Blood Test Detects Over 50 Different Cancers With a 0.7% False-Positive Rate
Industry Insight Jul 14, 2020 | By Laura Elizabeth Lansdowne, Senior Science Writer, Technology Networks.
Earlier this year, researchers reported in the journal Annals of Oncology that they had developed a blood test that was able to detect, with 99.3% specificity, more than 50 types of cancer and identify exactly where in the body the cancer had originated – in some cases before any symptoms of the disease had manifested.
The Circulating Cell-Free Genome Atlas (CCGA) study was designed to characterize cancer signals in the blood of participants, using genome-wide circulating cell-free DNA sequencing in combination with machine learning.
To learn more, Technology Networks spoke with the study’s corresponding author, Michael V. Seiden MD PhD, President of The US Oncology Network. Seiden explains what cell-free (cf)DNA is, discusses the challenges related to using cfDNA to detect cancer, elaborates on the study findings and touches on next steps and related studies currently underway.
Laura Lansdowne (LL): What is cell-free DNA (cfDNA) and how can it be harnessed to detect and localize cancers?
Michael Seiden (MS): Today, the majority of cancers are found too late when outcomes are often fatal, because most deadly cancers have no available screening tests. Tumors shed cfDNA, or fragments, into the bloodstream. GRAIL was formed based on scientific evidence around the associated opportunity for early detection of cancer in people without symptoms. GRAIL is pioneering a multi-cancer early detection test that uses a blood sample to detect more than 50 cancers, identify its location in the body, and significantly reduce the possibility of a false alarm.
LL: What challenges are related to using cfDNA to detect cancer?
MS: There are several challenges in using cfDNA. First, the amount of cfDNA in blood is modest so you need a very sensitive test. Second, normal cells shed DNA so the test must not only detect cfDNA it must also distinguish which cfDNA is from normal cell turnover versus from cancer cells. If cancer-associated cfDNA is detected the goal would be to provide some indication to the tissue from which the cancer originated to help the clinician in the clinical evaluation of the patient with cfDNA linked to cancer.
LL: Can you elaborate on the findings of the Circulating Cell-free Genome Atlas study published in Annals of Oncology?
MS: The report in Annals of Oncology, was a collaborative effort between academic and community centers working with GRAIL to evaluate the ability of the test to demonstrate detectable circulating DNA associated with that patient's cancer. In this study, the patients had a known or suspected cancer. The study also included a large population of individuals who had no cancer diagnosis. The study measured the ability of the test to detect a cancer signal (sensitivity) and of equal importance its ability to return a “no detected signal” in a large population of individuals thought to be free of cancer (specificity).
The test detected more than 50 cancers at a specificity of 99.3%, across all stages, and identified the tissue of origin (where the cancer is located in the body) with high accuracy when a cancer signal was detected. The low false-positive rate of less than 1% reduces unnecessary anxiety and the costs and medical risks of unneeded follow-up procedures.
GRAIL took a comprehensive approach to developing its multi-cancer early detection test by evaluating multiple proprietary next-generation sequencing (NGS)-based technologies in the first sub-study of the Circulating Cell-free Genome Atlas (CCGA) study. NGS involves looking at the genome, and there are many different ways to do this. GRAIL evaluated several approaches that looked at different aspects of genomic alterations, including targeted sequencing, mutations, and methylation. Through rigorous research, it was concluded that methylation patterns of cfDNA fragments provided the richest source of signal for early cancer detection.
GRAIL’s high-efficiency methylation-based technology preferentially targets the most informative regions of the genome and is designed to use its proprietary database and machine-learning algorithms to both detect the presence of cancer and identify the tissue of origin.
LL: Could you touch on some of the study limitations as well as next steps/related studies that are underway?
MS: All studies have limitations, first the test is better in detecting cfDNA in individuals with advanced-stage cancer as compared to individuals with early-stage cancers. Second, the test seemed less sensitive in detecting cfDNA in patients with prostate cancer and a common form of breast cancer, both of which are common tumor types. Finally, the study was done in individuals with a known cancer and hence the study does not provide data on how it would perform in a population that might have this test for cancer screening purposes.
To address this later issue, in February, GRAIL announced the initiation of the PATHFINDER Study, the first time that the investigational multi-cancer early detection test will be used to guide clinical care. This is an important step on the test’s path to commercialization.
The test is not intended to be a substitute for existing screening tests; it supplements them. Extending existing screening to include GRAIL’s test could reduce cancer deaths and decrease healthcare costs by detecting more cancers earlier.
LL: How does your approach differ to current commercially available cfDNA-based tests?
MS: Currently there are no commercially available tests that look to detect cfDNA as part of a screening test for cancer detection. This type of technology is used in obstetrics for detecting fetal DNA and has helped to replace amniocentesis for evaluating the fetus for potential genetic syndromes. In addition, a similar technology evaluating cfDNA test is used in oncology to evaluate patients with known cancer to help evaluate what key mutations an individual’s tumor might harbor which is an important component of applying precision medicine to cancer patients with its greatest application in the care of individuals with lung cancer.
Michael V. Seiden MD, PhD, was speaking with Laura Elizabeth Lansdowne, Senior Science Writer for Technology Networks.