Dataset Information

Combination of longitudinal biomarkers in predicting binary events.

ABSTRACT: In disease screening, the combination of multiple biomarkers often substantially improves the diagnostic accuracy over a single marker. This is particularly true for longitudinal biomarkers where individual trajectory may improve the diagnosis. We propose a pattern mixture model (PMM) framework to predict a binary disease status from a longitudinal sequence of biomarkers. The marker distribution given the disease status is estimated from a linear mixed effects model. A likelihood ratio statistic is computed as the combination rule, which is optimal in the sense of the maximum receiver operating characteristic (ROC) curve under the correctly specified mixed effects model. The individual disease risk score is then estimated by Bayes' theorem, and we derive the analytical form of the 95% confidence interval. We show that this PMM is an approximation to the shared random effects (SRE) model proposed by Albert (2012. A linear mixed model for predicting a binary event from longitudinal data under random effects mis-specification. Statistics in Medicine 31: (2), 143-154). Further, with extensive simulation studies, we found that the PMM is more robust than the SRE model under wide classes of models. This new PPM approach for combining biomarkers is motivated by and applied to a fetal growth study, where the interest is in predicting macrosomia using longitudinal ultrasound measurements.

SUBMITTER: Liu D

PROVIDER: S-EPMC4168317 | biostudies-literature | 2014 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Combination of longitudinal biomarkers in predicting binary events.

Liu Danping D Albert Paul S PS

Biostatistics (Oxford, England) 20140514 4

In disease screening, the combination of multiple biomarkers often substantially improves the diagnostic accuracy over a single marker. This is particularly true for longitudinal biomarkers where individual trajectory may improve the diagnosis. We propose a pattern mixture model (PMM) framework to predict a binary disease status from a longitudinal sequence of biomarkers. The marker distribution given the disease status is estimated from a linear mixed effects model. A likelihood ratio statistic ...[more]

PMID: 24831103

Similar Datasets

Project description:Background Serial imaging studies in the general population remain important to evaluate the usefulness of pathophysiologically relevant biomarkers in predicting progression of left ventricular (LV) remodeling and dysfunction. Here, we assessed in a general population whether these circulating biomarkers at baseline predict longitudinal changes in LV structure and function. Methods and Results In 592 participants (mean age, 50.8 years; 51.4% women; 40.5% hypertensive), we derived echocardiographic indexes reflecting LV structure and function at baseline and after 4.7 years. At baseline, we measured alkaline phosphatase, markers of collagen turnover (procollagen type I, C-terminal telopeptide, matrix metalloproteinase-1) and high-sensitivity cardiac troponin T. We regressed longitudinal changes in LV indexes on baseline biomarker levels and reported standardized effect sizes as a fraction of the standard deviation of LV change. After full adjustment, a decline in LV longitudinal strain (-14.2%) and increase in E/e' ratio over time (+18.9%; P≤0.019) was associated with higher alkaline phosphatase activity at baseline. Furthermore, longitudinal strain decreased with higher levels of collagen I production and degradation at baseline (procollagen type I, -14.2%; C-terminal telopeptide, -16.4%; P≤0.029). An increase in E/e' ratio over time was borderline associated with lower matrix metalloproteinase-1 (+9.8%) and lower matrix metalloproteinase-1/tissue inhibitor of metalloproteinase-1 ratio (+11.9%; P≤0.041). Higher high-sensitivity cardiac troponin T levels at baseline correlated significantly with an increase in relative wall thickness (+23.1%) and LV mass index (+18.3%) during follow-up ( P≤0.035). Conclusions We identified a set of biomarkers predicting adverse changes in LV structure and function over time. Circulating biomarkers reflecting LV stiffness, injury, and collagen composition might improve the identification of subjects at risk for subclinical cardiac maladaptation.

Dataset Information

Combination of longitudinal biomarkers in predicting binary events.

Publications

Combination of longitudinal biomarkers in predicting binary events.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets