Comparative analytical performance of multiple plasma Aβ42 and Aβ40 assays and their ability to predict positron emission tomography amyloid positivity

Background Receiver operating characteristic (ROC) curves are useful tools to evaluate classifiers in biomedical and bioinformatics applications. However, conclusions are often reached through inconsistent use or insufficient statistical analysis. To support researchers in their ROC curves analysis we developed pROC, a package for R and S+ that contains a set of tools displaying, analyzing, smoothing and comparing ROC curves in a user-friendly, object-oriented and flexible interface. Results With data previously imported into the R or S+ environment, the pROC package builds ROC curves and includes functions for computing confidence intervals, statistical tests for comparing total or partial area under the curve or the operating points of different classifiers, and methods for smoothing ROC curves. Intermediary and final results are visualised in user-friendly interfaces. A case study based on published clinical and biomarker data shows how to perform a typical ROC analysis with pROC. Conclusions pROC is a package for R and S+ specifically dedicated to ROC analysis. It proposes multiple statistical tests to compare ROC curves, and in particular partial areas under the curve, allowing proper ROC interpretation. pROC is available in two versions: in the R programming language or with a graphical user interface in the S+ statistical software. It is accessible at http://expasy.org/tools/pROC/ under the GNU General Public License. It is also distributed through the CRAN and CSAN public repositories, facilitating its installation.

To facilitate clinical trials of disease-modifying therapies for Alzheimer's disease, which are expected to be most efficacious at the earliest and mildest stages of the disease, supportive biomarker information is necessary. The only validated methods for identifying amyloid-β deposition in the brain-the earliest pathological signature of Alzheimer's disease-are amyloid-β positron-emission tomography (PET) imaging or measurement of amyloid-β in cerebrospinal fluid. Therefore, a minimally invasive, cost-effective blood-based biomarker is desirable. Despite much effort, to our knowledge, no study has validated the clinical utility of blood-based amyloid-β markers. Here we demonstrate the measurement of high-performance plasma amyloid-β biomarkers by immunoprecipitation coupled with mass spectrometry. The ability of amyloid-β precursor protein (APP)669-711/amyloid-β (Aβ)1-42 and Aβ1-40/Aβ1-42 ratios, and their composites, to predict individual brain amyloid-β-positive or -negative status was determined by amyloid-β-PET imaging and tested using two independent data sets: a discovery data set (Japan, n = 121) and a validation data set (Australia, n = 252 including 111 individuals diagnosed using 11C-labelled Pittsburgh compound-B (PIB)-PET and 141 using other ligands). Both data sets included cognitively normal individuals, individuals with mild cognitive impairment and individuals with Alzheimer's disease. All test biomarkers showed high performance when predicting brain amyloid-β burden. In particular, the composite biomarker showed very high areas under the receiver operating characteristic curves (AUCs) in both data sets (discovery, 96.7%, n = 121 and validation, 94.1%, n = 111) with an accuracy approximately equal to 90% when using PIB-PET as a standard of truth. Furthermore, test biomarkers were correlated with amyloid-β-PET burden and levels of Aβ1-42 in cerebrospinal fluid. These results demonstrate the potential clinical utility of plasma biomarkers in predicting brain amyloid-β burden at an individual level. These plasma biomarkers also have cost-benefit and scalability advantages over current techniques, potentially enabling broader clinical access and efficient population screening.