Reading a Dense-Breast Notification: What the DENSE Trial Can and Cannot Tell You

A dense-breast notification tells you something about the texture of your tissue on a mammogram, not that anything is wrong. Since September 2024, every mammography report in the United States carries a standardized note about breast density, and for a lot of women that note lands as a puzzle: it flags a fact but stops short of saying what to do next. The best randomized answer to the "what next" question comes from the DENSE trial, reported in the New England Journal of Medicine in 2019. In women with extremely dense breasts and a normal mammogram, being invited for a supplemental MRI cut roughly in half the cancers that would otherwise surface between screening rounds, and it did so at the cost of more recalls, more false alarms, and more benign biopsies.

Key points#

• A density note reports how tissue looks on imaging. It is disclosure, not a diagnosis and not an order for more tests.
• The DENSE trial randomized women with the densest breasts to be invited for a screening MRI or to stay with mammography alone.
• Adding MRI roughly halved interval cancers (from about 5.0 to 2.5 per 1000 screenings).
• The same MRI produced a high false-positive rate, and most women it flagged did not have cancer.
• DENSE studied category D breasts only, so its numbers do not automatically apply to everyone who receives a letter.

What the notification is, and what it is not#

The Food and Drug Administration amended the Mammography Quality Standards Act in a final rule dated March 10, 2023, with facilities required to comply from September 10, 2024. Every patient now receives one of two standardized statements. Women with dense tissue are told that dense breasts can make cancers harder to see on a mammogram, that other imaging may help find them, and that they should talk over their own situation with a clinician.

Two things about that wording are worth slowing down for. First, it states a fact that holds across a whole population, not a personal verdict; a density note does not say that a particular woman carries above-average risk that demands action. Second, the rule requires that the information be shared, not that any test be ordered. It is neither a recommendation for extra screening nor a clean bill of health. The central reading error is to treat the letter as an instruction rather than as one input into a conversation.

The question DENSE was built to answer#

Most arguments for extra imaging lean on a simple count: how many additional cancers did the test find? That number always looks good, because any test that produces more images will flag more spots, whether or not chasing them helps anyone. DENSE was designed to sidestep that trap. Instead of rewarding detection for its own sake, it asked whether the extra scan changes what the mammogram actually misses.

The trial ran across multiple centers in the Netherlands and enrolled 40,373 women aged 50 to 75 who had extremely dense tissue and a normal screening mammogram. Participants were assigned in a 1-to-4 ratio either to be invited for a supplemental MRI or to continue with mammography alone. The main outcome was the interval-cancer rate: cancers diagnosed in the roughly two years between routine mammograms. Those cancers matter because they usually announce themselves through symptoms rather than screening, and on average they behave more aggressively than tumors caught on a scheduled scan. Randomizing tens of thousands of women to settle a screening question is uncommon, and it is what gives DENSE its weight.

The numbers, both sides at once#

On the benefit side, the interval-cancer rate was 2.5 per 1000 screenings among women invited for MRI, against 5.0 per 1000 in the mammography-only group. That is a gap of 2.5 per 1000 (95 percent confidence interval 1.0 to 3.7), which reads as roughly a halving. Notice that the absolute difference is far smaller than the word "halved" suggests, the familiar pattern whenever the underlying event is already rare.

The effect depended heavily on whether women used the scan they were offered. Only 59 percent of those invited actually had the MRI. Among the women who went through with it, the interval-cancer rate dropped to 0.8 per 1000; among those who declined, it stayed at 4.9 per 1000, essentially the same as the control group. The MRI detected cancer in 16.5 per 1000 women screened, so it was genuinely turning up tumors the mammogram had passed over.

The cost is the part the phrase "other imaging may help" glosses over. The false-positive rate was 79.8 per 1000 screenings. Of the women recalled after MRI, only 17.4 percent were found to have cancer, and of those sent on to biopsy, only 26.3 percent did. Put plainly, most women the first-round MRI flagged were healthy, and many went through more imaging or a needle biopsy to confirm it. Serious harms during or right after the scan were rare, around 0.1 percent, but the anxiety and the follow-up procedures were common.

Who the trial actually studied#

DENSE looked at women in the densest category, category D, which is about the top tenth of tissue density. It did not study the much larger group with heterogeneously dense category C breasts, who also receive the density note. It also did not pit MRI against ultrasound or against newer tomosynthesis. So the numbers above describe a specific, narrow population. Stretching them onto every woman who opens a density letter pushes the evidence past what was measured.

There is a time dimension too. DENSE reported results from the first screening round over a two-year window. Whether cutting interval cancers eventually means fewer breast-cancer deaths is a longer question that a single round cannot settle, and later DENSE rounds showed the extra yield falling once the initial backlog of hidden cancers had been cleared.

The bottom line for a real decision#

DENSE is best read as a ledger, not a verdict. On one side sit fewer interval cancers and a credible path to catching aggressive tumors sooner. On the other sit false positives, benign biopsies, the stress of a recall, a contrast injection, and cost. Good randomized evidence quantified both columns instead of advertising only the flattering one. Whether supplemental screening is worth it for a given person turns on her personal risk, how she weighs a false alarm against a missed cancer, and what she can access, and that is a conversation to have with a clinician who knows the full history rather than a question the notification answers on its own.