In a lead-walled operating room in the basement of Howard University Hospital, brain surgeon Dr. Gary Dennis drilled four tiny holes in the back of Ellis Trappio's skull.

Dennis then carefully sawed out a 3-inch square of bone, exposing Trappio's brain -- and the walnut-sized tumor that had caused him to black out just a week earlier as he lay in bed watching "Nightline."

But after removing the tumor, Dennis did not close up the incision right away as usual. Instead, Dr. Alfred Goldson, chairman of radiotherapy at Howard, set a Plexiglas cone on the surface of Trappio's exposed brain where the tumor had been. The doctors wheeled their patient into position under the arm of a radiation machine called a linear accelerator. They aligned the Plexiglas cone with the accelerator column and left the room.

During the next three minutes, Trappio's brain was bombarded with a precisely-focused beam of electrons -- 1,500 rads worth, or the energy equivalent of 3,000 D-cell batteries.

Trappio, 56, is one of the first brain cancer patients in the world to be treated with a still-experimental procedure pioneered at Howard University Hospital. Called intra-operative radiation therapy (IORT), it combines surgery with radiation for treatment of some deep-seated or inoperable tumors.

Most radiation therapy is delivered from outside the body through the skin. In intraoperative radiation, after as much tumor as possible is removed surgically, the high-energy beam is focused directly on the cancer site in an effort to kill remaining cancer cells so they can't form a new tumor or spread to other parts of the body.

IORT can deliver a higher dose of radiation to the tumor itself, because surrounding organs sensitive to radiation -- such as the skin, intestines and liver -- can be held aside or shielded. The exposed tumor also can be felt and seen directly, rather than on a CAT scan screen.

Trappio's brain cancer had spread from his previously-treated tumors of the colon and lung. His chance of long-term survival was very poor. After talking with his doctors last April, Trappio decided to let them try IORT.

More than six months later, he is alive and active, holding fast to the "good faith" that he says helped pull him through.

Intraoperative radiation is still an experimental treatment -- promising but not a proven cure. Its use remains limited to cases where conventional therapy has been impossible or ineffective, such as some cancers of the pancreas, colon-rectum, cervix and brain.

First reported in 1915, IORT was ignored in the United States until the 1970s, when Goldson began his pioneering studies at Howard. He performed the first modern intraoperative radiation in the United States at Howard University in November 1976.

Since then, about 150 patients have been treated with IORT at Howard, more than at any other medical center in the United States. Goldson has learned to enunciate "Howard" clearly when speaking to out-of-town colleagues so they won't assume he's from Harvard.

"I call it the Howard-Harvard syndrome," jokes Goldson, at 38 the youngest full professor at Howard University Hospital.

Originally, intraoperative radiation was used because early radiation sources were too weak to penetrate the body and reach deep-seated tumors. By the 1950s and 1960s, however, development of powerful radiation machines such as linear accelerators allowed external radiation to penetrate deep within the body. Interest in IORT waned, except in Japan.

Ironically, the very power of those new machines raised a complication that led Goldson and a few other researchers to reconsider IORT. It's the Catch-22 of radiation therapy: A higher dose raises the chance of killing cancerous cells -- but also increases the risk of damaging healthy surrounding cells.

"You could cure every tumor in the world if you gave a million rads," Goldson says. "But you'd kill the patient, too."

In IORT, says Dr. Joel Tepper, associate professor of radiation therapy at Massachusetts terminate composition General Hospital and Harvard Medical School, "the lesser dose to the normal tissues around the tumor allows a heavier dose to the tumor itself."

The main disadvantage of IORT is that it's a one-shot treatment. Since the patient is in surgery, radiation doses cannot be "fractionated," or spread out over several weeks, as in conventional therapy.

Also, since most medical centers don't have an operating room combined with a linear accelerator patients have to be moved from the operating room to the radiation room while under anesthesia.

Howard has the first operating room in the country designed especially for intraoperative radiation patients. It was built in 1975, after Goldson and his mentor, Dr. Ulrich Henschke, then chairman of radiotherapy at Howard, persuaded the university to equip one operating room in its new hospital with a built-in linear accelerator.

Meanwhile, Goldson studied the first English-language translations of the University of Kyoto's IORT research and began tests on dogs to determine correct doses for IORT tumor treatment.

A self-described "good salesman," Goldson had to be at his most convincing to persuade reluctant surgeons to let him intervene with radiation during an operation.

At first they said, "You want to do what to my patient?" he recalls. Now, Howard's surgeons are among the most enthusiastic supporters of further investigation of IORT, he says.

Gradually, following reports of Howard's earlier studies, other leading medical centers got involved. Massachusetts General Hospital began IORT in 1978, the National Cancer Institute in 1979 and the Mayo Clinic in 1981. Several dozen other medical centers are now considering IORT.

A panel discussion on IORT at the recent annual meeting of the American Society for Therapeutic Radiology and Oncology included reports of new studies at Massachusetts General Hospital, the Mayo Clinic, the National Cancer Institute and a Sacramento community hospital.

Most of the results, though promising, are still statistically insignificant. Researchers agree on the need for controlled studies before IORT can be called a cure.

"There are very few home runs in cancer therapy," says Dr. Leonard Gunderson, associate professor of oncology at Mayo Medical School and a consultant to the Mayo Clinic. "This is more likely to be a single."

Goldson and his colleagues now believe that IORT's greatest contribution may turn out to be the ability to use less radical surgery in cancer treatment. Increasingly, IORT is used as a "boost" dose in combination with a series of conventional external doses beginning two weeks after surgery -- and sometimes with chemotherapy as well.

"You need the combination," Goldson says. "Against cancer you need every weapon you have.