People usually find out that they have cancer after developing symptoms or through a screening test such as a mammogram—signs that may appear only after the cancer has grown or spread so much that it can't be cured. But what if you could find out from a simple, highly accurate blood test that you had an incipient tumor? By sequencing the abnormal DNA that a tumor releases into a person's bloodstream, researchers are now one step closer to a universal cancer test. Although the technique is now only sensitive enough to detect advanced cancers, that may be a matter of money: As sequencing costs decrease, the developers of the method say, the test could eventually pick up early tumors as well.
The new work is part of a wave of research on using either cells shed into the blood by tumors or free-floating tumor DNA in blood to track the growth and spread of tumors and tailor treatments. The free tumor DNA tests generally rely on looking for known alterations in cancer genes to distinguish cancerous DNA from normal DNA. Seeking a way to detect tumor DNA without knowing its genetic makeup beforehand, postdoctoral researcher Rebecca Leary and others in the labs of Victor Velculescu and Luis Diaz at the Johns Hopkins University School of Medicine in Baltimore, Maryland, and collaborators at other institutions took advantage of an observation they and others have made: No matter the type of cancer, tumor cells almost invariably have substantially altered chromosomes, such as swapped pieces and extra copies of certain genes. This suggests that a test that could detect any chromosomal abnormalities in a person's blood could serve as a general test for cancer.
Now, the researchers have shown that their idea has promise. First, they isolated the free DNA in blood samples from 10 people with advanced colon or breast cancer. Then, using next-generation DNA sequencing methods, they read the entire genome of the DNA in the blood. (The approach was similar to a new test that can detect Down syndrome in a fetus from a pregnant woman's blood sample by looking for just an extra copy of chromosome 21.) The cancer patients all had DNA with chromosomal alterations in their blood, whereas none of 10 healthy controls tested positive, according to the team's report today in Science Translational Medicine.
"There are multiple uses of this approach," Velculescu says. His group initially hopes to track whether a patient's tumor is responding to treatment or regrows after surgery. The test could also be used to decide what drug a patient should get without biopsying her tumor—in some patients, the Hopkins team detected extra copies of two genes known to drive cancer, ERBB2 and CDK6, which can be targeted with existing drugs.
The test isn't that cheap or quick at the moment: Each of the 10 patients' tests in the study cost several thousand dollars just for sequencing and took a month, including the time for analysis. And early detection is still a ways off. The technique has to sift through large amounts of DNA from normal cells that is also floating in blood to find tumor-associated sequences; the portion of DNA in the cancer patients' blood that came from tumors ranged from 47.9% to as low as 1.4%. The test might have to work on blood samples with less than 0.1% tumor DNA to detect small, curable tumors, the researchers suggest. But that is just a matter of doing more sequencing, Velculescu says. And as sequencing costs continue to drop, "in the very near future, this could end up being extremely cheap," he adds.
"The approach has tremendous promise and, should the sequencing strategy become economical, it could have important applications in early cancer detection," says Daniel Haber of Massachusetts General Hospital in Boston who works on using circulating tumor cells to detect and monitor cancer.
Carlos Caldas of the Cancer Research UK Cambridge Research Institute, who, like the Hopkins group, is working on sequencing free tumor DNA in blood, says the new study is the latest showing "that circulating tumor DNA is going to have a great future in all aspects of cancer management. … This is an exploding field." He thinks such tests could reach the clinic within 5 to 10 years.