The director of the National Cancer Institute said yesterday that his agency will move quickly to publicize any harmful side effects from experimental cancer drugs, even if it means sounding an occasional false alarm.

The official, Vincent T. DeVita, told a Senate hearing that the cancer institute no longer will wait months before telling private doctors or the Food and Drug Administration that such drugs may be highly toxic or even fatal. In several cases, he said, his researchers have been remiss in failing to report disturbing results from animal or human tests just because they felt more evidence was needed.

"Our investigators are scientists and they like to be sure," DeVita said. "They don't want to be known as someone who pulls the alarm too often. But it's better to err on the side of reporting a potential reaction."

Nevertheless, he said, "I'm sure there will be other lapses in the future. . . . I'm sure we'll see a drug test or two in the future that will be started without NCI approval."

DeVita was questioned at a second day of hearings before the Senate Labor and Human Resources investigations subcommittee after reports in The Washington Post on the dangers of some experimental cancer drugs.

"The testimony on Tuesday convinced me that the NCI does not adequately protect patient safety," said Sen. Paula Hawkins (R-Fla.), the subcommittee chairman. "Either you promptly report life-threatening drug reactions or you don't. . . . Either you tell patients the truth or you don't."

DeVita said the institute recently reported that one cancer drug, MeCCNU, seemed to be causing leukemia in some patients, but that further study proved this theory untrue. He said he is now concerned that "we may be reporting so many adverse reactions which turn out not to exist that we clutter up the system."

But he acknowledged that the cancer institute had been too slow in reporting evidence that the same drug caused kidney damage in animals -- even though it was being given to 20 young cancer patients at the time.

Hawkins asked why the FDA had sent a routine scientific pamphlet to a New Jersey couple who complained that their 14-year-old son had died from kidney failure after taking MeCCNU. "I have apologized to them . . . .for the insensitive response they got," said J. Richard Crout, chief of the FDA's Bureau of Drugs.

In another instance, when an FDA official contacted the cancer institute for information, Crout's deputy wrote an internal memorandum stressing the agency's "special relationship" with the NCI. "The mere act of sending a letter to NCI . . . is an unusual if not unique step for us," the memo said.

"That does not imply that there is a lesser standard for the NCI," said Arthur Hull Hayes, the FDA commissioner. As for allegations that some FDA employes feel pressured into quickly approving untested drugs, Hayes said: "Some may feel they are being pushed . . . but I have never implied that they're supposed to practice bad science."

DeVita said his agency also had been lax in handling a study that found that the cancer drug Mitoxantrone caused heart failure in animals. "That study was sent to one of our staff members and it sat on his desk for a period of time," he said. "That's unacceptable."

But he said he is still not convinced that the drug is fatal to humans, even though several cancer patients who took the drug have died from heart failure.

DeVita also conceded that some of his researchers resent the FDA, saying: "It is not rare for anyone, including NCI staff, not to like to be investigated and regulated."

Edward N. Brandt Jr., assistant secretary of Health and Human Services, said the government should not write such stringent regulations that it stifles innovation in cancer research.