Bayesian network data imputation with application to survival tree analysis

Retrospective clinical datasets are often characterized by a relatively small sample size and many missing data. In this case, a common way for handling the missingness consists in discarding from the analysis patients with missing covariates, further reducing the sample size. Alternatively, if the mechanism that generated the missing allows, incomplete data can be imputed on the basis of the observed data, avoiding the reduction of the sample size and allowing methods to deal with complete data later on. Moreover, methodologies for data imputation might depend on the particular purpose and might achieve better results by considering specific characteristics of the domain. The problem of missing data treatment is studied in the context of survival tree analysis for the estimation of a prognostic patient stratification. Survival tree methods usually address this problem by using surrogate splits, that is, splitting rules that use other variables yielding similar results to the original ones. Instead, our methodology consists in modeling the dependencies among the clinical variables with a Bayesian network, which is then used to perform data imputation, thus allowing the survival tree to be applied on the completed dataset. The Bayesian network is directly learned from the incomplete data using a structural expectation–maximization (EM) procedure in which the maximization step is performed with an exact anytime method, so that the only source of approximation is due to the EM formulation itself. On both simulated and real data, our proposed methodology usually outperformed several existing methods for data imputation and the imputation so obtained improved the stratification estimated by the survival tree (especially with respect to using surrogate splits).

Untargeted Metabolomic Analysis of Human Plasma Indicates Differentially Affected Polyamine and L-Arginine Metabolism in Mild Cognitive Impairment Subjects Converting to Alzheimer’s Disease

This study combined high resolution mass spectrometry (HRMS), advanced chemometrics and pathway enrichment analysis to analyse the blood metabolome of patients attending the memory clinic: cases of mild cognitive impairment (MCI; n = 16), cases of MCI who upon subsequent follow-up developed Alzheimer's disease (MCI_AD; n = 19), and healthy age-matched controls (Ctrl; n = 37). Plasma was extracted in acetonitrile and applied to an Acquity UPLC HILIC (1.7μm x 2.1 x 100 mm) column coupled to a Xevo G2 QTof mass spectrometer using a previously optimised method. Data comprising 6751 spectral features were used to build an OPLS-DA statistical model capable of accurately distinguishing Ctrl, MCI and MCI_AD. The model accurately distinguished (R2 = 99.1%; Q2 = 97%) those MCI patients who later went on to develop AD. S-plots were used to shortlist ions of interest which were responsible for explaining the maximum amount of variation between patient groups. Metabolite database searching and pathway enrichment analysis indicated disturbances in 22 biochemical pathways, and excitingly it discovered two interlinked areas of metabolism (polyamine metabolism and L-Arginine metabolism) were differentially disrupted in this well-defined clinical cohort. The optimised untargeted HRMS methods described herein not only demonstrate that it is possible to distinguish these pathologies in human blood but also that MCI patients 'at risk' from AD could be predicted up to 2 years earlier than conventional clinical diagnosis. Blood-based metabolite profiling of plasma from memory clinic patients is a novel and feasible approach in improving MCI and AD diagnosis and, refining clinical trials through better patient stratification.

One-Year Safety of Olodaterol Once Daily via Respimat® in Patients with GOLD 2-4 Chronic Obstructive Pulmonary Disease: Results of a Pre-Specified Pooled Analysis

The novel long-acting β2-agonist olodaterol demonstrated an acceptable safety profile in short-term phase II clinical studies. This analysis of four randomized, double-blind, placebo-controlled, parallel-group, phase III studies (1222.11, NCT00782210; 1222.12, NCT00782509; 1222.13, NCT00793624; 1222.14, NCT00796653) evaluated the long-term safety of olodaterol once daily (QD) in a large cohort of patients with moderate to very severe (Global initiative for chronic Obstructive Lung Disease 2-4) chronic obstructive pulmonary disease (COPD). The studies compared olodaterol (5 or 10 μg) QD via Respimat®, formoterol 12 μg twice daily (BID) via Aerolizer® (1222.13 and 1222.14), and placebo for 48 weeks. Patients continued receiving background maintenance therapy, with ∼60% receiving concomitant cardiovascular therapy and 25% having a history of concomitant cardiac disease. Pre-specified analyses of pooled data assessed the adverse events (AEs) and serious AEs in the whole population, and in subgroups with cardiac disease, along with in-depth electrocardiogram and Holter monitoring. In total, 3104 patients were included in the safety analysis: 876 received olodaterol 5 μg, 883 received olodaterol 10 μg, 885 received placebos, and 460 received formoterol 12 μg BID. Overall incidence of on-treatment AEs (71.2%), serious AEs (16.1%), and deaths (1.7%) were balanced across treatment groups. Respiratory and cardiovascular AEs, including major adverse cardiac events, were reported at similar frequencies in placebo and active treatment groups. The safety profiles of both olodaterol 5 μg (marketed and registered dose) and 10 μg QD delivered via Respimat® are comparable to placebo and formoterol BID in this population, with no safety signals identified.

Dietary fat and breast cancer mortality: a systematic review and meta-analysis

Background: The influence of dietary fat upon breast cancer mortality remains largely understudied despite extensive investigation into its influence upon breast cancer risk 

Objective: To conduct meta-analyses of studies to clarify the association between dietary fat and breast cancer mortality Design: MEDLINE and EMBASE were searched for relevant articles published up to March 2012. Risk of all-cause or breast cancer specific death was evaluated by combining multivariable adjusted estimates comparing highest versus lowest categories of intake; and per 20 gram increase in intake of total and/or saturated fat (g/day) using random-effects meta-analyses. 

Results: Fifteen prospective cohort studies investigating total fat and/or saturated fat intake (g/day) and breast cancer mortality were included. There was no difference in risk of breast cancer specific death (n = 6; HR = 1.14; 95% CI: 0.86, 1.52; P = 0.34) or all cause death (n = 4; HR = 1.73; 95% CI: 0.82, 3.6; P = 0.15) for women in the highest versus lowest category of total fat intake. Breast cancer specific death (n = 5; HR = 1.63; 95% CI: 1.19, 2.24; p <0.01) was higher for women in the highest versus lowest category of saturated fat intake. 

Conclusions: These meta-analyses have shown that saturated fat intake negatively impacts upon breast cancer survival.