AUC-based Method Research Articles

AucPR: an AUC-based approach using penalized regression for disease prediction with high-dimensional omics data.

MotivationIt is common to get an optimal combination of markers for disease classification and prediction when multiple markers are available. Many approaches based on the area under the receiver operating characteristic curve (AUC) have been proposed. Existing works based on AUC in a high-dimensional context depend mainly on a non-parametric, smooth approximation of AUC, with no work using a parametric AUC-based approach, for high-dimensional data.ResultsWe propose an AUC-based approach using penalized regression (AucPR), which is a parametric method used for obtaining a linear combination for maximizing the AUC. To obtain the AUC maximizer in a high-dimensional context, we transform a classical parametric AUC maximizer, which is used in a low-dimensional context, into a regression framework and thus, apply the penalization regression approach directly. Two kinds of penalization, lasso and elastic net, are considered. The parametric approach can avoid some of the difficulties of a conventional non-parametric AUC-based approach, such as the lack of an appropriate concave objective function and a prudent choice of the smoothing parameter. We apply the proposed AucPR for gene selection and classification using four real microarray and synthetic data. Through numerical studies, AucPR is shown to perform better than the penalized logistic regression and the nonparametric AUC-based method, in the sense of AUC and sensitivity for a given specificity, particularly when there are many correlated genes.ConclusionWe propose a powerful parametric and easily-implementable linear classifier AucPR, for gene selection and disease prediction for high-dimensional data. AucPR is recommended for its good prediction performance. Beside gene expression microarray data, AucPR can be applied to other types of high-dimensional omics data, such as miRNA and protein data.

DI-026 Overall survival benefit with pomalidomide: area under curves-based reanalysis

BackgroundThe pomalidomide pivotal clinical trial was recently published in Lancet Oncology (PubMed PMID:24007748). Pomalidomide shows a significant benefit on overall survival (OS), with a difference between medians of 4.6 months...

The pharmacokinetics of PF‐734200, a DPP‐IV inhibitor, in subjects with renal insufficiency

PF-734200 is a potent, selective inhibitor of DPP-IV. This two-part study evaluated the pharmacokinetics (PK) of oral 20mg PF-734200 in subjects with varying degrees of renal insufficiency or with end-stage renal disease (ESRD) requiring chronic haemodialysis (HD). The study also assessed the HD clearance of PF-734200 in ESRD. Part 1 included subjects with normal renal function or renal insufficiency but not on HD. Subjects received a single dose of 20mg PF-734200 while fasting and serum and urine samples were collected. In part 2, period 1, 1h after HD, a single 20-mg dose was given to subjects with ESRD and serum samples were collected. After a 7-day washout, subjects received another dose followed by collection of serum samples (period 2), during which HD was initiated 4h after dosing. Dialysate samples were collected to quantify amount of drug removed, from which HD clearance was calculated. The fraction of drug dialysed was calculated using an AUC-based method. Systemic exposures of PF-734200 increased approximately 1.5-, 2.2-, 2.1- and 2.8-fold in subjects with mild, moderate, or severe renal insufficiency or ESRD, respectively, compared with subjects with normal renal function. The terminal half-life increased from 16.2h in subjects with normal renal function to 36.6h in subjects with ESRD. Approximately, 29% of PF-734200 in the body after a single-dose administration was dialysed by 4h HD. Systemic exposure of PF-734200 increases with decreasing renal function. The effect of HD on drug removal is modest.

New AUC-Based Method to Estimate Drug Fraction Removed by Hemodialysis

Background: A supplemental dose is often necessary after hemodialysis depending on the amount of drug removed by hemodialysis. However, there are different methods of estimating this amount, and most methods ignore drug rebound after hemodialysis. In this report we present a new area under the concentration curve (AUC)-based method that provides an estimate of the drug fraction removed by hemodialysis including drug rebound. Methods: Valganciclovir, the oral prodrug of ganciclovir, was administered to 6 patients with end-stage renal disease. Hemodialysis was performed after 32 h. The fraction of ganciclovir removed by hemodialysis was estimated using the new AUC-based method, a classical method (using the slope on and off hemodialysis), the back-extrapolation method, and a reference model (a two-compartment model with zero-order input and first-order elimination). Results: The AUC-based method and the back-extrapolation method provided accurate estimates of the fraction of ganciclovir removed by hemodialysis (47 ± 6 and 46 ± 5%, respectively) compared to the reference model (49 ± 3%). The classical method, which does not account for the rebound of ganciclovir concentrations after hemodialysis, overestimated the removed fraction by 9% (58 ± 3%). Conclusions: The new AUC-based method and the back-extrapolation method accurately estimate the drug fraction removed by hemodialysis for drugs with a rebound after hemodialysis. The AUC-based method is more robust and as efficient compared to the back-extrapolation method.

A general model of metabolite kinetics following intravenous and oral administration of the parent drug.

A model of metabolite pharmacokinetics is developed in terms of residence time distributions and derived non-compartmental measures. It provides quantitative insight into factors determining the concentration-time curve of metabolite following intravenous and oral administration of the precursor drug. The AUCs and higher curve moments (mean residence times and relative dispersions) are calculated/predicted and their dependence on mean absorption time, fraction of first-pass metabolism and intrinsic disposition residence times of the parent drug and metabolite, respectively, is discussed. An AUC-based method for the determination of the first-pass effect is proposed which is not influenced by drug absorption. The approach is valid for linear pharmacokinetic systems exhibiting hepatic and renal elimination of the precursor drug; it is not restricted to specific compartmental models. Limitations of previous concepts of metabolite kinetics are defined. Criteria are presented for the appearance of concave metabolite curves in a semi-logarithmic scale.