“When we look at the fragmentation in the cell-free DNA, essentially what we’re looking at is these differences in genome organization,” said Jillian Phallen, PhD, postdoctoral fellow at Johns Hopkins and a lead author on the study. “That results in size differences in the fragments, and it also results in differences in the placement of these fragments across the genome.” Among healthy people, she said, the size range of the fragments and their arrangement on the chromosomes are remarkably consistent, while those with cancer have much more variation both in fragment size and placement. “When you compare the 2, they’re fairly easy to distinguish.”
Blood tests analyzing ctDNA hold many advantages over tissue biopsies. Besides being less invasive, they may also give better representation of a cancer’s heterogeneity. Testing for cancer driver mutations in ctDNA can reveal the presence of cancer before diagnosis — but for smaller tumors, it may be difficult to collect enough circulating tumor DNA to detect the mutations.
“If the tumor is less than 10 mm in diameter, you will likely retrieve very little or no DNA from the circulation, thus making the diagnosis impossible,” said Dr Diamandis. In addition, testing for mutations limits the detection to only known mutations. Cancers with novel or less well-known mutations can slip by undetected.
Another approach, testing for methylation patterns, looks more broadly across the genome, avoiding the drawback of missing cancers that don’t harbor a specific mutation. GRAIL, a Menlo Park, California-based start-up, achieved high performance with a methylation-based test for early detection; this test also identified the cancer’s tissue of origin. GRAIL is currently conducting the Circulating Cell-free Genome Atlas (CCGA) study. At the 2019 American Society of Clinical Oncology (ASCO) Annual Meeting in Chicago, Illinois, they presented results showing that their methylation-based test successfully identified 12 different cancers with 76% sensitivity and 99% specificity.2 In 94% of cases, they were able to detect the malignancy’s tissue of origin, and in these cases, they provided a correct result 90% of the time.
Early-detection tests run the risk of overdiagnosis, and it’s important to find ways for the test to distinguish between likely harmless cancers and those that could become deadly.