ROCKVILLE, Md. (AP) _ TranScan Inc. predicts using electricity to map out possible breast cancer could cut 200,000 unnecessary biopsies every year, but government scientists say the company's machine needs more study before U.S. doctors try it.

Advisers to the Food and Drug Administration called the T-scan system a promising technology, but unanimously agreed Monday that there's not enough proof about how well it works _ and in which women.

``I read 25,000 mammograms a year and ... I'm not sure I'd know who to use this on,'' said Dr. Judy Destouet of Advanced Radiology in Baltimore.

Still, ``It is a technology that needs to be pursued,'' said Dr. Patricia Romilly-Harper of the Breast Cancer Center in Indianapolis.

TranScan says the nonsurgical test, already used in Israel, is valuable in helping doctors pick which women need a biopsy when their mammograms alone don't give enough information to decide.

``Who's going to get the T-scan are the ones you're on the fence about. Those to me happen every day,'' said Dr. Michele Rossman of Sinai Women's Health Center in Detroit, who helped study the device.

Some 800,000 biopsies are done every year, from which 180,000 women are diagnosed with breast cancer. The rest are benign lumps. The question is how to reduce the number of biopsies without missing women who have cancer.

Some doctors try ultrasound machines that use sound waves to picture lumps as an addition to the mammogram's X-rays.

The T-scan would offer a unique alternative _ using electricity to map the breast.

A painless, one-volt shot of electricity is sent into the hand, where it travels through the body into the breast. A hand-held probe is moved over the breast, where it measures the electrical conductivity of breast cells. Cancer cells conduct much less electricity than healthy cells _ so when the probe flashes its findings onto a computer screen, possible tumors show up as bright white spots.

In a study of 504 breasts, adding T-scans to regular mammograms allowed researchers to accurately predict _ 86 percent of the time _ that the woman had cancer, TranScan said. The company called it particularly useful in women under age 50, whose naturally denser breasts are more difficult for mammograms to penetrate.

The combination also allowed researchers to cut by 28 percent the number of biopsies performed on women whose breast lumps turned out to be benign, the company said.

But that 86 percent rate was not statistically better than mammograms alone _ and FDA scientists warned that the test wasn't always accurate, meaning that using it on the wrong women could deter a small number who did have cancer from getting a needed biopsy.

In addition, the FDA's advisers said the study did not reflect the way doctors diagnose breast cancer, leaving them uncertain about the machine's effect when physicians use it every day. Doctors today use multiple, magnified views of mammograms in addition to physical exams before deciding who needs a biopsy, while the study only told doctors the women's ages and provided just two mammogram images for the decision.

``There are too many leaps of faith for me to feel comfortable with it,'' said Dr. David Hackney of the University of Pennsylvania.