Mammograms catch the majority of breast cancers, but they aren’t perfect. In the general population, screening mammography detects roughly 85% of breast cancers, though that number drops significantly for women with dense breast tissue, where sensitivity falls to around 62%. About 9 out of every 100 screening mammograms result in a callback for additional imaging, and most of those turn out to be nothing serious.
What the Numbers Actually Mean
Two key metrics define mammogram accuracy: sensitivity (how well it catches real cancers) and specificity (how well it avoids false alarms). For the general screening population, sensitivity typically falls between 80% and 90%, meaning mammograms miss somewhere between 1 in 10 and 1 in 5 breast cancers. Specificity runs higher, generally above 90%, which means the vast majority of women without cancer receive a correct “all clear” result.
In practical terms, about 8.8 out of every 100 screening mammograms lead to a recall for additional imaging. Of those women called back, roughly 17% end up needing a biopsy. The rest are cleared with extra imaging alone. So while getting a callback is understandably stressful, the odds strongly favor a benign outcome.
Why Dense Breasts Change Everything
Breast density is the single biggest factor affecting mammogram accuracy. Dense breast tissue appears white on a mammogram, and so do tumors, which makes cancers harder to spot against that background. In a study of 240 women with dense breasts (the two highest density categories), mammography caught only 61.8% of cancers and missed 26 out of 68. Specificity remained relatively high at 91.9%, but the drop in detection is significant.
Nearly half of all women have dense breasts, and most states now require that mammography facilities notify you if your tissue falls into this category. If you have dense breasts, supplemental screening with ultrasound or MRI can catch cancers that mammograms miss. The combination of mammography plus ultrasound consistently outperforms mammography alone in this group.
How Age Affects Accuracy
Mammograms become more accurate as you get older, largely because breast density decreases with age. Younger women in their 40s tend to have denser tissue, which leads to more false positives and more missed cancers. On average, a woman who starts annual screening at age 40 will have one false-positive recall every 10 years. Starting at age 50 pushes that to about one every 11.5 years. The difference seems small, but over a decade of screening it adds up to more unnecessary callbacks, biopsies, and anxiety for younger women.
The current U.S. Preventive Services Task Force recommendation is biennial (every two years) screening mammography for women aged 40 through 74. This applies to those at average risk. Women with a family history of breast cancer, known genetic mutations, or prior chest radiation may need to start earlier or screen more frequently.
3D Versus 2D Mammography
Three-dimensional mammography, also called tomosynthesis, takes multiple X-ray images of the breast from different angles and assembles them into a layered picture. This helps radiologists see through overlapping tissue that might hide a tumor on a standard 2D image.
In a large registry study from the Chicago area, 3D mammography had the lowest recall rate of any modality at 9.4%, compared to 10.8% for standard digital mammography. Cancer detection rates were similar (0.40% for 3D versus 0.42% for 2D digital), but 3D achieved those detection numbers while flagging fewer women unnecessarily. Published research supports that 3D tomosynthesis offers greater specificity overall. Most major breast centers now use 3D as the default screening method.
What AI Screening Adds
Artificial intelligence is now being used alongside radiologists to read mammograms in several countries. A large-scale real-world study in Sweden, covering over 460,000 women screened between 2021 and 2023, found that AI-supported reading detected 17.6% more cancers per 1,000 women screened compared to standard radiologist reading. The AI group found 6.7 cancers per 1,000 women versus 5.7 in the control group.
Notably, the AI group achieved this improvement without increasing false alarms. Recall rates actually dropped slightly (37.4 per 1,000 versus 38.3), and the positive predictive value of those recalls rose from 14.9% to 17.9%. In other words, when AI flagged something, it was more likely to be a real cancer. AI also reduced radiologist workload by triaging low-risk mammograms, allowing doctors to focus their attention on the cases most likely to need it.
Why Mammograms Miss Some Cancers
A false-negative mammogram means a cancer was present but the screening didn’t catch it. One study found a false-negative rate of about 5.1%, and the reasons broke down into four categories. The most common, responsible for about 39% of missed cases, involved the tumor itself: cancers that were extremely small (under 3 millimeters), cancers that mimicked normal tissue patterns, or cancers spread across multiple locations in the breast.
Patient-related factors accounted for about 27% of misses, primarily dense breast tissue but also difficulty holding still during the exam. Technical issues like improper breast positioning caused about 15% of missed cases. The remaining 19% came down to the radiologist, either failing to notice a subtle abnormality or misinterpreting what they saw. This is one reason many screening programs use double reading (two radiologists reviewing each image) or AI-assisted review.
The Overdiagnosis Question
Mammograms can also be “too accurate” in a specific sense. Overdiagnosis occurs when screening detects a real cancer that would never have grown large enough to cause symptoms or threaten your life. These slow-growing or dormant cancers still get treated, often with surgery, radiation, or medication, even though they may never have needed it.
Overdiagnosis rates increase substantially with age. A National Cancer Institute analysis estimated that among women aged 70 to 74 diagnosed through screening, about 31% represented overdiagnosis. For women aged 75 to 84, that figure rose to 47%. For women with a life expectancy under five years, more than half of screen-detected cancers were estimated to be overdiagnosed. This is a key reason why screening recommendations for women over 74 remain uncertain, and why the decision to continue screening at older ages involves weighing potential harms against benefits.
For women in the core screening age range of 40 to 74, the balance of evidence still favors regular mammography. The cancers caught early through screening are more treatable and less likely to require aggressive therapy. But no screening test is flawless, and understanding the real numbers helps you interpret your results with the right expectations.