Clin Biochem 2018 Sep 27;59:69-77. Epub 2018 Jun 27.
Service of Medical Biochemistry, Department of Medical Biology, CHU de Québec - Université Laval, Quebec, Canada; Department of Molecular Biology, Medical Biochemistry, and Pathology, Faculty of Medicine, Université Laval, Quebec, Canada; Human and Molecular Genetics Research Unit, Research Center, CHU de Québec, Quebec, Canada; PEGASUS, Quebec, Canada.
Objectives: Non-invasive prenatal aneuploidy testing (NIPT) by next-generation sequencing of circulating cell-free DNA in maternal plasma relies on chromosomal ratio (chr) measurements to detect aneuploid values that depart from euploid ratios. Diagnostic performances are known to depend on the fraction of fetal DNA (FF) present in maternal plasma, although how this translates into specific quantitative changes in specificity/positive predictive values and which other variables might also be important is not well understood.
Design & Methods: To explore this issue, theoretical relationships between FF and various measures of diagnostic performances were assessed for a range of parameter values. Empirical data from three NIPT assays were then used to validate theoretical calculations.
Results: For a given positivity threshold, dramatic changes in specificity and positive predictive values (PPV) as a function of both FF and the coefficient of variation (CV) of the chr measurement were observed. Theoretically predicted and observed chr z-scores agreed closely, confirming the determinant impact of small changes in both FF and chr CV.
Conclusions: Evaluation of NIPT assay performances therefore requires knowledge of the FF distribution in the population in which the test is intended to be used and, in particular, of the precise value of the assay chr CV for each chromosome or genomic region of interest. Laboratories offering NIPT testing should carefully measure these parameters to ensure test reliability and clinical usefulness in interpreting individual patients' results.