Some 20 years ago Carolyn Magin found a lump in one of her breasts. It was removed and found to be benign. The following year she found another. It, too, was benign.

In the passing years, she has had no more surgery, but she does have fibrocystic breast disease. It is a relatively common condition and although it can be painful, especially before menstrual periods, it is regarded as benign and not, in itself, serious. There is, however, some evidence indicating an increased risk of breast cancer among women with this condition.

In Carolyn Magin's case, this suspicion is coupled with a strong family history of breast cancer. Two maternal aunts, for example, were among its victims. And there is a strong body of evidence of familial predisposition to malignant breast tumors, evidence that continues to mount.

So for the past 20 years or so Carolyn Magin and her doctors and, more recently, her radiologist, have kept her under careful observation. Twice a year she travels from her home near Harrisburg, Pa. (which she describes now as being "10 miles from Three Mile Island") to specialists in New York City. She has a thermogram every six months and a mammogram every 18 months. There is, always, at least a little kernel of worry in her mind.

As Dr. John Peter Minton, surgical oncologist, recalls, it was in the early '70s and he had just given a paper to a group of science writers in Noglaes, Ariz. The paper dealt with his efforts to treat a condition in which the production of breast milk could not be stopped.

After his presentation he was approached by a man who asked him if he'd "looked into cyclic AMPs." No, he allowed to the "old fella" whom he did not know; no, he hadn't thought about cyclic AMPs. And that might have been the end of it, except he learned quickly that the "fella" was Dr. Earl Sutherland, the now-deceased physiologist from Vanderbilt University who had only a month before won a Nobel Prize for his discovery of cyclic AMPs, the chemcial in cells that acts as messenger between hormones and organs, the signal that transmits instructions from hormone to cell. Dr. Minton, who is with the Department of Surgery and Physiological Chemistry at Ohio State University College of Medicine, found himself thinking quite a lot about cyclic AMPs (and their cousin-chemicals, called cyclic GMPs).

From biopsied tissue he discovered that the level of these substances was highest in cancerous tissue, next highest in tissue from patients with benign breast disease and lowest in normal tissue. Other studies showed that the substance which normally breaks down cyclic AMP and GMP after they have finished delivering their assigned message to cell or organ, is blocked by still another group of chemicals commonly found in some popular foods and beverages. These are called methylxanthines and include caffeine, theophylline and theobromine. They are found in coffee (even so-called decaffeinated coffee), tea (even herbal teas), cola drinks and chocolate. They are also found in a number of over-the-counter analgesics and cold remedies.

These substances, in effect, keep the signals in the cells from being turned off, Dr. Minton and his associates theorized, perhaps stimulating cells to do things they weren't supposed to.Like creating fibrous cysts, perhaps, or even cancers . . .

Carolyn Magin heard about a link between coffee, tea, colas and chocolate and benign breast disease early last spring. She is, she notes, "not a heavy coffee and tea drinker, drinks no Cokes and eats little chocolate."

So in May, she simply eliminated them from her diet.

In a study conducted at Ohio State, and published last year, Dr. Minton found distinct improvement in a majority of patients with benign breast disease who abstained totally from methylxanthine-containing products. He also found that women who smoked had distinctly poorer results, but when they stopped smoking as well, the benefits of the regimen increased appreciably. Younger women had quicker results, the study showed. (Many of the women had initial withdrawal symptoms from the methylxanthine, mostly in the form of headaches lasting as long as a week.)

This study and subsequent research, Dr. Minton said in an interview, led him to conclude a "clear relationship" between the consumption of methylxanthine and breast disease, more clear-cut in some cases than others.

In addition, his research showed a link to smoking.

Other recent studies -- showing that Mormons and Seventh Day Adventists who neither smoke nor consume many of the methylxanthine-containing drinks, have, as a group, fewer cancers than the population as a whole -- tend to support Minton's findings.

When Carolyn Magin went for her semi-annual thermogram in August, the technician exclaimed over the apparent "cooling" of her cystic disease as delineated in this photogrpahic technique. Only then did she tell the technician about her dietary experiment. The technician told the doctors, and they compared her thermograms from previous years. There was an apparent tend toward improvement. They are still watching.

No one, of course, with a breast lump should under any circumstances try to see if it will go away by itself, or try to make it do so with dietary changes. Anyone who finds a lump in her breast should see a doctor immediately. Caroly Magin was under more or less constant medical surveillance, so she ran no risk in putting herself on a methylxanthine-free regimen. And Dr. Minton says he has received literally hundreds of letters and calls from all over the country recounting often "amazing" results.

He notes, however, that when consumption of the products is resumed, the disease returns.

He does not say that methylxanthine causes cancer or that its elimination from the diet prevents it. But in the report on his study he writes: "The possibility that methylxanthine consumption is related to the development of malignacy should be considered . . ."

Carolyn Magin is due for another checkup in a month or so. She is hopeful. "I like my life," she says. "I want to keep living it."