Skip to main content

Cell-Free DNA Assays Could Result in a New Blood Test for Early Detection of Cancer

August 2018 Vol 11, Special Issue

Chicago, IL—Early results from a study of genomewide sequencing support the promise of using plasma cell–free DNA (cfDNA) to develop an early cancer detection test with high specificity, said Geoffrey R. Oxnard, MD, Thoracic Oncologist, Dana-Farber Cancer Institute, Boston, at ASCO 2018.

"cfDNA-based blood tests represent an untapped opportunity for cancer detection," said Dr Oxnard. Cancer genotyping using plasma cfDNA has already been adopted for the detection of specific actionable mutations, but it has only been validated for advanced cancer. Evidence that the use of cfDNA may be feasible for the early detection of lung cancer has been limited thus far.

The CCCA Study

The Circulating Cell-free Genome Atlas (CCGA) study is a prospective, multicenter, observational study launch­ed for the purpose of developing a noninvasive assay for cancer detection. In the CCGA study, blood was collected from 749 controls and 878 patients with newly diagnosed untreated cancer. The study included a training set of 1733 patients with clinically evaluable cancers, including 127 patients with lung cancer, and 580 controls. An independent test set for validation included 47 patients with lung cancer and 368 controls. The patients and controls were enrolled from the same center to avoid geographic variability.

The ability of 3 parallel sequencing approaches to detect cancer in the 127 patients with stage I to stage IV lung cancer was explored.

The 3 prototype sequencing assays performed were:

  1. Paired cfDNA and white blood cell (WBC)-targeted sequencing for somatic mutations, including single nucleotide variants or indels
  2. Paired cfDNA and WBC whole-genome sequencing for copy number variation
  3. cfDNA whole-genome bisulfite sequencing to detect abnormal methylation patterns.

Using a next-generation sequencing panel of 813 clinically actionable variants without WBC filtering would detect many cancer cases with genetic variants, but the false-positive rate would be high, said Dr Oxnard. Using a CCGA-targeted assay with WBC filtering across 507 genes, "you find more than half of lung cancers detected, and you would suppress that false-positive rate in the range of 2%," he added.

Results were presented from 118 patients with lung cancer and 561 controls in the training sample. After correction for WBC-matched variants, 5 patients had a cancerlike signal across all 3 assays, 2 of whom were subsequently diagnosed with cancer.

Across the 3 assays, for early-stage curable lung cancer, "approximately half of lung cancers are detected, and for advanced lung cancer, approximately 90% of lung cancers can be detected within the blood," he said. The findings from the training set were validated in the independent test set.

Dr Oxnard said that the cfDNA assays "can detect lung cancer across stages, across histologies, [and] across populations.…There are similar sensitivities observed across the 3 approaches, but they do replicate the ­independent test set."

Dr Oxnard added, "WBC-derived mutations and copy number variations are a significant source of false positives, and they must be accounted for if you're going to achieve high specificity and a low false-positive rate in such an assay."

A New Blood Test for Lung Cancer?

Together, "These results support the promise of using cfDNA-based assays to develop an early cancer detection test with high specificity, and further assay improvement development is ongoing," Dr Oxnard said.

Thousands of patients in the CCGA study remain to be analyzed for further optimization and the development of a cancer diagnostic tool.

ASCO expert David Graham, MD, FASCO, Medical Director, Levine Cancer Institute, Charlotte, NC, commented, "This is an important first step toward an easier way to detect lung cancer at earlier, hopefully more curable, stages. In this first report of a very exciting new methodology, cfDNA from a blood draw detected 51% of early-stage lung cancers with an extremely high specificity."

Last modified: August 30, 2021