Cell-Free DNA Sequencing Could Help Early Detection of Lung Cancer

Geoffrey R. Oxnard, MD

Early-stage lung cancers are detectable in cell-free DNA (cfDNA) using a genome-wide sequencing approach, according to findings of The Circulating Cancer Genome Atlas (CCGA) study.

This prospective, longitudinal cohort trial was launched for the development of a noninvasive assay for cancer detection, said Geoffrey R. Oxnard, MD, lead author.

The study enrolled 12,292 patients, of which 70% with cancer and 30% without. Two initial cohorts were created with a training set (1733 clinically evaluable samples, 127 positive for lung cancer) and a test set (n = 980 clinically evaluable samples, 47 positive for lung cancer).

The 3 prototype sequencing assays included paired cfDNA and white blood cell (WBC) targeted sequencing for single nucleotide variants/indels; paired cfDNA and WBC whole genome sequencing (WGS) for copy number variation; and cfDNA whole genome bisulfite sequencing (WGBS) for methylation. Results showed that the WGBS assay detected 41% of stage I-IIIA cancers and 89% of stage IIIB-IV cancers in the population of 127 patients with lung cancer. Additionally, the signal for cancer increased with higher stages of tumor advancement. In the same population, the WGS and targeted assays were similarly effective, which detected 38% and 51% of early-stage cancers and 87% and 89% of late-stage cancers, respectively.

“We are able to aim for high specificity and achieve about 98% specificity,” said Oxnard, adding that the next steps would be to apply these discoveries into a universal diagnostic test for the ultimate clinical utility of all patients with lung cancer.

The findings show that similar sensitivities were observed across histological subtypes, such as adenocarcinoma, squamous, and small cell.

OncLive: What was the rationale for the CCGA study?

In an interview with OncLive, Oxnard, an associate professor of medicine at Harvard Medical School and thoracic oncologist at Dana-Farber Cancer Institute, discussed the promise of genome-wide sequencing in cfDNA.Oxnard: Currently, for the care of patients with advanced lung cancer, we are using blood tests to find mutations. Plasma genotypes, which is analysis of the cfDNA. If you found a mutation—an EGFR or resistance mutation—you have enough there to start a therapy. These blood tests have been widely adopted, and they've made an impact. The question is, "Can we use this same methodology for cancer detection?"

This requires a very different approach. It's broad sequencing that is looking for any signal of cancer. However, maybe we could fill an unmet need. We know that CT screening for lung cancer can be impactful and lead to an improvement in survival. It's not widely adopted, though. Why is that? It could be logistics or false positives—that kind of thing. Perhaps blood tests are an untapped opportunity.

Therefore, that brings us into this study, which enrolled 15,000 participants. We are reporting on a prespecified analysis to see how these approaches are working. These are comparable groups—so similar age, gender, and other demographics. There are 3 parallel sequencing methods. One is a big targeted assay of 507 genes. Second, we're using broad sequencing or genome-wide sequencing. We are also doing genome-wide methylation. [The latter 2] of these have probably never been done, or at least never been published on for cfDNA.

Another key component is the sequencing of WBCs. These are rich with mutations, whether it be p53 or KRAS. There are similar detection rates across the 3 sequencing methods. The detection rates validate in the training sets and test sets. We can find smokers, nonsmokers, and symptomatic lung cancers of all stages and histologies. I was skeptical a while ago. However, we can really detect cancer with high specificity.

When is the best time to perform this type of blood test?

Can this be applied to other tumor types besides lung cancer?

Is there anything else you would like to add?

Therefore, what's the next step? What we need to do is turn these broad sequencing methods into a diagnostic. We need to turn all of these discoveries into a test that can actually benefit patients. Clinical utility is the goal. We want this to actually improve outcomes for patients in the long run.It's hard to anticipate right now. We don't know the final performance of this assay. It is still in early stages. Intuitively, any screening approach is best applied to at-risk individuals. It depends on the false positive rate and a number of other things. Could it be complimentary to other screening methods? How do you work it into practice? These are all important to consider. We won't know until we have a final test.It’s true that by looking through these genomic signals in the blood, you can find other cancers besides lung cancer. In fact, based upon the signals you see, you can anticipate what types of tumors they are. You can assume that people who are at risk for lung cancer are at risk for all sorts of cancers. All these cancers are genomically different. Therefore, this blood test has brought some potential, but also potential confusion.It is worth mentioning that such a cancer detection test is different than what we are currently using in the clinic. If you tried to do this at home with your average [next-generation sequencing] panel, you'll fall into all sorts of problems. You're going to find p53 mutations that aren't cancer at all, they're just related to the white blood cells. We need to be motivated by the progress in cancer genotyping, but there are more adjustments to cover.

Oxnard G, Maddala T, Hubbell E, et all. Genome-wide sequencing for early stage lung cancer detection from plasma cell-free DNA (cfDNA): The Circulating Cancer Genome Atlas (CCGA) study. J Clin Oncol. 2018;36(suppl; abstr LBA8501).