Scott Hilderbrand, a chemist at Massachusetts General Hospital's Center for Molecular Imaging Research, is also developing targeted fluorescent probes. He notes that the individual pieces of the technique developed by Tsien's team have been shown to work by other researchers, "but to be able to do this with one agent is one of the more major advances of this approach."

The researchers hope they may be able to add yet another feature to their molecule. "We'd love to think that with this fluorescence we've gotten 100 percent of the cells, but that's not the case," says surgeon Quyen Nguyen, one of the papers' first authors. She says they're working to attach a third branch to the molecule, one that becomes toxic in the presence of bright light. "At the end of the surgery, you could shine a bright light that targets the fluorescent molecule and makes it phototoxic. That way, I can kill the residual cells," Nguyen says.

One of the drawbacks of using MMPs, however, is that they are not expressed in all cancers, and are present in some noncancerous tissues, too, including the liver and in areas of inflammation. "The mice they tested this in have only normal tissue and cancer tissue, but the human body is not so simple" says Hisataka Kobayashi, a molecular imaging specialist at the National Cancer Institute in Bethesda, MA, and another person working on the targeting cancer with fluorescent probes.

That is why the researchers are targeting their probe for use in surgical guidance, where an experienced surgeon can distinguish between cancerous tissue and inflamed areas elsewhere in the body. Nguyen says the same team is working on using the molecule to deliver therapeutics targeted directly to cancer cells, but notes that this is going to be a bit trickier.

The researchers are looking into other potential uses for their molecule, such as lighting up arterial plaques in order to identify ones most at risk of causing a stroke or heart attack. Avelas Biosciences, a new startup based in San Diego, has licensed the probe technology in 2009 and hopes to have something ready for human testing within two to three years.