Binary Classification with Imbalanced Data.

When the binary response variable contains an excess of zero counts, the data are imbalanced. Imbalanced data cause trouble for binary classification. To simplify the numerical computation to obtain the maximum likelihood estimators of the zero-inflated Bernoulli (ZIBer) model parameters with imbalanced data, an expectation-maximization (EM) algorithm is proposed to derive the maximum likelihood estimates of the model parameters. The logistic regression model links the Bernoulli probabilities with the covariates in the ZIBer model, and the prediction performance among the ZIBer model, LightGBM, and artificial neural network (ANN) procedures is compared by Monte Carlo simulation. The results show that no method can dominate the other methods regarding predictive performance under the imbalanced data. The LightGBM and ZIBer models are more competitive than the ANN model for zero-inflated-imbalanced data sets.

Forecasting Hospitalizations Due to COVID-19 in South Dakota, USA.

Anticipating the number of hospital beds needed for patients with COVID-19 remains a challenge. Early efforts to predict hospital bed needs focused on deriving predictions from SIR models, largely at the level of countries, provinces, or states. In the USA, these models rely on data reported by state health agencies. However, predicting disease and hospitalization dynamics at the state level is complicated by geographic variation in disease parameters. In addition, it is difficult to make forecasts early in a pandemic due to minimal data. Bayesian approaches that allow models to be specified with informed prior information from areas that have already completed a disease curve can serve as prior estimates for areas that are beginning their curve. Here, a Bayesian non-linear regression (Weibull function) was used to forecast cumulative and active COVID-19 hospitalizations for SD, USA, based on data available up to 2020-07-22. As expected, early forecasts were dominated by prior information, which was derived from New York City. Importantly, hospitalization trends differed within South Dakota due to early peaks in an urban area, followed by later peaks in rural areas of the state. Combining these trends led to altered forecasts with relevant policy implications. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s41666-021-00094-8.

MUC13 mucin augments pancreatic tumorigenesis.

The high death rate of pancreatic cancer is attributed to the lack of reliable methods for early detection and underlying molecular mechanisms of its aggressive pathogenesis. Although MUC13, a newly identified transmembrane mucin, is known to be aberrantly expressed in ovarian and gastro-intestinal cancers, its role in pancreatic cancer is unknown. Herein, we investigated the expression profile and functions of MUC13 in pancreatic cancer progression. The expression profile of MUC13 in pancreatic cancer was investigated using a recently generated monoclonal antibody (clone PPZ0020) and pancreatic tissue microarrays. The expression of MUC13 was significantly (P < 0.005) higher in cancer samples compared with normal/nonneoplastic pancreatic tissues. For functional analyses, full-length MUC13 was expressed in MUC13 null pancreatic cancer cell lines, MiaPaca and Panc1. MUC13 overexpression caused a significant (P < 0.05) increase in cell motility, invasion, proliferation, and anchorage-dependent or -independent clonogenicity while decreasing cell-cell and cell-substratum adhesion. Exogenous MUC13 expression significantly (P < 0.05) enhanced pancreatic tumor growth and reduced animal survival in a xenograft mouse model. These tumorigenic characteristics correlated with the upregulation/phosphorylation of HER2, p21-activated kinase 1 (PAK1), extracellular signal-regulated kinase (ERK), Akt, and metastasin (S100A4), and the suppression of p53. Conversely, suppression of MUC13 in HPAFII pancreatic cancer cells by short hairpin RNA resulted in suppression of tumorigenic characteristics, repression of HER2, PAK1, ERK, and S100A4, and upregulation of p53. MUC13 suppression also significantly (P < 0.05) reduced tumor growth and increased animal survival. These results imply a role of MUC13 in pancreatic cancer and suggest its potential use as a diagnostic and therapeutic target.

Expression and functions of transmembrane mucin MUC13 in ovarian cancer.

MUC13, a transmembrane mucin, is normally expressed in gastrointestinal and airway epithelium. Its aberrant expression has been correlated with gastric colon and cancer. However, the expression and functions of MUC13 in ovarian cancer are unknown. In the present study, the expression profile and functions of MUC13 were analyzed to elucidate its potential role in ovarian cancer diagnosis and pathogenesis. A recently generated monoclonal antibody (clone PPZ0020) was used to determine the expression profile of MUC13 by immunohistochemistry using ovarian cancer tissue microarrays and 56 additional epithelial ovarian cancer (EOC) samples. The expression of MUC13 was significantly (P < 0.005) higher in cancer samples compared with the normal ovary/benign tissues. Among all ovarian cancer types, MUC13 expression was specifically present in EOC. For the functional analyses, a full-length MUC13 gene cloned in pcDNA3.1 was expressed in a MUC13 null ovarian cancer cell line, SKOV-3. Here, we show that the exogenous MUC13 expression induced morphologic changes, including scattering of cells. These changes were abrogated through c-Jun NH(2) kinase (JNK) chemical inhibitor (SP600125) or JNK2 siRNA. Additionally, a marked reduction in cell-cell adhesion and significant (P < 0.05) increases in cell motility, proliferation, and tumorigenesis in a xenograft mouse model system were observed upon exogenous MUC13 expression. These cellular characteristics were correlated with up-regulation of HER2, p21-activated kinase 1, and p38 protein expression. Our findings show the aberrant expression of MUC13 in ovarian cancer and that its expression alters the cellular characteristics of SKOV-3 cells. This implies a significant role of MUC13 in ovarian cancer.

Combined staining of TAG-72, MUC1, and CA125 improves labeling sensitivity in ovarian cancer: antigens for multi-targeted antibody-guided therapy.

Single antigen-targeted intraperitoneal radioimmunotherapy for ovarian cancer has shown limited success. Due to the heterogeneous expression of tumor antigens on cancer cells, a multi-antigen targeting approach appears logical to augment the therapeutic efficacy of antibody-guided therapy. In the interest of developing this novel approach, ovarian cancer tissue microarray slides containing cancer and benign/non-neoplastic tissue samples (n=92) were processed for single-, double-, and triple-antigen labeling using antibodies for the tumor-associated antigens TAG-72, MUC1, and CA125. Among all ovarian cancer types, 72%, 61%, and 50% of the samples showed immunolabeling for TAG-72, MUC1, and CA125, respectively. Expression level of these antigens was significantly (p<0.005) higher in advanced stage carcinomas compared with early stage. Of the 48 epithelial ovarian cancer samples, individual anti-TAG-72, MUC1, and CA125 antibody probing showed labeling in 89.5%, 87.5%, and 73.0% of the cases, respectively. In the majority of the cancer samples (>70%), a heterogeneous labeling pattern was observed (only 30-40% of the cancer cells within the sample were labeled). However, upon combining the three antigens (triple-antigen labeling), 98% of the epithelial ovarian cancer samples were labeled and >95% of the cancer cells within each sample were labeled. Our data indicate that the heterogeneous expression of cancer antigens appears to be a major obstacle in antibody-guided therapy, and this can be overcome by multiple antigen targeting. Therapeutic efficacy of antibody-guided therapy for ovarian cancer treatment will be enhanced by the combined targeting of TAG-72, MUC1, and CA125.