David D. Thiel, MD, discussed the results of a CCGA study and the next steps with this research.
David D. Thiel, MD
A cell-free DNA (cfDNA) test showed high sensitivity and specificity in detecting cancer and identifying the tissue of origin in individuals with confirmed or suspected cancer diagnoses, according to findings from The Circulating Cell-free Genome Atlas (CCGA) study.
Over 15,000 individuals with a confirmed pathologic diagnosis of cancer or a high clinical suspicion of cancer were enrolled in the study, which was broken down into 3 substudies. The initial findings in substudy 1 showed a >99% specificity in cancer detection and >90% accuracy in determining the tissue of origin among patients with a confirmed cancer diagnosis and a cancer signal detected.
The test also demonstrated comparable specificity and sensitivity in a subgroup of patients with a high clinical suspicion of cancer. In substudy 2, the test showed 99.8% specificity (95% CI, 99.4%-99.9%) in the training group and 99.3% specificity (95% CI, 98.3%-99.8%) in the validation group. Among patients with a high clinical suspicion of cancer, the test showed 100% specificity (95% CI, 92.7%-100%) in the training group and 100% specificity (95% CI, 78.2%-100%) in the validation group.
In substudy 2, the sensitivity of the test was 55.2% (95% CI, 52.7%-57.7%) in the training group and 54.9% (95% CI, 51.0%-58.8%) in the validation group. In patients with a high clinical suspicion of cancer, the sensitivity of the test was 40.2% (95% CI, 32.7-48.2%) in the training group and 46.7% (95% CI, 35.1-58.6%) in the validation group.
“The implications of these findings are two-fold,” said lead study author, David D. Thiel, MD. “We’re starting to show that we can use cfDNA to identify people who have cancer and those who don’t. Secondly, we’re starting to get to the point where we may be able to start defining where the cancer is coming from.
In an interview with OncLive, Thiel, who is the chair of the Department of Urology at Mayo Clinic, discussed the results of the CCGA study and the next steps with this research.
OncLive: Could you shed light on the goal of this study?
Thiel: The goal of the CCGA trial is to develop a single blood test we can use to screen the population for multiple cancers, as opposed to multiple tests that each screen for a single cancer. We have direct screening programs in specific cancers, such as prostate cancer with prostate-specific antigen and colon cancer with colonoscopy, but these [screening modalities are] disease-specific. [We wanted to] find a way to do a point-of-care blood draw where we can identify a tumor marker—something to let us know that the individual may have cancer, as well as the tissue of origin of the suspected cancer. That was the goal of the CCGA trial.
How was the trial designed, and what were the findings?
The trial included over 15,000 participants, mostly from US medical centers. The study was divided fairly well between male and female participants, race, and academic versus non-academic medical centers. We included a broad array of patients, some with cancer, some without. We tried to determine the markers that could help us identify participants with cancers and those without.
[The study was broken down by substudies]. The first substudy was designed to identify the best assay we could use to detect and identify these cancers. In the second substudy, we tested this approach in patients with cancer and those without to see whether we could determine who had cancer and who didn’t.
The study that we presented is a breakout of [substudy] 2, which took patients with high clinical suspicion of cancer. We wanted to see whether we could figure out which of those participants had cancer versus those who didn’t, and more specifically, what the tissue of origin was. We wanted to determine where the cancer was coming from.
In the study that we presented, the test showed 100% specificity [of detecting cancer and the tissue of origin]. If a cancer signal was detected, the participant most likely had cancer. [The assay] did a good job of identifying participants with cancer and those without, as well as determining where the tissue of origin would be. Overall, the findings from the high clinical suspicion group aligned well with those from substudy 2.
We haven’t yet delineated [how to determine] the subtype of the cancer. For instance, if we determine that a patient may have cancer in a certain organ, [we can’t tell whether] it’s adenocarcinoma, squamous carcinoma, sarcoma, etc. We haven’t gotten that far yet.
Also, [the assay] may identify a tissue of origin, but that could [tissue] be a metastatic lesion. For instance, if the patient has bladder cancer that is metastatic to the liver, the tissue of origin may [isolate the] liver [tissue] rather than the primary tumor. [The assay] cannot necessarily differentiate primary tumors from metastatic lesions. Regardless, the assay predicted the tissue of origin even in stage IV cancers with approximately 90% accuracy.
How did the results from substudy 2 compare with those in participants who had a high clinical suspicion of cancer?
We put [patients] into the high clinical suspicion group according to symptoms, blood tests, and radiology findings. We found that the sensitivity and specificity [of the assay] matched up to the initial substudy 2 analysis. This tells us that this test could be [an aid] in diagnostics in people with a high clinical suspicion of cancer.
What would be more helpful as we continue to develop this [tool] is to find more stage I/II cancers. In this study, we detected a few stage III/IV cancers, but we could detect those numerous ways. A lot of people would facetiously say, “I could tell, just by blood work and symptoms, whether somebody has stage III/IV cancer.”
What we are really interested in finding is, if we’re going to develop a true screening study for cancer, is cancers that are curable when we find them. I am specifically referring to stage I/II tumors. As we refine the test, it will have to be able to preferentially detect stage I/II tumors if we’re going to make a difference in curing patients and detecting cancers earlier than when they present with metastatic disease.
What are the next steps?
The next step would be to evaluate whether we could take patients with high clinical suspicion, and those without high clinical suspicion of cancer, and give these results back to the provider to see what they do with this information and whether it helped with diagnosis. We want to know whether the results can put them on the right path to delineate what [needs to be done]. The PATHFINDER trial (NCT04241796) is designed to get the results of these tests back to the provider, and will evaluate their next steps and determine whether that [information] will help them in diagnosis.
What additional information would you like to provide about the trial?
The strength of the trial is its high enrollment, notably with a large non-cancer cohort that helped to ensure sufficiently high specificity for population-scale implementation of the test. We enrolled over 15,000 patients at numerous medical centers, which were divided between various races, sexes, etc. The other benefit of this study will be its long-term follow-up. [With this extended follow-up we’ll know], what the clinical implications are and whether [we can be sure] that the people who didn’t have cancer won’t develop it in the future. That will be interesting to see as this longitudinal study plays out. We had this high clinical suspicion group that didn’t have cancer, but how is that to say that these patients won’t develop cancer 2 or 3 years down the road? It will be interesting to evaluate that as we go. Interestingly, in a previous analysis, we observed participants considered as false positives go on to develop cancer up to 15 months after testing.
Could you provide perspective on all of the research that’s being done to develop noninvasive screening modalities for cancer?
I’m impressed by how many investigators are working on the same principle, of trying to figure out noninvasive global screening tests for cancer, which we can use to figure out who has cancer and who doesn’t, as well as what the tissue of origin is.
A number of different avenues are being explored. We’re using cfDNA. There is work with protein markers, which is not new, but it’s being refined. There are many different avenues to get to where we all want to go, which is to be able to detect cancers earlier or before they even form. These are interesting avenues where science will continue to explode.
Thiel DD, Chen X, Kurtzman KN, et al. Prediction of cancer and tissue of origin in individuals with suspicion of cancer using a cell-free DNA multi-cancer early detection test. Presented at: 2020 AACR Virtual Annual Meeting I; April 27-28, 2020. Abstract CT021.