Blood test seems an effective aid in cancer screening

NEW YORK (Reuters Health) - A multicancer blood test appears to have met its initial goal of being an effective addition to standard screening for cancer, in a study of nearly 10,000 women.

The aim of the study “was to demonstrate reliability and safety to show the blood test could lead to the diagnosis of cancers and get patients to treatment aimed at curing them,” said Dr. Kenneth Kinzler of Johns Hopkins University School of Medicine, in Baltimore, Maryland, in a statement,

The new tool, a multi-analyte blood test called DETECT-A from Thrive Earlier Detection Corp., relies on DNA and protein biomarkers, Dr. Kinzler and colleagues explain in Science.

Current standard-of-care (SOC) approaches are effective, they add, and a new multi-cancer test should not supplant or discourage SOC cancer screening. They also point out that studies of blood-based multi-cancer tests have involved people who were already known to have cancer at the time of testing.

To evaluate the test, the team restricted the study to women aged 65 to 75 years with no personal history of cancer and with a high adherence to SOC screening. All participants were enrolled through the Geisinger Health System.

The study had a special focus on ovarian cancer, which lacks SOC screening and “typically has a favorable prognosis only when detected early,” the researchers note.

All cancer types were included in the analysis, except leukemia and those of the skin and central nervous system, which “have very low likelihood of being detected by a blood-based screening test.”

In the baseline blood test 490 (4.9%) of 9,911 participants scored positively and were retested. Of these, 134 were confirmed. Sixty percent of tests not confirmed were due to clonal hematopoiesis of indeterminate potential, the authors say, “which was analyzed more rigorously in the confirmation test than in the baseline test.”

In the group who had positive blood testing, PET-CT was recommended in 116 and other imaging in 11. In total, 63 did not have imaging concerning for cancer and underwent no further testing. Of the 64 who had such imaging, 26 were subsequently shown to have cancer through biopsy or other unequivocal evidence.

Altogether 96 cancer diagnoses were made within the first 12 months of enrollment. In addition to the 26 first detected by blood testing, an additional 24 were first detected by SOC screening. The 46 cases detected by neither method were mostly identified through diagnostic testing prompted by symptoms.

Senior author Dr. Nickolas C. Papadopoulos, also of Johns Hopkins, told Reuters Health by email that the DETECT-A test has subsequently given way to CancerSEEK. “Both tests,” he said, “detect mutations in circulating DNA and elevated levels of certain proteins.”

However, in the DETECT-A study, “It calls positives based on thresholds. In other words, if a mutation, or a protein level is above a certain pre-determined threshold.” On the other hand, CancerSEEK “utilizes a machine learning algorithm that takes into account the values of all analytes (mutations plus protein levels) to identify a positive. We know that this alone improves sensitivity and specificity.”

“Furthermore,” he concluded, “the test has been and is being refined both in its context of analytes and development of the algorithms.”

Dr. Papadopoulos and other authors are founders of and hold equity in Thrive Earlier Detection, and Geisinger has an equity stake in Thrive.

SOURCE: Science, online April 28, 2020.