Development of PCR assays to detect signature circulating tumor DNA methylation markers and KRas mutations for pancreatic ductal adenocarcinoma (PDAC).

Abstract

524 Background: PDAC is a cancer of high mortality. Accurate and cost-effective PCR assays detecting diagnostic and prognostic PDAC markers in blood are desirable for conducting PDAC screening in high-risk populations, evaluating treatments for patients and surveying for post-treatment relapse. Methods: We combined PDAC-diagnosing ctDNA methylation markers with PDAC-driving Kras mutations as targets to develop said PCR tests. Methylation markers were previously validated to classify PDAC plasma at a high accuracy by a targeted methylation sequencing assay. Seven known PDAC driver mutations in KRas exon2 were selected. Taqman-based quantitative methylation-specific PCR (MSP) and ARMS PCR were designed for methylation markers and KRas mutations, respectively, and were validated for technical performances. A step of pre-amplification of input DNA by conventional PCR was included prior to quantitative PCR to improve sensitivity. Over 200 clinical PDAC and control tissue and plasma samples were used to determine their analytical performances in classifying PDAC plasma. Results: All targets were detected at a limit-of-detection of 0.25% or better on standards DNA in PCR. To test clinical plasma samples, methylation markers were first filtered by comparing their levels in PDAC-, para-tumor tissues and whole blood cells. Twenty-four most discriminatory markers for PDAC tissues were selected and used to classify 110 PDAC plasma samples from equal number of normal samples. They were evaluated for their sensitivity at a pre-set specificity of 90%. The 4 best-performing markers were selected to build and cross-validate a general-linear-model (GLM) PDAC classifier, which achieved a sensitivity of 66% at 90% specificity (AUC = 0.829) in cross-validation. Adding status of KRas mutations further improved prediction accuracy by increasing sensitivity to 75% (AUC = 0.89), demonstrating the methylation markers and KRas mutations complement each other in detection PDAC plasma. Conclusions: Our results suggest that a small number of our DNA methylation markers can classify PDAC plasma at a reasonably high accuracy by MSP. Incorporating KRas driver mutations further improved classification accuracy. Together they are promising to be further translated into diagnostics for PDAC early screening, treatment assessment and postoperative surveillance.