For many cancer patients, survival depends on whether their doctors can successfully remove the tumors in their entirety. But relying on conventional imaging technologies makes this difficult, and surgeons often inadvertently leave cancer cells behind.

Researchers at the University of Washington’s Fred Hutchinson Cancer Research Center have developed an imaging agent that allows physicians to identify and surgically remove smaller areas of tumor tissue than is possible with standard imaging techniques. In tests, the fluorescent imaging agent helped them detect several kinds of cancer in mice, including those of the brain, prostate, and intestines. The researchers hope to begin testing the imaging system in humans in as few as 18 months.

The researchers call the agent “tumor paint” because, when injected into the bloodstream, it seems to cling to and illuminate only tumor cells. The tumor paint allows for what’s called molecular imaging: it tells researchers about a tissue’s chemical activities, rather than showing them what a tissue looks like, as do traditional imaging technologies such as magnetic resonance imaging (MRI). Molecular imaging for cancer could help doctors find and remove traces of tumor cells that are invisible during surgeries guided by preoperative MRIs, the current standard.

The tumor-paint compound has two major parts. One is a protein called chlorotoxin, which binds to cancer cells. This protein, which is made by TransMolecular, a company based in Cambridge, MA, is currently in phase II clinical trials as a targeted treatment for brain cancer. To make chlorotoxin visible to near-infrared cameras, the University of Washington researchers bound it to a fluorescent protein. Near-infrared light can travel only one or two centimeters through the body, which means the University of Washington agent couldn’t be used to diagnose cancer noninvasively in people. But doctors could use near-infrared imaging during surgery to remove a tumor, or use an endoscope to look for colon cancers.

Researchers developing molecular imaging agents for cancer have run up against two major challenges, says James Olson, who led the imaging research and is a practicing physician and professor of pediatric oncology at the University of Washington. It has been difficult to synthesize imaging agents that bind only to cancer cells. And most that have been developed are limited to detecting only one cancer type.

The University of Washington agent solves both problems. It seems to target an enzyme that breaks down the tissues that surround and hold tumors in, enabling them to grow and spread. This enzyme is made by many tumor types.

Olson says that performing near-infrared imaging during surgery could have a great impact on the treatment of tumors that are not sensitive to radiation and chemotherapy, including many prostate, ovarian, and breast cancers. For many common cancers, what determines how long a patient survives is whether doctors succeed in completely removing a tumor. “You assume clusters [of cancer cells] are left behind for all [these surgeries],” says Olson.

Even a small amount of tumor tissue can be deadly. For children with medulloblastoma, for example, if the surgeon leaves behind less than 1.5 cubic centimeters of cancer tissue, the survival rate is 80 percent. If the surgeon leaves more than this amount, survival drops as low as 50 percent. But distinguishing small amounts of tumor tissue during surgery is difficult, especially during brain surgery.

Olson says that near-infrared imaging could be used at what is normally the endpoint of cancer surgeries, when doctors, guided by preoperative images and their own sense of touch and sight, believe they have removed the extent of a tumor. Surgeons could then wheel an infrared imaging system up to a patient who has been administered the fluorescent tumor label, then find and remove any remaining traces of cancer.

Using their near-infrared imaging agent in mice, the University of Washington researchers were able to detect clusters of as few as 200 cancer cells. Olson says that in his 15 years of oncology he has seen “nothing else close” to this sensitivity. This resolution is “pretty impressive,” agrees Khalid Shah, head of the molecular neuroscience imaging lab at Massachusetts General Hospital’s Center for Molecular Imaging Research.

The high resolution of the tumor paint should help surgeons see which tissues to remove–and also what to leave behind. This is a particular problem during brain surgery, when removing normal tissue can mean loss or impairment of major functions. But removing cancerous tissue while sparing normal tissue is a difficult trade-off during all cancer surgeries, says Olson. Surgeons must “decide whether or not to do extensive surgery and take out lots of lymph nodes,” which carry cancer cells from their original site to other places throughout the body. This is a difficult choice: while removing lymph nodes can prevent cancer from coming back, loss of many lymph nodes can cause swelling, aching, and numbness long after patients recover from surgery.

The advantage of the University of Washington imaging agent is its specificity, says Michael Egan, president of TransMolecular, the company that makes chlorotoxin. TransMolecular will work with Olson to move the agent to clinical trials. Olson and Egan are confident that the agent will prove safe for use in humans, given the success of chlorotoxin in human trials and their results in the mouse study. Autopsies and evaluations of the mice’s organ function showed no evidence of toxicity.