprehensively profiled LMWPs in serum using data-independent acquisition and reliably achieved quantitative results for 75% serous LMWPs. When applying this strategy to quantify LMWPs in human serum samples, we could identify 405 proteins on average per sample, of which 136 proteins were with a MW less than 30 kDa and 293 proteins were with a MW less than 65 kDa. Of note, pre- and post-operative gastric carcinoma (GC) patients showed differentially expressed serous LWMPs, which was also different from the pattern of LWMP expression in healthy controls. In conclusion, our results showed that LMWPs could efficiently distinguish GC patients from healthy controls as well as between pre- and post-operative statuses, and more importantly, our newly developed LMWP profiling platform could be used to discover candidate LMWP biomarkers for disease diagnosis and status monitoring.

Nobiletin promotes the pyroptosis of breast cancer via regulation of miR-200b/JAZF1 axis.

Nobiletin is a polymethoxylated flavone present in citrus fruits, which has been reported to have inhibitory effects on tumorigenesis of cancers. However, the biological function of nobiletin in breast cancer (BC) is largely unknown. To investigate the effect of nobiletin on growth of BC cells, the cell viability of BC was measured by MTT assay. In addition, gene and protein expressions were detected by qRT-PCR and western blot, respectively. The apoptosis and pyroptosis of BC cells were tested by flow cytometry. Finally, the correlation between miR-200b and JAZF1 was detected by dual luciferase report. The data indicated that nobiletin inhibited the proliferation of BC cells in a dose-dependent manner. Moreover, miR-200b mimics-induced pyroptosis of BC cells was further increased by nobiletin. Meanwhile, JAZF1 was found to be the target of miR-200b. Moreover, nobiletin induced apoptosis and pyroptosis of BC cells via miR-200b/JAZF1/NF-κB axis. In conclusion, nobiletin inhibited the tumorigenesis of BC via regulation of miR-200b/JAZF1 axis. Thus, nobiletin might serve as a new agent for the treatment of BC.

Automated mammographic mass detection using deformable convolution and multiscale features.

Designing computer-assisted diagnosis (CAD) systems that can precisely identify lesions from mammography images would be useful for clinicians. Considering the morphological variation in breast cancer, it is necessary to extract robust features from the mammogram. Here, we propose a mass detection CAD system that is based on Faster R-CNN. First, we applied a novel convolution network in the backbone of Faster R-CNN, namely deformable convolution network (DCN), which improves the detection of lesions with varying shapes and sizes. Second, the original Faster R-CNN uses the output of the last layer of the backbone as a single-scale feature map. To facilitate the detection of small lesions, we used a multiscale feature pyramid network of multiple cross-scale connections between the different output layers of the backbone, called the neural architecture search-feature pyramid network (NAS-FPN). Thus, we were able to integrate the best features into the model. We then evaluated our method by using the datasets the Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM) and INbreast, respectively. Our method yielded a true positive rate of 0.9345 at 2.2805 false positive per image on CBIS-DDSM and a true positive rate of 0.9554 at 0.3829 false positive per image on INbreast. Graphical abstract.

Scinderin-knockdown inhibits proliferation and promotes apoptosis in human breast carcinoma cells.

Previous studies have reported that scinderin (SCIN) affects multiple cellular processes, including proliferation, migration and differentiation in cancer. However, the specific role of SCIN in breast cancer (BC) cells is unknown. Immunohistochemistry was used to investigate SCIN expression in 46 BC and 21 mammary fibroadenoma or fibroadenomatoid hyperplasia tissue samples. SCIN expression was ablated in MDA-MB-231 and T-47D cells using lentivirus-mediated small interfering RNA technology. Cell proliferation was tested using Celigo and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays. Cell apoptosis was analyzed by measuring Caspase 3/7 activity and annexin-V staining. The results of the present study demonstrated that SCIN expression was elevated in BC tissues compared with mammary fibroadenoma or fibroadenomatoid hyperplasia tissues. Specifically, higher SCIN expression was observed in Ki-67-positive BC tissues (78.6%) compared with Ki-67-negative BC tissues. Furthermore, knockdown of SCIN expression in the BC cell lines significantly suppressed cell proliferation and induced apoptosis. The data presented in the present study indicate that SCIN serves an important role in the development of breast cancer.

Clinical and genetic characterization of hereditary breast cancer in a Chinese population.

BACKGROUND: Breast cancer develops as a result of multiple gene mutations in combination with environmental risk factors. Causative variants in genes such as BRCA1 and/or BRCA2 have been shown to account for hereditary nature of certain breast cancers. However,other genes, such as ATM, PALB2, BRIP1, CHEK, BARD1, while lower in frequency, may also increase breast cancer risk. There are few studies examining the role of these causative variants. Our study aimed to examine the clinical and genetic characterization of hereditary breast cancer in a Chinese population. METHODS: We tested a panel of 27 genes implicated in breast cancer risk in 240 participants using Next-Generation Sequencing. The prevalence of genetic causative variants was determined and the association between causative variants and clinico-pathological characteristics was analyzed. RESULTS: Causative variant rate was 19.2% in the breast cancer (case) group and 12.5% in the high-risk group. In the case group 2.5% of patients carried BRCA1 causative variant, 7.5% BRCA2 variants, 1.7% patients had MUTYH, CHEK or PALB2 variants, and 0.8% patients carried ATM, BARD1, NBN, RAD51C or TP53 variants. In the high-risk group 5.8% women carried MUTYH causative variants, 2.5% had causative variants in ATM, 1.7% patients had variants in BRCA2 and 0.8% in BARD1, BRIP1 or CDH1. There was no significant difference in the presence of causative variants among clinical stages of breast cancer, tumor size and lymph nodes status. However, eight of the 12 BRCA1/2 causative variants were found in the TNBC group. CONCLUSIONS: We found increased genetic causative variants in the familial breast cancer group and in high-risk women with a family history of breast cancer. However, the variant MUTYH c.892-2A > G may not be directly associated with hereditary breast carcinoma.

A novel loss-of-function heterozygous BRCA2 c.8946_8947delAG mutation found in a Chinese woman with family history of breast cancer.

INTRODUCTION: Breast cancer is the most frequent female malignancy worldwide. Among them, some cases have hereditary susceptibility in two leading genes, BRCA1 and BRCA2. Heterozygous germ line mutations in them are related with increased risk of breast, ovarian and other cancer, following autosomal dominant inheritance mode. METHODS AND RESULTS: For purpose of early finding, early diagnosis and early treatment, mutation detecting of BRCA1/2 genes was performed in unselected 300 breast or ovarian patients and unaffected women using next-generation sequencing and then confirmed by Sanger sequencing. A non-previously reported heterozygous mutation c.8946_8947delAG (p.D2983FfsX34) of BRCA2 gene was identified in an unaffected Chinese woman with family history of breast cancer (her breast cancer mother, also carrying this mutation). The BRCA2-truncated protein resulted from the frame shift mutation was found to lose two putative nuclear localization signals and a Rad51-binding motif in the extreme C-terminal region by bioinformatic prediction. And then in vitro experiments showed that nearly all the mutant protein was unable to translocate to the nucleus to perform DNA repair activity. This novel mutant BRCA2 protein is dysfunction. CONCLUSIONS: We classify the mutation into disease causing and conclude that it is the risk factor for breast cancer in this family. So, conducting the same mutation test and providing genetic counseling for this family is practically meaningful and significant. Meanwhile, the identification of this new mutation enriches the Breast Cancer Information Core database, especially in China.

Composition and Interface Engineering of Alloyed MoS2 x Se2(1- x ) Nanotubes for Enhanced Hydrogen Evolution Reaction Activity.

Hierarchical MoS2 x Se2(1- x ) nanotubes assembled from several-layered nanosheets featuring tunable chalcogen compositions, expanded interlayer spacing and carbon modification, are synthesized for enhanced electrocatalytic hydrogen evolution reaction (HER). The chalcogen compositions of the MoS2 x Se2(1- x ) nanotubes are controllable by adjusting the selenization temperature and duration while the expanded (002) interlayer spacing varies from 0.98 to 0.68 nm. It is found that the MoS2 x Se2(1- x ) (x = 0.54) nanotubes with expanded interlayer spacing of 0.98 nm exhibit the highest electrocatalytic HER activity with a low onset potential of 101 mV and a Tafel slope of 55 mV dec(-1) . The improved electrocatalytic performance is attributed to the chalcogen composition tuning and the interlayer distance expansion to achieve benefitting hydrogen adsorption energy. The present work suggests a potential way to design advanced HER electrocatalysts through modulating their compositions and interlayer distances.

IMP3 Predicts Invasion and Prognosis in Human Lung Adenocarcinoma.

PURPOSE: Insulin-like growth factor II mRNA-binding protein 3 (IMP3) is an oncofetal protein associated with several aggressive and advanced cancers. Whether IMP3 can predict invasion, and prognosis in patients with human lung adenocarcinoma (LAC) remains unclear. METHODS: Ninety-five LAC and 75 non-tumor lung tissue samples were included in a tissue microarray. IMP3 expression was assessed by immunohistochemical examination. Correlation between IMP3 expression levels, clinicopathological characteristics, and overall prognosis was evaluated. In a separate in vitro study, RNA interference method was applied for knockdown of IMP3 gene in human LAC cell lines. Invasive potential of LAC cells was then evaluated by transwell migration assay. RESULTS: IMP3 immunoreactivity was observed in 39 out of 95 (41.1 %) LAC patients, but not in non-tumor lung tissues. IMP3 expression levels were closely associated with histological grade (P = 0.037), TNM stage (P = 0.034), and lymph node metastasis (P = 0.011). Patients presenting with positive IMP3 expression (P = 0.000), an advanced TNM stage (P = 0.000), and lymph node metastasis (P = 0.001) had a worse overall survival, compared to those lacking these characteristics. Both IMP3 expression (hazard ratio [HR], 2.310; 95 % confidence interval [CI] 1.192-4.476; P = 0.013) and TNM stage (HR 2.338; 95 % CI 1.393-3.925; P = 0.001) were independent predictors of poor prognosis. The invasive potential of LAC cells was significantly inhibited by IMP3 knockdown. CONCLUSION: IMP3 appears to play an important role in tumor invasion in patients with LAC and may serve as a useful prognostic biomarker in these patients.

SEPTIN2 and STATHMIN Regulate CD99-Mediated Cellular Differentiation in Hodgkin's Lymphoma.

Hodgkin's lymphoma (HL) is a lymphoid neoplasm characterized by Hodgkin's and Reed-Sternberg (H/RS) cells, which is regulated by CD99. We previously reported that CD99 downregulation led to the transformation of murine B lymphoma cells (A20) into cells with an H/RS phenotype, while CD99 upregulation induced differentiation of classical Hodgkin's lymphoma (cHL) cells (L428) into terminal B-cells. However, the molecular mechanism remains unclear. In this study, using fluorescence two-dimensional differential in-gel electrophoresis and matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS), we have analyzed the alteration of protein expression following CD99 upregulation in L428 cells as well as downregulation of mouse CD99 antigen-like 2 (mCD99L2) in A20 cells. Bioinformatics analysis showed that SEPTIN2 and STATHMIN, which are cytoskeleton proteins, were significantly differentially expressed, and chosen for further validation and functional analysis. Differential expression of SEPTIN2 was found in both models and was inversely correlated with CD99 expression. STATHMIN was identified in the A20 cell line model and its expression was positively correlated with that of CD99. Importantly, silencing of SEPTIN2 with siRNA substantially altered the cellular cytoskeleton in L428 cells. The downregulation of STATHMIN by siRNA promoted the differentiation of H/RS cells toward terminal B-cells. These results suggest that SEPTIN2-mediated cytoskeletal rearrangement and STATHMIN-mediated differentiation may contribute to changes in cell morphology and differentiation of H/RS cells with CD99 upregulation in HL.

A fine decision tree consisted of CK5/6, IMP3 and TTF1 for cytological diagnosis among reactive mesothelial cells, metastatic adenocarcinoma of lung and non-lung origin in pleural effusion.

The utility of combination with CK5/6, IMP3 and TTF1 to differentiate among reactive mesothelial cells (RMs), metastatic adenocarcinoma of lung (LAC) and non-lung (NLAC) origin was investigated by using immunocytochemistry (ICC) and conventional PCR (C-PCR) in pleural effusion. A total of 108 cell blocks (32 RMs, 51 LAC and 25 NLAC were evaluated by ICC for CK5/6, IMP3 and TTF1 protein expression. In addition, we further performed C-PCR for amplification of CK5/6, IMP3 and TTF1 DNA from 28 specimens (9 MAC and 7 RMs, 6 LAC and 6 NLAC) for molecular diagnosis. CK5/6 staining was observed in the majority of reactive specimens (78.1%) and was rare in adenocarcinoma cells (14.5%), whereas the opposite was true for IMP3 and TTF1. We found a high frequency of TTF1 positivity (76.5%) in LAC, but not in NLAC (4.0%); while there was no significant difference of IMP3 expression in LAC (88.2%) and NLAC (88.0%). The 487 bp DNA fragments of IMP3 was expected to be amplified in 6/9 of adenocarcinoma cases showed negative in ICC; and the 394 bp DNA fragments of CK5/6 was also expected to be amplified in 4/7 of RMs cases showed negative in ICC. Consistent with ICC results, there was significant difference of TTF1 expression in the LAC and NLAC compared with IMP3 expression. The combination with CK5/6, IMP3 and TTF1 immunostaining appears to be useful to improve the accuracy of cytological diagnoses between RMs, metastatic adenocarcinoma of lung and non-lung origin in pleural effusion. In addition, C-PCR may act as a useful supplemental approach for ICC, especially negative cases in ICC for differential cytological diagnosis.

[Optimization of streptozotocin dosing for establishing tumor-bearing diabetic mouse models].

OBJECTIVE: To determine the optimal dosing of streptozotocin (STZ) for establishing lymphoma-bearing diabetic mouse models. METHODS: A total of 200 healthy male Balb/c mice were randomized into 4 groups (n=50) for intraperitoneal injection of a single dose of vehicle solution (control) or 75, 150, or 200 mg/kg STZ. The changes of body weight and blood glucose were observed regularly, and the success rate of modeling, mortality rate, and survival of the mice were recorded after the injections. The mice with successfully induced diabetes received subcutaneous or tail vein injection of A20 lymphoma cells, and the rate of tumorigenesis, mortality rate, and survival time were observed at 1 month and 3 months after tumor cell injection. RESULTS: Compared with the control group, the mice receiving STZ injection at 150 and 200 mg/kg showed significantly decreased body weight and increased blood glucose (P<0.05), while STZ at 75 mg/kg did not produced such obvious changes. STZ injection at 200 mg/kg resulted in a significantly higher mortality rate and shorter survival time than STZ at 150 mg/kg (P<0.05). In the control group and 150 and 200 mg/kg STZ groups, the rate of tumorigenesis or mortality rate showed no significant differences after subcutaneous injection of A20 lymphoma cells (P>0.05), but differed significantly at 3 months after tail vein injection of the tumor cells (P<0.05). CONCLUSION: Intraperitoneal injection of STZ at 150 mg/kg is associated with a low mortality rate, a high successful modeling rate of diabetes and a long survival time in mice, and is therefore optimal for establishing diabetic mouse models bearing transplanted tumors.

IMP3 as a supplemental diagnostic marker for Hodgkin lymphoma.

Insulin-like growth factor II mRNA-binding protein 3 (IMP3) is ubiquitously expressed in embryos, mediating organogenesis, RNA trafficking, and cell growth, and is generally down-regulated in adult tissue. However, IMP3 has recently been shown to be overexpressed in some malignant epithelial neoplasms and to be a useful diagnostic and/or prognostic biomarker for several carcinomas. To determine whether IMP3 might also be an accurate biomarker of Hodgkin lymphoma, we examined 81 Hodgkin lymphomas for immunoreactivity to IMP3 as compared to commonly used markers such as CD30, CD15, PAX5, and MUM1. Consequently, in 98.8% (80/81) of Hodgkin lymphomas, the malignant Hodgkin and Reed-Sternberg cells were selectively reactive for IMP3, with 72.8% (59/81) of the tumors showing strong, diffuse cytoplasmic staining. Positive staining of the Hodgkin lymphomas was also seen for CD30 (82.7%, 67/81), CD15 (65.4%, 53/81), PAX5 (84.0%, 68/81), and MUM1 85.2% (69/81), but significantly fewer cells showed strong staining intensity for CD30 (32.1%, 26/81), CD15 (17.3%, 14/81), PAX5 (12.3%, 10/81), and MUM1 (29.6%, 24/81). Furthermore, the IMP3 staining was selectively restricted to Hodgkin and Reed-Sternberg cells, with a clearly negative background, and complementary to CD30 staining. Our findings show that IMP3 may be a useful diagnostic marker of Hodgkin lymphoma, helping to improve diagnostic accuracy for this malignancy.