Explore the vast range of titles available.

Background In the intensive care unit (ICU), delirium is a common, acute, confusional state associated with high risk for short- and long-term morbidity and mortality. Machine learning (ML) has promise to address research priorities and improve delirium outcomes. However, due to clinical and billing conventions, delirium is often inconsistently or incompletely labeled in electronic health record (EHR) datasets. Here, we identify clinical actions abstracted from clinical guidelines in electronic health records (EHR) data that indicate risk of delirium among intensive care unit (ICU) patients. We develop a novel prediction model to label patients with delirium based on a large data set and assess model performance. Methods EHR data on 48,451 admissions from 2001 to 2012, available through Medical Information Mart for Intensive Care-III database (MIMIC-III), was used to identify features to develop our prediction models. Five binary ML classification models (Logistic Regression; Classification and Regression Trees; Random Forests; Naïve Bayes; and Support Vector Machines) were fit and ranked by Area Under the Curve (AUC) scores. We compared our best model with two models previously proposed in the literature for goodness of fit, precision, and through biological validation. Results Our best performing model with threshold reclassification for predicting delirium was based on a multiple logistic regression using the 31 clinical actions (AUC 0.83). Our model out performed other proposed models by biological validation on clinically meaningful, delirium-associated outcomes. Conclusions Hurdles in identifying accurate labels in large-scale datasets limit clinical applications of ML in delirium. We developed a novel labeling model for delirium in the ICU using a large, public data set. By using guideline-directed clinical actions independent from risk factors, treatments, and outcomes as model predictors, our classifier could be used as a delirium label for future clinically targeted models.

Background Cluster analysis is the most common unsupervised method for finding hidden groups in data. Clustering presents two main challenges: (1) finding the optimal number of clusters, and (2) removing “outliers” among the objects being clustered. Few clustering algorithms currently deal directly with the outlier problem. Furthermore, existing methods for identifying the number of clusters still have some drawbacks. Thus, there is a need for a better algorithm to tackle both challenges. Results We present a new approach, implemented in an R package called Thresher , to cluster objects in general datasets. Thresher combines ideas from principal component analysis, outlier filtering, and von Mises-Fisher mixture models in order to select the optimal number of clusters. We performed a large Monte Carlo simulation study to compare Thresher with other methods for detecting outliers and determining the number of clusters. We found that Thresher had good sensitivity and specificity for detecting and removing outliers. We also found that Thresher is the best method for estimating the optimal number of clusters when the number of objects being clustered is smaller than the number of variables used for clustering. Finally, we applied Thresher and eleven other methods to 25 sets of breast cancer data downloaded from the Gene Expression Omnibus; only Thresher consistently estimated the number of clusters to lie in the range of 4–7 that is consistent with the literature. Conclusions Thresher is effective at automatically detecting and removing outliers. By thus cleaning the data, it produces better estimates of the optimal number of clusters when there are more variables than objects. When we applied Thresher to a variety of breast cancer datasets, it produced estimates that were both self-consistent and consistent with the literature. We expect Thresher to be useful for studying a wide variety of biological datasets.

Cleft Lip and Palate Transmembrane Protein 1-Like (CLPTM1L), resides in a region of chromosome 5 for which copy number gain has been found to be the most frequent genetic event in the early stages of non-small cell lung cancer (NSCLC). This locus has been found by multiple genome wide association studies to be associated with lung cancer in both smokers and non-smokers. CLPTM1L has been identified as an overexpressed protein in human ovarian tumor cell lines that are resistant to cisplatin, which is the only insight thus far into the function of CLPTM1L. Here we find CLPTM1L expression to be increased in lung adenocarcinomas compared to matched normal lung tissues and in lung tumor cell lines by mechanisms not exclusive to copy number gain. Upon loss of CLPTM1L accumulation in lung tumor cells, cisplatin and camptothecin induced apoptosis were increased in direct proportion to the level of CLPTM1L knockdown. Bcl-xL accumulation was significantly decreased upon loss of CLPTM1L. Expression of exogenous Bcl-xL abolished sensitization to apoptotic killing with CLPTM1L knockdown. These results demonstrate that CLPTM1L, an overexpressed protein in lung tumor cells, protects from genotoxic stress induced apoptosis through regulation of Bcl-xL. Thus, this study implicates anti-apoptotic CLPTM1L function as a potential mechanism of susceptibility to lung tumorigenesis and resistance to chemotherapy.

Acalabrutinib (ACP-196) is a second-generation inhibitor of Bruton agammaglobulinemia tyrosine kinase (BTK) with increased target selectivity and potency compared to ibrutinib. In this study, we evaluated acalabrutinib in spontaneously occurring canine lymphoma, a model of B-cell malignancy similar to human diffuse large B-cell lymphoma (DLBCL). First, we demonstrated that acalabrutinib potently inhibited BTK activity and downstream effectors in CLBL1, a canine B-cell lymphoma cell line, and primary canine lymphoma cells. Acalabrutinib also inhibited proliferation in CLBL1 cells. Twenty dogs were enrolled in the clinical trial and treated with acalabrutinib at dosages of 2.5 to 20mg/kg every 12 or 24 hours. Acalabrutinib was generally well tolerated, with adverse events consisting primarily of grade 1 or 2 anorexia, weight loss, vomiting, diarrhea and lethargy. Overall response rate (ORR) was 25% (5/20) with a median progression free survival (PFS) of 22.5 days. Clinical benefit was observed in 30% (6/20) of dogs. These findings suggest that acalabrutinib is safe and exhibits activity in canine B-cell lymphoma patients and support the use of canine lymphoma as a relevant model for human non-Hodgkin lymphoma (NHL).

DNA methylation and BLC-2 are potential therapeutic targets in acute myeloid leukemia (AML). We investigated pharmacologic interaction between the DNA methyltransferase inhibitor 5-azacytidine (5-AZA) and the BCL-2 inhibitor ABT-737. Increased BCL-2 expression determined by reverse phase protein analysis was associated with poor survival in AML patients with unfavorable cytogenetics ( n  = 195). We found that 5-AZA, which itself has modest apoptotic activity, acts synergistically with ABT-737 to induce apoptosis. The 5-AZA/ABT-737 combination enhanced mitochondrial outer membrane permeabilization, as evidenced by effective conformational activation of BAX and ∆ψ m loss. Although absence of p53 limited apoptotic activities of 5-AZA and ABT-737 as single agents, the combination synergistically induced apoptosis independent of p53 expression. 5-AZA down-regulated MCL-1, known to mediate resistance to ABT-737, in a p53-independent manner. The 5-AZA/ABT-737 combination synergistically induced apoptosis in AML cells in seven of eight patients. 5-AZA significantly reduced MCL-1 levels in two of three samples examined. Our data provide a molecular rationale for this combination strategy in AML therapy.

Background There have been many recent breakthroughs in processing and analyzing large-scale data sets in biomedical informatics. For example, the CytoGPS algorithm has enabled the use of text-based karyotypes by transforming them into a binary model. However, such advances are accompanied by new problems of data sparsity, heterogeneity, and noisiness that are magnified by the large-scale multidimensional nature of the data. To address these problems, we developed the Mercator R package, which processes and visualizes binary biomedical data. We use Mercator to address biomedical questions of cytogenetic patterns relating to lymphoid hematologic malignancies, which include a broad set of leukemias and lymphomas. Karyotype data are one of the most common form of genetic data collected on lymphoid malignancies, because karyotyping is part of the standard of care in these cancers. Results In this paper we combine the analytic power of CytoGPS and Mercator to perform a large-scale multidimensional pattern recognition study on 22,741 karyotype samples in 47 different hematologic malignancies obtained from the public Mitelman database. Conclusion Our findings indicate that Mercator was able to identify both known and novel cytogenetic patterns across different lymphoid malignancies, furthering our understanding of the genetics of these diseases.

Background Integration of transcriptomic and metabolomic data improves functional interpretation of disease-related metabolomic phenotypes, and facilitates discovery of putative metabolite biomarkers and gene targets. For this reason, these data are increasingly collected in large (> 100 participants) cohorts, thereby driving a need for the development of user-friendly and open-source methods/tools for their integration. Of note, clinical/translational studies typically provide snapshot (e.g. one time point) gene and metabolite profiles and, oftentimes, most metabolites measured are not identified. Thus, in these types of studies, pathway/network approaches that take into account the complexity of transcript-metabolite relationships may neither be applicable nor readily uncover novel relationships. With this in mind, we propose a simple linear modeling approach to capture disease-(or other phenotype) specific gene-metabolite associations, with the assumption that co-regulation patterns reflect functionally related genes and metabolites. Results The proposed linear model, metabolite ~ gene + phenotype + gene:phenotype, specifically evaluates whether gene-metabolite relationships differ by phenotype, by testing whether the relationship in one phenotype is significantly different from the relationship in another phenotype (via a statistical interaction gene:phenotype p -value). Statistical interaction p -values for all possible gene-metabolite pairs are computed and significant pairs are then clustered by the directionality of associations (e.g. strong positive association in one phenotype, strong negative association in another phenotype). We implemented our approach as an R package, IntLIM, which includes a user-friendly R Shiny web interface, thereby making the integrative analyses accessible to non-computational experts. We applied IntLIM to two previously published datasets, collected in the NCI-60 cancer cell lines and in human breast tumor and non-tumor tissue, for which transcriptomic and metabolomic data are available. We demonstrate that IntLIM captures relevant tumor-specific gene-metabolite associations involved in known cancer-related pathways, including glutamine metabolism. Using IntLIM, we also uncover biologically relevant novel relationships that could be further tested experimentally. Conclusions IntLIM provides a user-friendly, reproducible framework to integrate transcriptomic and metabolomic data and help interpret metabolomic data and uncover novel gene-metabolite relationships. The IntLIM R package is publicly available in GitHub ( https://github.com/mathelab/IntLIM ) and includes a user-friendly web application, vignettes, sample data and data/code to reproduce results.

Despite many attempts to establish pre-treatment prognostic markers to understand the clinical biology of esophageal adenocarcinoma (EAC), validated clinical biomarkers or parameters remain elusive. We generated and analyzed tumor transcriptome to develop a practical biomarker prognostic signature in EAC. Untreated esophageal endoscopic biopsy specimens were obtained from 64 patients undergoing surgery and chemoradiation. Using DNA microarray technology, genome-wide gene expression profiling was performed on 75 untreated cancer specimens from 64 EAC patients. By applying various statistical and informatical methods to gene expression data, we discovered distinct subgroups of EAC with differences in overall gene expression patterns and identified potential biomarkers significantly associated with prognosis. The candidate marker genes were further explored in formalin-fixed, paraffin-embedded tissues from an independent cohort (52 patients) using quantitative RT-PCR to measure gene expression. We identified two genes whose expression was associated with overall survival in 52 EAC patients and the combined 2-gene expression signature was independently associated with poor outcome (P<0.024) in the multivariate Cox hazard regression analysis. Our findings suggest that the molecular gene expression signatures are associated with prognosis of EAC patients and can be assessed prior to any therapy. This signature could provide important improvement for the management of EAC patients.

Acute myeloid leukemia (AML) is believed to arise from leukemic stem-like cells (LSC) making understanding the biological differences between LSC and normal stem cells (HSC) or common myeloid progenitors (CMP) crucial to understanding AML biology. To determine if protein expression patterns were different in LSC compared to other AML and CD34+ populations, we measured the expression of 121 proteins by Reverse Phase Protein Arrays (RPPA) in 5 purified fractions from AML marrow and blood samples: Bulk (CD3/CD19 depleted), CD34-, CD34+(CMP), CD34+CD38+ and CD34+CD38-(LSC). LSC protein expression differed markedly from Bulk (n =31 cases, 93/121 proteins) and CD34+ cells (n = 30 cases, 88/121 proteins) with 54 proteins being significantly different (31 higher, 23 lower) in LSC than in either Bulk or CD34+ cells. Sixty-seven proteins differed significantly between CD34+ and Bulk blasts (n = 69 cases). Protein expression patterns in LSC and CD34+ differed markedly from normal CD34+ cells. LSC were distinct from CD34+ and Bulk cells by principal component and by protein signaling network analysis which confirmed individual protein analysis. Potential targetable submodules in LSC included the proteins PU.1(SP1), P27, Mcl1, HIF1α, cMET, P53, Yap, and phospho-Stats 1, 5 and 6. Protein expression and activation in LSC differs markedly from other blast populations suggesting that studies of AML biology should be performed in LSC.

Fludarabine, cyclophosphamide, and rituximab (FCR) has become a gold-standard chemoimmunotherapy regimen for patients with chronic lymphocytic leukaemia. However, the question remains of how to treat treatment-naive patients with IGHV-unmutated chronic lymphocytic leukaemia. We therefore aimed to develop and validate a gene expression signature to identify which of these patients are likely to achieve durable remissions with FCR chemoimmunotherapy. We did a retrospective cohort study in two cohorts of treatment-naive patients (aged ≥18 years) with chronic lymphocytic leukaemia. The discovery and training cohort consisted of peripheral blood samples collected from patients treated at the University of Texas MD Anderson Cancer Center (Houston, TX, USA), who fulfilled the diagnostic criteria of the International Workshop on Chronic Lymphocytic Leukemia, had received at least three cycles of FCR chemoimmunotherapy, and had been treated between Oct 10, 2000, and Oct 26, 2006 (ie, the MDACC cohort). We did transcriptional profiling on samples obtained from the MDACC cohort to identify genes associated with time to progression. We did univariate Cox proportional hazards analyses and used significant genes to cluster IGHV-unmutated samples into two groups (intermediate prognosis and unfavourable prognosis). After using cross-validation to assess robustness, we applied the Lasso method to standardise the gene expression values to find a minimum gene signature. We validated this signature in an external cohort of treatment-naive patients with IGHV-unmutated chronic lymphocytic leukaemia enrolled on the CLL8 trial of the German Chronic Lymphocytic Leukaemia Study Group who were treated between July 21, 2003, and April 4, 2006 (ie, the CLL8 cohort). The MDACC cohort consisted of 101 patients and the CLL8 cohort consisted of 109 patients. Using the MDACC cohort, we identified and developed a 17-gene expression signature that distinguished IGHV-unmutated patients who were likely to achieve a long-term remission following front-line FCR chemoimmunotherapy from those who might benefit from alternative front-line regimens (hazard ratio 3·83, 95% CI 1·94–7·59; p<0·0001). We validated this gene signature in the CLL8 cohort; patients with an unfavourable prognosis versus those with an intermediate prognosis had a cause-specific hazard ratio of 1·90 (95% CI 1·18–3·06; p=0·008). Median time to progression was 39 months (IQR 22–69) for those with an unfavourable prognosis compared with 59 months (28–84) for those with an intermediate prognosis. We have developed a robust, reproducible 17-gene signature that identifies a subset of treatment-naive patients with IGHV-unmutated chronic lymphocytic leukaemia who might substantially benefit from treatment with FCR chemoimmunotherapy. We recommend testing the value of this gene signature in a prospective study that compares FCR treatment with newer alternative therapies as part of a randomised clinical trial. Chronic Lymphocytic Leukaemia Global Research Foundation and the National Institutes of Health/National Cancer Institute.