A Method for Increasing the Robustness of Stable Feature Selection for Biomarker Discovery in Molecular Medicine Developed Using Serum Small Extracellular Vesicle Associated miRNAs and the Barrett's Oesophagus Disease Spectrum.

The biomarker development field within molecular medicine remains limited by the methods that are available for building predictive models. We developed an efficient method for conservatively estimating confidence intervals for the cross validation-derived prediction errors of biomarker models. This new method was investigated for its ability to improve the capacity of our previously developed method, StaVarSel, for selecting stable biomarkers. Compared with the standard cross validation method, StaVarSel markedly improved the estimated generalisable predictive capacity of serum miRNA biomarkers for the detection of disease states that are at increased risk of progressing to oesophageal adenocarcinoma. The incorporation of our new method for conservatively estimating confidence intervals into StaVarSel resulted in the selection of less complex models with increased stability and improved or similar predictive capacities. The methods developed in this study have the potential to improve progress from biomarker discovery to biomarker driven translational research.

Detection of hypermethylated BCAT1 and IKZF1 DNA in blood and tissues of colorectal, breast and prostate cancer patients.

BACKGROUND: Detection of circulating cell-free DNA (ccfDNA) methylated in BCAT1 and IKZF1 is sensitive for detection of colorectal cancer (CRC), but it is not known if these biomarkers are present in other common adenocarcinomas. OBJECTIVE: Compare methylation levels of BCAT1 and IKZF1 in tissue and plasma from breast, prostate, and colorectal cancer patients. METHODS: Blood was collected from 290 CRC, 32 breast and 101 prostate cancer patients, and 606 cancer-free controls. Tumor and matched normal tissues were collected at surgery: 26 breast, 9 prostate and 15 CRC. DNA methylation in BCAT1 and IKZF1 was measured in blood and tissues. RESULTS: Either biomarker was detected in blood from 175/290 (60.3%) of CRC patients. The detection rate was higher than that measured in controls (48/606 (8.1%), OR = 18.2, 95%CI: 11.1-29.0). The test positivity rates in breast and prostate cancer patients were 9.4% (3/32) and 6.9% (7/101), respectively, and not significantly different to that measured in gender-matched controls (8.0% (33/382) females (OR = 0.84, 95%CI: 0.23-3.1) and 7.6% (26/318) males (OR = 0.86, 95%CI: 0.65-2.1). In tumor and non-neoplastic tissues, 93.5% (14/15) of CRC tumors were methylated in BCAT1 and/or IKZF1 (p< 0.004). Only 11.5% (3/26) and 44.4% (4/9) (p= 0.083) of breast and prostate tumors were hypermethylated in these two genes. CONCLUSIONS: Detection of circulating DNA methylated in BCAT1 and IKZF1 is sensitive and specific for CRC but not breast or prostate cancer.