While this experiment showed that the nanoparticles could snag at least some cancer cells within the body, it's not yet clear what proportion of cancer cells were captured and removed. Tests to pinpoint that proportion are planned.

Cornell's King suspects that the technology may be better suited to diagnosing, rather than treating, metastasis. "I think that this technology has much more potential for diagnostics and for detecting cancer cells," he says. "I'm not fully convinced that it could be used to really significantly filter out cancer cells as a therapy."

A similar technology that uses antibody-coated beads to separate out cancer cells has already proved effective in vitro, but the new study's authors believe that the magnetic nanoparticles will be less likely to cause an unwanted immune response and are thus better suited for use inside the body. And because they seem to bind more strongly than antibodies to their targets, says McDonald, they may be better able to pull out cancer cells.

"The ideal would be to try to get everything, but I doubt that would happen," says McDonald. "But we believe that we could significantly reduce the number and thus lower the probability of metastasis."

For now, the treatment seems uniquely suited to ovarian cancer; most other tumors metastasize through cells floating in the bloodstream rather than in the abdominal fluid. But eventually, the team hopes to adapt the particles for use in blood, perhaps extending their use not only to other cancer types, but also to viral diseases such as HIV. To do so, say the researchers, they will need to develop highly specific targeting molecules for each disease to ensure that healthy blood cells are spared.

To test the feasibility of using the nanoparticles in the bloodstream, Ken Scarberry, a graduate student at Georgia Tech and coauthor of the study, reports watching them in action in an artificial circulatory system that passed under a fluorescent microscope. When a magnet was placed near the microscope's lens, "you could see that all of the cells immediately got sequestered over to the side and did not move as the fluid continued to flow," says Scarberry. "This technology has so many possibilities. Right now, I think we're just scratching the surface with it."