Unsupervised Object-Centric Learning From Multiple Unspecified Viewpoints.

Visual scenes are extremely diverse, not only because there are infinite possible combinations of objects and backgrounds but also because the observations of the same scene may vary greatly with the change of viewpoints. When observing a multi-object visual scene from multiple viewpoints, humans can perceive the scene compositionally from each viewpoint while achieving the so-called "object constancy" across different viewpoints, even though the exact viewpoints are untold. This ability is essential for humans to identify the same object while moving and to learn from vision efficiently. It is intriguing to design models that have a similar ability. In this article, we consider a novel problem of learning compositional scene representations from multiple unspecified (i.e., unknown and unrelated) viewpoints without using any supervision and propose a deep generative model which separates latent representations into a viewpoint-independent part and a viewpoint-dependent part to solve this problem. During the inference, latent representations are randomly initialized and iteratively updated by integrating the information in different viewpoints with neural networks. Experiments on several specifically designed synthetic datasets have shown that the proposed method can effectively learn from multiple unspecified viewpoints.

Clinical application of a body area network-based smart bracelet for pre-hospital trauma care.

Objective: This study aims to explore the efficiency and effectiveness of a body area network-based smart bracelet for trauma care prior to hospitalization. Methods: To test the efficacy of the bracelet, an observational cohort study was conducted on the clinical data of 140 trauma patients pre-admission to the hospital. This study was divided into an experimental group receiving smart bracelets and a control group receiving conventional treatment. Both groups were randomized using a random number table. The primary variables of this study were as follows: time to first administration of life-saving intervention, time to first administration of blood transfusion, time to first administration of hemostatic drugs, and mortality rates within 24 h and 28 days post-admission to the hospital. The secondary outcomes included the amount of time before trauma team activation and the overall length of patient stay in the emergency room. Results: The measurement results for both the emergency smart bracelet as well as traditional equipment showed high levels of consistency and accuracy. In terms of pre-hospital emergency life-saving intervention, there was no significant statistical difference in the mortality rates between both groups within 224 h post-admission to the hospital or after 28-days of treatment in the emergency department. Furthermore, the treatment efficiency for the group of patients wearing smart bracelets was significantly better than that of the control group with regard to both the primary and secondary outcomes of this study. These results indicate that this smart bracelet has the potential to improve the efficiency and effectiveness of trauma care and treatment. Conclusion: A body area network-based smart bracelet combined with remote 5G technology can assist the administration of emergency care to trauma patients prior to hospital admission, shorten the timeframe in which life-saving interventions are initiated, and allow for a quick trauma team response as well as increased efficiency upon administration of emergency care.

Compositional Scene Representation Learning via Reconstruction: A Survey.

Visual scenes are composed of visual concepts and have the property of combinatorial explosion. An important reason for humans to efficiently learn from diverse visual scenes is the ability of compositional perception, and it is desirable for artificial intelligence to have similar abilities. Compositional scene representation learning is a task that enables such abilities. In recent years, various methods have been proposed to apply deep neural networks, which have been proven to be advantageous in representation learning, to learn compositional scene representations via reconstruction, advancing this research direction into the deep learning era. Learning via reconstruction is advantageous because it may utilize massive unlabeled data and avoid costly and laborious data annotation. In this survey, we first outline the current progress on reconstruction-based compositional scene representation learning with deep neural networks, including development history and categorizations of existing methods from the perspectives of the modeling of visual scenes and the inference of scene representations; then provide benchmarks, including an open source toolbox to reproduce the benchmark experiments, of representative methods that consider the most extensively studied problem setting and form the foundation for other methods; and finally discuss the limitations of existing methods and future directions of this research topic.

Nek2B activates the wnt pathway and promotes triple-negative breast cancer chemothezrapy-resistance by stabilizing ß-catenin.

BACKGROUND: The chemotherapy-resistance of triple-negative breast cancer (TNBC) remains a major challenge. The Nek2B kinase and ß-catenin serve as crucial regulators of mitotic processes. The aim of this study was to test the correlation between Nek2B and TNBC chemotherapy sensitivity, and to determine the regulation of Nek2B on ß-catenin and wnt/ß-catenin signal pathway. METHODS: Gene Expression Omnibus(GEO) databases were used to gather gene exprsssion data of TNBC patients who undergoing chemotherapy. The co-expression of Nek2B and ß-catenin in TNBC surgical sections and cells were analysed by immunohistochemistry, Q-RT-PCR, Western-blot and immunofluorescent staining. The impact of the expression of Nek2B and ß-catenin in prognosis was also assessed using the Kaplan-Meier curves. CCK8 assay was used to detect the IC50 value of TNBC cell line. The endogenous binding capacity of Nek2B and ß-catenin and phosphorylation of ß-catenin by Nek2B were detected using co-immunoprecipitation (CO-IP). Chromatin immune-precipitation (ChIP) analysis and Luciferase Assays were used to evaluate the binding ability of the Nek2B, ß-catenin and TCF4 complex with LEF-1 promoter. Nek2B-siRNA and Nek2B plasmid were injected into nude mice, and tumorigenesis was monitored. RESULTS: We found that overexpression of Nek2B and ß-catenin in TNBC samples, was associated with patients poor prognosis. Patients with positive Nek2B expression were less sensitive to paclitaxel-containing neoadjuvant chemotherapy. Interestingly, in a panel of established TNBC cell line, Nek2B and ß-catenin were highly expressed in cells exhibiting paclitaxel resistance. Our data also suggest that ß-catenin binded to and was phosphorylated by Nek2B, and was in a complex with TCF4. Nek2B mainly regulates the expression of ß-catenin in TNBC nucleus. Nek2B, ß-catenin and TCF4 can be binded with the WRE functional area of LEF-1 promoter. Nek2B can activite wnt signaling pathway and wnt downstream target genes. The tumors treated by Nek2B siRNA associated with paclitaxel were the smallest in nude mouse, and Nek2B can regulate the expression of ß-catenin and wnt downstream target genes in vivo. CONCLUSION: Our study suggested that Nek2B can bind to ß-catenin and the co-expression correlated with TNBC patients poor prognosis. It appears that Nek2B and ß-catenin might synergize to promote chemotherapy resistance.

Epigenetic silencing of ASPP1 confers 5-FU resistance in clear cell renal cell carcinoma by preventing p53 activation.

Inactivation of p53 has been shown to correlate with drug resistance in tumors. However, in clear cell renal cell carcinoma (ccRCC), p53 is rarely mutated, yet the tumors remain highly insensitive to the conventional chemotherapeutic drugs. The underlying mechanisms responsible for the non-genetic p53 inactivation remain obscure. Here, we report, for the first time, that Apoptosis Stimulating of P53 Protein 1 (ASPP1) was remarkably downregulated at both mRNA (about 3.9-fold) and protein (about 4.9-fold) levels in ccRCC human specimens in comparison with the paired normal controls. In addition, lower ASPP1 was closely related to the higher grade of tumors and shorter life expectancy of ccRCC patients, both with p < 0.001. We also find that CpG island hypermethylation at promoter region contributed to the suppression of ASPP1 expression in ccRCC that contained relatively low levels of ASPP1. Further functional studies demonstrated that forced expression ASPP1 not only significantly inhibited the growth rate of ccRCC, but also promoted sensitivity of ccRCC to the conventional chemotherapeutic drug 5-fluorouracil (5-FU)-induced apoptosis. Moreover, ASPP1 expression was accompanied with the apoptosis-prone alterations of p53 targets expression and p53 target PIG3 luciferase reporter activation. In contrast, ASPP1 knockdown promoted cell growth and prevent 5-FU-induced p53 activation and apoptosis. In conclusion, our results suggest that ASPP1 silencing is one of dominate mechanisms in inhibiting wild type p53 in ccRCC. ASPP1, therefore, may be potentially used as a promising biomarker for prognosis and therapeutic intervention in ccRCC.

Lin28A and androgen receptor expression in ER-/Her2+ breast cancer.

The aim of this study was to examine the expression of Lin28A and androgen receptor (AR) in ER-/Her2+ breast cancer, and to research the association of Lin28A and AR co-expression status with patients' prognosis. The expression of Lin28A and AR in formalin-fixed and paraffin-embedded surgical sections from 305 patients with ER-/Her2+ breast cancer was analyzed by immunohistochemistry, and the co-expression patterns in breast cancer cells were investigated by immunofluorescent staining. The impact of the expression of Lin28A and AR in prognosis was also assessed by the Kaplan-Meier, univariate, and multivariate logistic regression models. This study included 305 cases ER-/Her2+ breast cancer patients. Lin28A and AR were expressed in 240 cases (78.7 %) and 220 cases (72.1 %), respectively. Lin28A tended to be higher in AR-positive patients (75.0 %). Lin28A and AR co-expression (Lin28+/AR+) was significantly associated with high tumor grade (G3) (p = 0.023) and high Ki67 index (p = 0.020). The mRNA and protein expression levels of Lin28A and AR were higher in MDA-MB-453 cells (ER-/Her2+) than in the MDA-MB-231 cells (ER-/Her2-). In univariate analysis, Lin28A+/AR+ was significant risk factors associated with unfavorable OS (p = 0.049) and RFS (p = 0.019). Kaplan-Meier analysis showed that Lin28A+/AR+ expression showed lower RFS rates compared with Lin28A-/AR+ (p = 0.043) and Lin28A-/AR- patients(p = 0.019). Multivariate cox model showed that Lin28A+/AR+ remained an independent negative prognostic factor for RFS. Our study showed that Lin28A and AR co-expressed in ER-/Her-2+ breast cancer and correlated with poor prognosis. The possibility that Lin28A may drive AR expression via a positive feedback mechanism remains to be tested.

Prognostic analysis in a Chinese population with T1-2N1 breast cancer: Did patients with 1 or 2, and 3 positive axillary lymph nodes have similar survival outcomes?

BACKGROUND AND OBJECTIVES: There is a paucity of data examining whether 1-3 positive lymph nodes patients have similar survival outcomes. The present studies separately analyse survival outcomes of T1-2N1 breast cancer patients according to the number of positive lymph nodes. METHODS: A total of 1,030 patients with T1-2N1 breast cancer were available for analysis. Survival estimates were calculated using the Kaplan-Meier method, univariate, and multivariate logistic regression models RESULTS: Kaplan-Meier analysis showed progressively worse survival with the increased number of positive lymph nodes. Log-rank test P values were 0.003 (1 vs. 2 positive LNs), <0.0001 (1 vs. 3), and 0.006 (2 vs. 3) for recurrence-free survival (RFS). Log-rank test P values were 0.045 (1 vs. 2), <0.0001 (1 vs. 3), and 0.018 (2 vs. 3) for metastasis-free survival (MFS). Log-rank test P values were 0.101 (1 vs. 2), <0.0001 (1 vs. 3), and 0.005 (2 vs. 3) for overall survival (OS). Multivariate analysis showed that 3 and 2 positive lymph nodes had worse survival compared with 1 positive axillary lymph nodes. CONCLUSIONS: Our study does suggest that T1-2N1 patients showed progressively worse survival outcomes with the increased number of positive lymph nodes.

Survivin and HLA-I expression predicts survival of patients with clear cell renal cell carcinoma.

Altered expression of survivin and leukocyte antigen class I (HLA-I) proteins is associated with tumor progression. This study investigated their expressions in clear cell renal cell carcinoma (ccRCC) tissues for association with a clinical significance of ccRCC patients. Ninety ccRCC and 20 normal tissue samples (i.e., control) were immunohistochemically stained for survivin and HLA-I expression for an association with clinicopathological data and survival of ccRCC patients. Survivin protein was expressed in 82.2 % (74/90) of ccRCC tissue samples compared to 0 % in the normal tissues, and HLA-I protein was expressed in 90 % (18/20) of the normal tissues vs. 67.8 % (61/90) in ccRCC samples. Survivin expression was associated with tumor grade, stage, and lymph node metastasis (p = 0.000, p = 0.016, and p = 0.001, respectively). Conversely, lost HLA-I expression did not have any associations with clinicopathological data (p > 0.05). Survivin-negative patients had a higher tumor-free survival rate than patients with survivin expression (p = 0.037). Patients with normal HLA-I levels had a higher tumor-free survival rate than those with reduced HLA-I levels (p = 0.02). The uni- and multivariate analyses indicated that expression of survivin and HLA-I, individually and in combination, was an independent predictor for survival of ccRCC patients. Overexpression of survivin but reduced HLA-I expression is useful in the prediction of tumor-free survival of ccRCC patients.

Down-regulation of human leukocyte antigen class I (HLA-I) is associated with poor prognosis in patients with clear cell renal cell carcinoma.

Human leukocyte antigen class I (HLA-I) molecules are transmembrane glycoproteins that have been reported to be down-regulated in multiple types of human malignancies, including clear cell renal cell carcinoma (CCRCC). However, only one study has investigated its prognostic value in CCRCC. In the present study, HLA-I protein expression was analyzed in 120 archived, paraffin-embedded CCRCC samples and 10 adjacent normal tissues using immunohistochemistry. The correlation between HLA-I expression and clinicopathological factors was evaluated by the χ(2) test. Patients' overall survival was analyzed by the Kaplan-Meier method. HLA-I down-regulation was observed in 38.3% (46/120) of renal tumor samples, but only in 10% (1/10) of adjacent normal tissues. Statistical analysis showed a significant correlation of HLA-I expression with TNM stage, lymph node metastasis, and Fuhrman grade. Patients with tumors displaying down-regulation of HLA-I showed significantly shorter overall survival (P=0.021, log-rank test). More importantly, multivariate analysis indicated that down-regulation of HLA-I was an independent prognostic factor for CCRCC patients (P=0.033). Overall, our data suggest that HLA-I down-regulation is associated with tumor progression and a poor prognosis in CCRCC patients, and emphasize the importance of HLA-I in natural and therapeutic immune surveillance of patients with CCRCC.

Peroxisome proliferator-activated receptor gamma is frequently underexpressed in renal cell carcinoma.

AIM: To examine peroxisome proliferator-activated receptor gamma (PPARgamma) mRNA expression quantitatively in human renal cell carcinoma (RCC) cell lines and RCC tissue, as well as in corresponding normal kidney tissue. METHODS: We examined PPARgamma mRNA expression quantitatively in six human RCC cell lines by real-time reverse transcription-polymerase chain reaction. In addition, we evaluated the relationship between cell growth inhibition by PPARgamma ligands and the level of PPARgamma mRNA expression. We compared the expression of PPARgamma mRNA in 47 RCC tissues with that in corresponding normal kidney tissue, and investigated the relationship between clinicopathological features and the level of PPARgamma mRNA expression. RESULTS: Among the six RCC cell lines, five showed decreased PPARgamma mRNA expression. There was no relationship between the inhibitory effects of PPARgamma ligands and PPARgamma mRNA expression levels. Of the tissues from 47 RCC patients, 25 (53%) showed decreased expression of PPARgamma mRNA compared to corresponding normal kidney tissue, and one was equivalent to normal tissue. These patients had distant metastasis at diagnosis more frequently than the remaining patients with high expression. There was also a trend for these patients to have a higher stage. CONCLUSIONS: Most RCC cell lines showed decreased expression of PPARgamma mRNA. However, the level of PPARgamma mRNA expression did not affect cell growth inhibition by PPARgamma ligands. More than half of the tissues from RCC patients had low expression of PPARgamma mRNA, and such carcinomas might have more aggressive behavior.

Ligands for peroxisome proliferator-activated receptor gamma have potent antitumor effect against human renal cell carcinoma.

OBJECTIVES: To examine whether peroxisome proliferator-activated receptor gamma (PPARgamma) is expressed in human renal cell carcinoma (RCC) cells, and whether activation of PPARgamma by its ligands can have multiple antitumor effects on human RCC cells in vitro. METHODS: We examined the expression of PPARgamma in four human RCC cell lines by reverse transcriptase-polymerase chain reaction and immunocytochemical staining. The effects of two PPARgamma ligands, pioglitazone and 15-deoxy-Delta12,14-prostaglandin J2, on cell proliferation were investigated by 3-[4,5-dimethylthiazol-2-thiazoly]-2,5-diphenyltetrazolium bromide assay. The induction of apoptosis by the ligands was examined using the terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling method and Annexin V assay. Furthermore, we investigated whether these ligands suppressed the production of angiogenic factors, vascular endothelial growth factor and basic fibroblast growth factor, by enzyme-linked immunosorbent assay. RESULTS: PPARgamma and retinoid X receptor, which forms a heterodimer with PPARgamma, were expressed in all RCC cell lines. In addition, immunocytochemical studies showed expression of PPARgamma protein in the RCC cells. PPARgamma ligands inhibited the cell growth in all cells in a dose-dependent manner. These ligands also induced apoptosis. Furthermore, secretion of both vascular endothelial growth factor and basic fibroblast growth factor was inhibited by these ligands in a dose-dependent and time-dependent manner. CONCLUSIONS: Ligands for PPARgamma have multiple antitumor effects in human RCC cells in vitro. Activation of the PPARgamma pathway may be a new strategy for treatment of patients with RCC.

Antiproliferative and antiangiogenic activities of genistein in human renal cell carcinoma.

OBJECTIVES: To determine whether genistein, an isoflavone that is plentiful in soy beans, could inhibit the growth of human renal cell carcinoma (RCC) cells in vitro, induce apoptosis, and suppress neovascularization in vivo induced by human RCC cells. METHODS: We investigated the effect of genistein on cell proliferation in four human RCC cell lines, SMKT R-1, R-2, R-3, and R-4. The terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end-labeling assay was performed to examine whether genistein could induce apoptosis in SMKT R-1 cells. To evaluate the effect of genistein on in vivo angiogenesis, we used a new mouse dorsal air sac model in which we could evaluate it simply and quantitatively. Radioisotope-labeled red blood cells were injected into a tail vein in mice bearing a Millipore filter chamber containing genistein, and the vascular volume was examined by measuring the radioactivity of the mouse dorsal skin. RESULTS: Treatment with genistein for 48 hours inhibited cell proliferation in a dose-dependent manner and 100 microg/mL genistein inhibited it in a time-dependent manner. A dose of 50 microg/mL genistein clearly induced cell apoptosis. The vascular volume after implantation of the Millipore filter chamber containing RCC cells increased to threefold that without RCC cells. Genistein in the Millipore filter chamber significantly decreased the neovascularization induced by human RCC cells in vivo. CONCLUSIONS: The results of this study demonstrated that genistein inhibited cell proliferation, induced apoptosis, and suppressed in vivo angiogenesis in human RCC cells. Genistein may be a promising antitumorigenic and antiangiogenic agent for the treatment and prevention of RCC.