Soon after Pennsylvania's breast density notification law took effect in 2014, Jules Sumkin found himself wanting to spare women from getting a letter that might alarm or perplex them.
Twenty-eight states, including New Jersey and Delaware, now have laws that require mammography centers to inform women with dense breast tissue that it may increase the risk of cancer and obscure a malignancy on a mammogram, so they may want to talk to their doctors about extra imaging options.
The letters don't mention quandaries that Sumkin, chair of radiology at the University of Pittsburgh Medical Center, knows too well: There are no guidelines for extra imaging, or any evidence that it saves lives. And the assessment of density is often a matter of opinion.
When evaluating mammograms, radiologists usually concur on the American College of Radiology's highest and lowest ratings - extremely dense and mostly fatty. But 80 percent of women fall in the murky middle, with either scattered density or more uniform "heterogeneous" density. Under the laws, women with extreme or heterogenous density get notified, but not those with scattered density.
"I felt myself leaning toward grading two [scattered density] if I was on the edge," recalled Sumkin. "A law was influencing how I was interpreting a mammogram."
He had company: A study he co-authored found that half of his 16 colleagues thought they were classifying more women as scattered density as a result of the law. Yet 90 percent actually were doing so.
The breast density movement, launched seven years ago with a Connecticut law, has been driven by cancer survivors who learned they had dense breasts only after their malignancies were missed on mammograms. They - and politicians - portray the laws as empowering women with information and saving lives.
But now, it looks as if the skeptics - basically, the entire medical community - were right to be leery.
Not only are density assessments even more subjective than experts believed, but women also are getting conflicting advice about what to do. Supplementary screening is costly and inconsistently performed, and increases false alarms.
And while more screening finds more cancer, experts say this may fuel "overdiagnosis and overtreatment" - eradicating cancers that would never cause harm if left undetected.
In an editorial last year, Jennifer S. Haas, an internal medicine physician at Harvard Medical School, said the laws, although well-intentioned, "create the unsubstantiated anticipation that additional testing is better for women. . . . Instead, the laws may result in substantial personal harms and societal costs."
Mammography is the only screening proven to reduce breast cancer deaths, but it's far from perfect. Fatty tissue appears dark and transparent, while dense, or fibroglandular, tissue appears white and opaque - a potential mask for a malignancy.
Kathryn Thomas, 65, of Harrisburg, lobbied for Pennsylvania's law with the wrenching story of her metastatic cancer diagnosis. "My hope," she said then, "is to prevent another woman from going through my personal tragedy - 'normal' mammograms yet a hidden invasive cancer."
But density by itself "is not a major cancer risk factor," says a group of California radiologists and cancer specialists who created a website, www.breastdensity.info, in response to that state's law.
Extreme density doubles the average risk of breast cancer. For example, having extremely dense tissue raises an average 50-year-old woman's chance of being diagnosed in the next 10 years from 2.4 percent to 4.8 percent, which means she is more than 95 percent likely to be fine.
Extreme density is found in fewer than 10 percent of women, according to a study using national databases. For the 40 percent of women with the far more common heterogeneous density, it barely nudges their risk, studies suggest.
And again, density ratings are not precise. A new study of 30 breast imaging centers across the United States, including the University of Pennsylvania's, found vast variation in radiologists' perceptions.
Among 83 radiologists who read at least 500 screening mammograms between 2011 and 2013, the percentage of women they deemed to be dense ranged from about 6 percent to 85 percent - a 14-fold difference.
"I think it begs for a standardized quantitative technique," said Emily Conant, chief of breast imaging at Penn and a co-author of the study, published last month in Annals of Internal Medicine. "Women should be told density is a subjective judgment."
Density letters tell women to talk to their doctors about what to do.
"There's no doubt we get frequent calls from women asking, "What the heck is going on with my mammogram report?' " said Mark Finnegan, an ob-gyn at Lankenau Medical Center. "Invariably, it's the density rating. And then I have to go through the whole spiel that there is no standard guidance."
The extra screening options are an ultrasound, an MRI, or 3-D mammograms, but insurance generally won't cover these when mammograms are normal.
A number of studies, including by the American College of Radiology Imaging Network, have found that adding ultrasound and MRI to mammography detects more - but still not all - cancers. Intensified screening also increases recalls for diagnostic imaging and biopsies that find no malignancies.
In New Jersey - one of only four states that require insurance coverage of extra screening - breast center radiologists at St. Barnabas Medical Center in Essex County compared the 18 months before and 18 months after the 2014 law took effect. The study, published in June, found that ultrasound usage soared by 650 percent (to 11,486 screens), MRI usage jumped 60 percent (to 4,134), while the number of cancers detected rose 5 percent (to 621).
Radiologist Linda Sanders, who directs the breast center and led the study, was surprised to see ultrasounds spiked even among patients without dense breasts, which she and her co-authors said "suggests clinicians feel obligated to order extra imaging to prevent legal liability."
Sanders' study did not look at how many women were recalled for what was ultimately deemed a false alarm. However, many women who were recalled underwent an MRI to rule out cancer, an approach that slightly reduced the rate of biopsies needed to find a malignancy.
Still, that illustrates another area of subjectivity: radiologists' technological preferences.
Sanders thinks MRI is much better than ultrasound both for ruling out cancer and diagnosing it.
Because MRI costs more than $1,000 per exam, current guidelines recommend it to supplement mammograms only for women at very high risk of breast cancer, such as those with a cancer-promoting gene defect.
Sanders has concerns, however, about an emerging approach, so-called mini-MRI, because it abbreviates the standard exam, eliminating imaging that might be crucial to assess some cases.
At Penn, meanwhile, Conant and her colleagues believe mini-MRI can be effective - and more affordable for women paying out of pocket.
"It's what we're promoting for women who can't get the full MRI but want supplemental screening," Conant said. "We're price pointing it at $299, the same as an ultrasound."
Another innovation touted by some centers is an automated ultrasound system. It replaces the hand-held transducer with a tent-like version that fits over the breast, making the exams faster and more consistent. Aria Health in Northeast Philadelphia was an early adopter.
Although the recall rate with the system has been high - about 20 percent - getting insurance coverage has not been a problem, said radiologist Sonja Cerra-Gilch, Aria's chief of breast imaging.
At Pitt, women with dense breasts are advised to have newer 3D mammography, or tomosynthesis, Sumkin said. It is rapidly becoming standard because studies show it improves detection while reducing false alarms.
Pitt is also studying tomosynthesis plus ultrasound for dense-breasted women.
"Because we really don't know what to do," Sumkin said.