High performance methylated DNA markers for detection of colon adenocarcinoma

Abstract

Background Colon cancer (CC) is treatable if detected in its early stages. Improved CC detection assays that are highly sensitive, specific, and available at point of care are needed. In this study, we systematically selected and tested methylated markers that demonstrate high sensitivity and specificity for detection of CC in tissue and circulating cell-free DNA. Methods Hierarchical analysis of 22 candidate CpG loci was conducted using The Cancer Genome Atlas (TCGA) COAD 450K HumanMethylation database. Methylation of 13 loci was analyzed using quantitative multiplex methylation-specific PCR (QM-MSP) in a training set of fresh frozen colon tissues (N = 53). Hypermethylated markers were identified that were highest in cancer and lowest in normal colon tissue using the 75th percentile in Mann–Whitney analyses and the receiver operating characteristic (ROC) statistic. The cumulative methylation status of the marker panel was assayed in an independent test set of fresh frozen colon tissues (N = 52) using conditions defined and locked in the training set. A minimal marker panel of 6 genes was defined based on ROC area under the curve (AUC). Plasma samples (N = 20 colorectal cancers, stage IV and N = 20 normal) were tested by cMethDNA assay to evaluate marker performance in liquid biopsy. Results In the test set of samples, compared to normal tissue, a 6-gene panel showed 100% sensitivity and 90% specificity for detection of CC, and an AUC of 1.00 (95% CI 1.00, 1.00). In stage IV colorectal cancer plasma versus normal, an 8-gene panel showed 95% sensitivity, 100% specificity, and an AUC of 0.996 (95% CI 0.986, 1.00) while a 5-gene subset showed 100% sensitivity, 100% specificity, and an AUC of 1.00 (95% CI 1.00, 1.00), highly concordant with our observations in tissue. Conclusions We identified high performance methylated DNA marker panels for detection of CC. This knowledge has set the stage for development and implementation of novel, automated, self-contained CC detection assays in tissue and blood which can expeditiously and accurately detect colon cancer in both developed and underdeveloped regions of the world, enabling optimal use of limited resources in low- and middle-income countries.

Optimized screening of DNA methylation sites combined with gene expression analysis to identify diagnostic markers of colorectal cancer

“…In this study, we screened as few methylation sites as possible from a large number of methylation sites that can distinguish between cancer and normal groups. In contrast to that of other studies [33], we used the random forest method to calculate the ability of different sites to distinguish between the cancer and control groups. To avoid the in uence of collinearity, we adopted the ridge regression method and nally screened three sites.…”

“…Using the classi er based on DNA methylation, 15.5% (171/1104) of the cases was classi ed into an unambiguous molecular subgroup, which was not possible based on histopathology [35]. A panel of 13 methylated markers could also be effective in the diagnosis of colon cancer; however, the selection of methylated markers in this study was excessive, and the dimensionality reduction process was complicated [33]. The established NaiveBayes model with high accuracy and few characteristic loci has achieved good results in both independent samples and 10-fold cross-validation.…”

The African Research Group for Oncology: A decade fostering colorectal cancer research in Nigeria

“…Several strategies have been investigated, building iteratively on knowledge generated through the group's research efforts. These include clinical risk stratification tools for primary care providers, 25 stool-based screening with qualitative fecal immunohistochemical testing (FIT), 26,27 a novel urinary metabolomics approach to identify a CRC biomarker panel, 28 development of a point-of-care molecular diagnostics test using methylated DNA markers, 29 and a state-wide early detection program at the primary and secondary care levels.…”

DNA methylation biomarkers in colorectal cancer: Clinical applications for precision medicine

“…Multipanel methylation qPCR assays on ccfm DNA revealed BCAT1 , IKZF1 and IRF4 as suitable and flexible in the detection of advanced adenomas and neoplasia 138 . Recently, multiplexed methylation‐specific qPCR on 13 candidate loci in fresh‐frozen tissue yielded 95% to 100% sensitivity and specificity, respectively 139 . When single or few samples have to be analyzed, qPCR is one of the easiest and most applied techniques for DNA methylation detection.…”

“…138 Recently, multiplexed methylation-specific qPCR on 13 candidate loci in fresh-frozen tissue yielded 95% to 100% sensitivity and specificity, respectively. 139 When single or few samples have to be analyzed, qPCR is one of the easiest and most applied techniques for DNA methylation detection. Alternatively, methylation-specific high-resolution melting (HRM) analysis has the capacity to detect cancer-related de novo methylation down to 0.1%.…”