Online Program Home
  My Program

Abstract Details

Activity Number: 75 - Statistical Genomics in Cancer
Type: Contributed
Date/Time: Sunday, July 30, 2017 : 4:00 PM to 5:50 PM
Sponsor: Section on Statistics in Genomics and Genetics
Abstract #324952
Title: Detecting Circulating Tumor DNA Sequences
Author(s): Victor Song*
Companies: University of North Texas
Keywords: Logit classifier ; ROC curve ; Likelihood Ratio Classifier ; Negative binomial distribution

The liquid biopsy procedure to screen for early-stage cancers and monitor treatment responses involves the detection of cancer biomarkers in bodily fluids such as blood and urine. Recent advances in liquid biopsy techniques have involved detecting the presence of tumor-derived cell-free DNA (circulating tumor DNA, ctDNA) in a simple blood test. This detection procedure is minimally invasive and provides an attractive and reliable alternative to tissue biopsy. One of the major challenges in ctDNA analysis lies in its relatively low concentration and the difficulty in detecting ctDNA from the background cell-free DNA fragments derived from normal cells (cfDNA). Despite some recent progress, accurate detection methods remain elusive. In this paper, I develop two types of probabilistic classifiers for distinguishing ctDNA from cfDNA. The performance of the proposed classifiers is evaluated and measured by the receiver operating characteristic (ROC) curve. Its accuracy is demonstrated by the area under the ROC curve. Both types of classifiers are easy to compute and fairly accurate with the potential to become relatively cheap and applicable tools for the early detection of cancers.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

Copyright © American Statistical Association