Single‐molecule epiallelic profiling of DNA derived from routinely collected Pap specimens for noninvasive detection of ovarian cancer

Abstract

AbstractRecent advances in molecular analyses of ovarian cancer have revealed a wealth of promising tumour‐specific biomarkers, including protein, DNA mutations and methylation; however, reliably detecting such alterations at satisfactorily high sensitivity and specificity through low‐cost methods remains challenging, especially in early‐stage diseases. Here we present PapDREAM, a new approach that enables detection of rare, ovarian‐cancer‐specific aberrations of DNA methylation from routinely‐collected cervical Pap specimens. The PapDREAM approach employs a microfluidic platform that performs highly parallelized digital high‐resolution melt to analyze locus‐specific DNA methylation patterns on a molecule‐by‐molecule basis at or near single CpG‐site resolution at a fraction (< 1/10th) of the cost of next‐generation sequencing techniques. We demonstrate the feasibility of the platform by assessing intermolecular heterogeneity of DNA methylation in a panel of methylation biomarker loci using DNA derived from Pap specimens obtained from a cohort of 43 women, including 18 cases with ovarian cancer and 25 cancer‐free controls. PapDREAM leverages systematic multidimensional bioinformatic analyses of locus‐specific methylation heterogeneity to improve upon Pap‐specimen‐based detection of ovarian cancer, demonstrating a clinical sensitivity of 50% at 99% specificity in detecting ovarian cancer cases with an area under the receiver operator curve of 0.90. We then establish a logistic regression model that could be used to identify high‐risk patients for subsequent clinical follow‐up and monitoring. The results of this study support the utility of PapDREAM as a simple, low‐cost screening method with the potential to integrate with existing clinical workflows for early detection of ovarian cancer.Key points We present a microfluidic platform for detection and analysis of rare, heterogeneously methylated DNA within Pap specimens towards detection of ovarian cancer. The platform achieves high sensitivity (fractions <0.00005%) at a suitably low cost (∼$25) for routine screening applications. Furthermore, it provides molecule‐by‐molecule quantitative analysis to facilitate further study on the effect of heterogeneous methylation on cancer development.