A mammogram is a breast X-ray. It can sometimes find a cancer in a tiny, early form - early enough to be curable.

But for two years a fierce argument has been building among cancer specialists over whether the wide acceptance of such screening has been premature. Many specialists are asking whether X-raying women under 50 who have no sign of cancer may not cause more breast cancer than it cures.

American Cancer Society officials are saying - in the March Reader's Digest, for example - that eight women in 10 aged 35 to 49 fall into high-risk groups that should have these X-rays as often as their doctors say.

Recently, scientific advisers to the government's National Cancer Institute said that advice is wrong.

The advisers urged a halt to any routine screening of women under 50. This is a step the cancer institute (part of the National Institutes of Health in Bethesda) and the private cancer society (the largest voluntary health organization and a strong influence on doctors) had supposedly agreed on last summer.

The federal advisers implied - and in interviews some of them charged - that much such screening is still being done.

They also urged rigorous efforts to cut radiation doses used in hospitals, clinics and doctors' offices.

They said, too, that an ongoing, four-year-old, projected 10-year, $45 million federal "demonstration" - a program at 27 centres, including Georgetown University, in which 20,000 women aged 40 to 64 are being tested for breast cancer by several methods, including mammography - will neither prove nor disprove mammography's safety or value at any age.

The advisers said the cancer institute should go back to square one and start some true "clinical trials" in which volunteers are monitored by mammography and other methods so there can be a genuine comparison of results.

"The present demonstration program is a mistake," said one of the advisers, Dr. Brian Henderson of the University of Southern California.

The argument is certain to leave the average woman more confused than ever over whether to have mammograms. But there are two recommendations on which virtually all doctors agree:

Every woman should examine her breasts monthly and be examined by a doctor periodically, preferably yearly. Such probings find most breast cancer.

No woman should refuse a mammogram if a doctor needs it to help decide whether to operate.

Mammography won its reputation as a potential cancer-finder when the Health Insurance Plan of Greater New York, starting in 1963, gave breast X-rays to 20,000 women aged 40 to 64 and compared the results with those of 42,000 women who did not have the X-rays.

It was on the basis of this study that some cancer society and cancer institute doctors enthusiastically, but perhaps prematurely, decided that mammography worked and should be extended to "demonstrate" whom it helped most. "The HIP study was a very good study," Henderson said. But it wasn't set up to answer questions about mammography."

Last summer, therefore, an advisory group headed by Dr. Lester Breslow, dean of the University of California at Los Angeles School of Public Health, first told the cancer institute it should rethink its policy. His group said that the HIP study only suggested a "possible" benefit of mammography at age 50 and older but didn't prove it, that it showed no measurable benefit at younger ages and that new knowledge of radiation risks dictated a halt to routine mammography at these ages.

The cancer institute and the cancer society issued joint guidelines recommending such a halt. But the guidelines vaguely added: "We do not recommend withholding mammography from a woman 35-50 years if she and the physician agree that it is in her best interest."

Dr. Arthur Holleb, the society's medical director, soon began telling women under 50 that they are at increased risk and may need one or repeated mammograms if they fall into any of these groups: women who have had breast cancer or exploratory breast surgery, a family history of breast cancer, breast problems (chronic cystic mastitis, lumps, thickenings, nipple discharges or abnormalities) or early menstrual history; women who have never had a baby or had their first baby at 30 or older, and women with "fear of breast cancer" who need reassurance.

In the Reader's Digest, Dr. Benjamin Byrd, 1976 cancer society president, lists similar groups and adds women with "unusually large breasts difficult to examine."

"About 80 per cent of women 35 to 50 meet one or another of these criteria," he said.

But the cancer institute advisory committee on epidemiology, headed by Breslow, recently challenged this "notion that certain women under age 50 belong to 'high-risk' groups" and should be mammographed. It said most of the risk factors Holleb and Byrd cite are "small," not by themselves enough to justify routine X-raying - and only women who have had a breast cancer have a clearly higher risk demanding repeated X-rays.

Three advisory groups - Breslow's, one on pathology and one on radiation and cancer - joined in the renewed recommendation that routine X-raying of younger women's breasts be "discontinued."

There apparently will be no cancer institute "policy" until a fourth group finishes studying current demonstration results and a "consensus" meeting is held.

But Dr. Diane Fink, the institute's director of cancer control, agreed with the advisers. She said that on most risk factors "we just don't have enough evidence to state a policy."

Dr. John Bailar III, long an institute critic of overuse of mammography, said screening may be justified in younger women with "a combination of two or three real risk factors." Samuel Shapiro, Johns Hopkins statistician, thinks "any combination of three or more of five factors" - early menstruation, fewer than three pregnancies, age 30 or older at first pregnancy, history of breast conditions and a family history of breast cancer - may justify early mammography.

With the Breslow group, Fink pointed out that adding radiation to some risk factors may multiply a woman's chance of breast cancer.

A single, low-dose mammographic exam, said the advisers, probably increases the average woman's chance of developing breast cancer "by much less than 1 per cent at age 35, and by a progressively smaller percentage with increasing age." But this may add up to 3.5 to 7.5 cases of breast cancer a year per million women from the 10th year after being X-rayed through the remainder of life.

The measure of radiation absorbed within the body is the "rad" (for "radiation absorbed dose"). Doctors at one cancer conference last year said women should ask their radiologists" the X rays will deliver more than one rad at the midpoint of each breast - and go elsewhere if the number will be higher.

"Or go elsewhere if the radiologist doesn't know the answer," Henderson said. "If he doesn't know, he probably isn't giving as low a dose as he should."

There is medical disagreement with the federal advisers' conclusions. Many doctors advocate mammograms for various groups of younger women.

But they cannot stand on firm scientific ground until the government collects more answers, said the advisers recently.

"Until then," said Bailar, "women getting mammographic screening should be informed of the risk."