C
ould a blood test detect cancer in healthy people? Grail, a Menlo Park, Calif.-based company, has raised $1.6 billion in venture capital to prove the answer is yes. And at the world’s largest meeting of cancer doctors, the company is unveiling data that seem designed to assuage the concerns and fears of its doubters and critics. But outside experts emphasize there is still a long way to go.
The data, from a pilot study that Grail is using to develop its diagnostic before running it through the gantlet of two much larger clinical trials, are being presented Saturday in several poster sessions at the
annual meeting of the American Society of Clinical Oncology. The data show that the company’s test can detect cancer in the blood with relatively few false positives and that it is fairly accurate at identifying where in the body the tumor was found. Another abstract seems to show that the test is more likely to identify tumors if they are more deadly. One big worry with a cancer blood test is that it would lead to large numbers of patients being diagnosed with mild tumors that would be better off untreated.
“The progress of the technology is impressive,” said Dr. Len Lichtenfeld, the acting chief medical officer of the American Cancer Society. But he also urged caution. “Grail is one organization that is pursuing this goal. We will get there. But we still have to prove the technology, and we still have to learn how to apply the technology.”
Dr. Eric Topol, the director and founder of the Scripps Research Translational Institute, called the results “encouraging” and commented that Grail is “trending toward credibility.” He agreed that if these early results hold up, the blood test could be more predictive than existing screening tools, like prostate-specific antigen or mammography. But he, too, emphasized the need for much more research. “All of this requires a large, prospective assessment,” he wrote via email.
Grail is running a preliminary study called the Circulating Cell-Free Genome Atlas (CCGA), which is being conducted in 15,000 patients. The goal from the beginning was to use this study to optimize a diagnostic test. This would then be tested in two more studies: one of 100,000 women enrolled at the time of their first mammogram, and a second of 50,000 men and women between the ages of 50 and 77 in London who have not been diagnosed with cancer. These huge studies are one reason Grail has raised so much money.
But the data being reported at the ASCO meeting are from a tiny sliver of that first study: an initial analysis of 2,301 participants from the training phase of the sub-study, including 1,422 people known to have cancer and 879 who have not been diagnosed. These data are being used to pick exactly what test Grail will run.
Dr. Steven Joffe, the chief of the division of medical ethics at the University of Pennsylvania Perelman School of Medicine, called it “important and necessary work, but very different from the envisioned use as a screening tool.”
The first big surprise is
exactly what test the company is using. When it was spun out of DNA sequencing giant Illumina in 2016, Grail was focused on genes. The idea was that little bits of DNA shed by cancer cells could be detected in the blood. But Grail is now not looking at the genes themselves, but patterns called methylation, which is used by the body to change how the genetic code is read.
“You start out with a few cells, but then they become tremendously differentiated into brain cells, heart cells, skin cells, all of those things,” said Dr. Alex Aravanis, Grail’s head of research and development. “Methylation is the fundamental signal that determines those cell identities and cell fates.”
Grail had previously presented strategies of using DNA in the blood to detect cancer: sequencing the entire genetic code, or a targeted panel of genes, or using methylation. Aravanis said that methylation was not only the most accurate method, but the best at telling where the tumor originated — whether it was breast cancer, lung cancer, or pancreatic cancer, for instance, that the patient had.
The test was set up so that it would have a 99% specificity — meaning that for every 100 people told they had cancer, 1% would actually not have the disease. The test could detect 34% of cancers at stage 1, when the cancer is least risky, 77% at stage 2, and 84% at stage 3. For 94% of patients, the diagnostic gave the tissue of origin of the test, and it was right nine times out of 10.
Ability to correctly identify where the cancer was varied by tumor type. At 99% specificity, the test identified 59% of early-stage lung cancers, 74% of colorectal cancers, and 78% of pancreatic cancers. The test got better when later-stage cancers were included: identifying location for 92% of lung cancers, 97% of colorectal cancers, and 79% of pancreatic cancers.
With these data, it’s possible to start to imagine what the use of a Grail cancer blood test would look like in the real world. Dr. Anne-Renee Hartman, Grail’s vice president of clinical development, said that between 1% and 1.5% of people older than 50 develop cancer each year. So a test that looked at 100,000 individuals and detected cancer in 70% of them would find 700 cancers. If it had a 99% specificity, it would tell 1,000 people who do not have cancer that they had the disease. The test would tell doctors where to look, and they would have to calm those whose tumors could not be found and treat those whose tumors were findable.
This may not sound great. The test would still be scaring more people unnecessarily than it’s helping. But Hartman and Aravanis pointed out that that’s far better than existing tests like mammography, CT screening, or PSA. Those, Aravanis said, have false positive rates that are 10 times higher.
Still, while the Grail executives said they envision the company’s test used to screen large numbers of people, they also pointed out that their big studies contain plenty of people at higher risk, like heavy smokers or people with mutations known to increase their risk of cancer. Anirban Maitra, a pancreatic cancer researcher at MD Anderson Cancer Center, said that if you look just at pancreatic cancer, not all cancers, it’s likely almost 1,000 people who don’t have cancer would be identified for every 15 that are diagnosed. “It may be better to apply tests of this nature in a pre-selected high risk population (mutation carriers, or cohorts being followed for cancer surveillance due to some concurrent high risk features) before going all in on a general population,” he said.
Dr. Charu Aggarwal, an assistant professor of medicine at the University of Pennsylvania, said she, too, would like to see the test tried in patients at higher risk. But she also was surprised and impressed by the data. “I’m really intrigued,” she said.
Aravanis also argued that Grail has another advantage here: that it appears to be detecting deadlier early-stage tumors, no matter where they are in the body. “All cancers are not the same. We really want to find those cancers that are dangerous that grow, progress and cause clinical harm,” he said.
Grail is taking the next steps. Hartman, its head of clinical development, said that the company plans to soon start experiments where it will return results back to patients and their doctors, so they can understand what patients do with the information. Until now, Grail has been taking patients’ blood and watching, but not telling them about the results of its still-experimental tests.
Dr. Otis Brawley, a professor of oncology and epidemiology at Johns Hopkins University, said he was impressed by the data.
“This is truly exciting in that it’s a company that has consistently tried to do the right thing in developing a screening test,” he said. “Many companies have tried to cut corners.”
Aravanis also said that the company is cognizant that its final test can’t be too expensive, and that it is targeting small enough areas of the genome that it can make the test affordable. “We want this to be broadly accessible when deployed,” he said. “Being able to price it in a way that’s compatible with that is very important to us.” He gave a range from $600 to a few thousand dollars, and said that it would get to the low end of that range if possible.
There will definitely be competition. Yesterday,
Thrive, a Cambridge., Mass.-based company, announced that it had raised $110 million to develop a competing test developed by researchers at Johns Hopkins University. That test combines DNA and protein data to tell where in the body tumors are. Thrive says its test will cost hundreds of dollars. The race is on.
Dominic_Tyer