Cancer therapy is too often a matter of chance. Despite advances in cancer genetics, physicians have only limited information to make decisions about individual patients. People undergo treatments with their fingers crossed, not knowing if they will be helped or harmed.

A group at Massachusetts General Hospital might have a way to personalize cancer care more effectively. It is currently refining and testing a lab-on-a-chip that can sample and analyze the circulating tumor cells from just a teaspoon of a patient's blood, obviating the need, in many patients, for sometimes dangerous biopsies. "The chip will allow rational decision making for cancer patients," says Mehmet Toner, leader of the team that engineered the device at Boston's BioMicroElectroMechanical Systems Resource Center. Toner likens his new system to the way AIDS patients have their viral load and T cells measured so that their medication can be adjusted. "It could be the same for cancer," he offers.

Most carcinomas shed malignant cells that enter the bloodstream and disseminate, sometimes latching on to new areas where they form tumors. These circulating tumor cells (CTCs) constitute just a tiny fraction of blood cells, often fewer than one in a million, in patients with metastatic disease; they are even less abundant in patients with limited, early-stage tumors that have not overtly spread. The researchers realized that CTCs, though rare, offer a potential window into the real-time dynamics of a tumor's biology.

The team has adapted microfluidics technology, developed during the past 25 years to analyze tiny amounts of fluid and gas, to capture those uncommon cells. Like other microfluidics instruments, the CTC-chip, as the group calls it, comprises a silicon-etched chip fitted with microscopic columns, a chamber to enclose the fluid and chip, and a pneumatic pump. The columns, or microposts, function as miniature test tubes where cells and chemicals can mix, adhere and undergo evaluation.

The CTC-chip relies on 78,000 microposts to grab cancer cells from a mix of normal blood components as they meander through the system via exquisitely controlled suction. The posts are coated with antibodies to the epithelial cell adhesion molecule (EpCAM). Nearly all carcinoma cells bear EpCAM at the surface, where it plays a key role in directing how cells bind to one another, signal and migrate. Normal blood cells lack EpCAM, so only the malignant cells stick to the antibodies on the microposts.

In their first test of the chip, described in the December 20, 2007, Nature, the researchers used blood samples from 116 patients with lung, prostate, pancreatic, breast or colorectal cancer and successfully isolated CTCs in all but one case. The CTC-chip finds cancer cells as rare as one in a billion blood cells, making it at least 100-fold more powerful than the most widely used conventional method, which requires incubating a large blood sample with antibody-coated microbeads. The cells are also in better condition for analysis than those prepared using microbeads.

In another trial, the investigators used the chip to evaluate tumor genetics in 27 patients with lung cancer. In work published in the July 24, 2008, New England Journal of Medicine, they identified relevant genetic abnormalities in CTCs from most cases and noted in some patients emerging mutations that confer resistance to tyrosine kinase inhibitors, the type of medication the patients were taking. In the past, repeat biopsy would have been necessary to establish these kinds of genetic changes.

"It's an enormous advance in our ability to monitor patients," comments Roy Herbst, a thoracic oncologist at the University of Texas M.D. Anderson Cancer Center, who was not involved in the work. By providing a noninvasive method to follow the quantity and quality of tumor cells, the chip "offers the possibility of personalized medicine and hence effective therapy," he says.

For the 215,000 people in the U.S. diagnosed with lung cancer every year, the need for better diagnostic tools is particularly urgent, says Toner's collaborator Thomas Lynch, chief of the center for thoracic cancers at Mass General. In lung cancer patients, even a small biopsy bears risk of blood loss, infection and, in rare instances, collapse of the affected lung.

Herbst cautions that the findings need to be validated in larger clinical trials and at other medical centers. At Mass General, the researchers are now evaluating how well the chip measures cancer growth and responses to treatment in patients with breast, ovarian and prostate cancers.

By direct examination of cancer cells in blood, the CTC-chip might also uncover new targets for therapy and help determine when and how metastases arise. Toner sees unlimited possibilities once the chip has proved itself in bigger clinical studies. In the future, he remarks, it could become a screening tool to find nascent cancers and even "could be used at annual checkups."

Getting Personal with Cancer Care

Cancer therapies often fail because of genetic differences among individuals. In personalized cancer care, physicians would tailor treatment to the particular features of each patient's cancer. For example, colon cancer patients often receive treatments with antibodies directed at a growth factor receptor. Although this therapy, costing nearly $10,000 per month, can prolong life and make some patients feel better, it works only in those who have the normal version of K-ras, a gene for a signaling protein crucial for tumor growth. So if physicians examine the tumors for K-ras mutations before prescribing these drugs, they can direct care to those it would help, sparing others the costs, hassle and possible side effects of treatment. E.S.

Note (3/30/09): The headline was modified after publication to correct an error.