Researchers have found a compound that tumors make when they are likely to spread, and said they hope to use to it predict which patients are most at risk of dying from their cancers.
And experiments in mice show there may be a way to block the protein, preventing cancer from spreading and becoming deadly.
The findings, published in the Journal of Clinical Investigation, are at a very early stage. But a team at the National Institutes of Health, the University of Hong Kong and elsewhere said on Tuesday they will work to develop both a test and, perhaps, a treatment.
The protein is called CPE-delta N and ordinarily plays a role in processing insulin and other hormones.
"This form is present in large amounts in primary tumors that have spread or metastasized," said Y. Peng Loh of the NIH's National Institute of Child Health and Human Development.
"As everyone knows, cancer cells break away from the primary tumor, pass through surrounding tissues into lymph and blood vessels and these cancer cells then travel through the body and form tumors elsewhere."
This spread of tumors usually kills cancer patients. Early stage cancers that can be completely removed or destroyed usually do not kill the patient.
They tested the tumors of 18 patients with stage 2 liver cancer, which has spread but only within the liver.
"These patients would normally be told by their physicians that the cancer not likely to recur," Loh told reporters in a telephone briefing. They would not get chemotherapy after surgery.
Thirteen of the patients had low levels of CPE-delta N, and 10 of them were still cancer-free three years after surgery. However, three with low CPE-delta N levels did have their cancer come back, giving the test a 77 percent accuracy level in clearing patients.
Five of the original 18 had high levels of CPE-delta N and four of them did have their cancer come back, Loh said -- a 90 percent accuracy rate.
Ideally, if the results are confirmed, patients with high CPE-delta N levels could get extra chemotherapy or radiation to control the risk of spread, Loh said.
Other tests showed similar results, the researchers said, notably in 14 patients with pheochromocytoma, a rare tumor of the adrenal glands, and paraganglioma, another rare tumor. Both mostly affect children and adolescents.
"Testing for CPE-delta N, if combined with existing diagnostic methods, offers the possibility of more accurately estimating the chances that a cancer will spread," said NICHD director Dr. Alan Guttmacher.
The researchers also suppressed CPE-delta N in mice using an experimental approach called antisense and found tumors implanted into the animals did not spread. Antisense is a way to block the activity of genes.
"This offers the potential for developing a cure for certain types of cancers using antisense to CPE-delta N," Loh said.
Loh said her team is also studying how the CPE-delta N gene gets activated in tumors in the first place.
"If we can find out what will shut it down we might be able to find some small molecule that might shut down this gene in some other way besides antisense," she said. Small molecules can usually be made into drugs that can be taken as pills and would be far easier to develop into treatments than antisense.