Using a blood sample to diagnose a budding cancer in a person with no symptoms is a Holy Grail of oncology.
So much so that a leading company in the field chose the name Grail. With a mission “to detect cancer early, when it can be cured,” the California-based start-up has attracted $1.6 billion in venture capital from investors including Microsoft cofounder Bill Gates and Amazon founder Jeff Bezos.
The decades-long quest for a revolutionary “liquid biopsy” has become a race in recent years, driven by the latest genetic analysis technologies. While testing platforms vary, the basic premise is that fragments of DNA shed by a tumor into the bloodstream can be isolated and sequenced to identify mutations specific to the cancer.
But is the premise valid? Can scraps of DNA briefly circulating in the blood truly reveal early cancers of many types?
One skeptic, a prominent cancer diagnostics researcher, says “not likely.” The fundamental reason is simple, as he explained in an interview:
“The amount of cancer DNA in the bloodstream is so little, a tube of blood would not contain any cancer DNA,” said Eleftherios P. Diamandis, head of clinical biochemistry at the University of Toronto, University Health Network, and Mount Sinai Hospital in Toronto.
Diamandis doesn’t doubt the usefulness of cancer DNA tests — once a diagnosis has been made. The technology has proven valuable in predicting the aggressiveness of particular cancers, monitoring patients for relapse or for resistance to a treatment, and for selection of targeted treatments. For example, Roche’s cobas EGFR Mutation test is approved to identify lung cancer patients who would benefit from certain targeted therapies because their tumors have a mutation in the EGFR gene.
But detecting a nascent cancer is another story. Last year, Diamandis and his associate Clare Fiala published a series of critiques in journals.
In addition to being costly and complex, the circulating tumor DNA test “will likely suffer from the same issues of low sensitivity and specificity [accuracy] as traditional biomarkers when applied to population screening and early cancer detection,” they concluded in an article in the journal BMC Medicine.
Traditional biomarkers generally look for proteins shed into the blood or urine. The CA125 protein, for example, is used to monitor ovarian cancer during and after treatment. The prostate-specific antigen (PSA) test is used for early detection of prostate cancer and monitoring for relapse after treatment. But reliability can be a problem because noncancerous conditions can increase the levels of these and other biomarkers.
Surgical biopsies can clarify ambiguous tests, but taking a tissue sample is invasive, expensive, and, depending on the tumor location, difficult or even dangerous. That’s another reason using DNA from a simple blood draw has such appeal.
While early cancer detection with tumor DNA remains a dream, prenatal diagnosis with fetal DNA has become a reality. Down syndrome and other genetic defects can be detected using a maternal blood sample because the mother’s bloodstream contains lots of fetal DNA fragments.
Diamandis and Fiala used previously published studies of prenatal and cancer diagnostics to estimate how big a tumor would have to be to shed enough DNA that fragments could be captured in a standard 10-milliliter (about three teaspoons) vial of blood.
Their bottom line: The tumor would have to be at least 1 centimeter in diameter (a third of an inch) and weigh 1 gram (three-hundreds of an ounce). Although that’s an early-stage cancer, it’s big enough to show up on imaging such as MRI or mammography.
With a much tinier clump of malignant cells, there is unlikely to be a single fragment of tumor DNA in a standard blood sample, the researchers calculated.
“You can ask me, ‘Then how come all these big scientists are publishing positive results in journals?’ ” Diamandis said. “It’s because they choose to study big tumors that are already diagnosed. The idea is that if you prove the concept, then you can make it better later.”
In October, Grail announced new data from using its technology to analyze blood from healthy and newly diagnosed cancer patients. The test was able to “detect more than 20 cancer types across all stages with a single, very low false positive rate.” However, the earlier the cancer stage, the less effective the test. It detected 93% of advanced, incurable cancers, but 18% of early (stage I) cancers.
Diamandis and Fiala pointed out that even if a liquid biopsy proved to be 99% accurate at ruling out cancer, there would be 1,000 false alarms for every 100,000 people screened. Those falsely alarmed people would likely undergo imaging or biopsies, or both, before breathing sighs of relief.
Recent studies have revealed other dilemmas besides the rarity of tumor DNA in incipient cancers: Some mutated DNA in the blood is benign and comes from white blood cells. And some healthy people carry cancer-associated mutations without developing cancer.
Yet another issue — one that plagues current cancer screening tests — is that early detection is not necessarily life-saving, and can sometimes lead to unnecessary treatment. That’s why experts now agree that men should understand the pros and cons of screening for prostate cancer before opting to do it, because many prostate tumors would not cause problems if left undetected, and treatment comes with risks of permanent side effects.