Find a breast cancer tumor when it is tiny, and a woman will probably beat the disease. Find that same malignancy when it is larger or has spread to other organs, and she is far more likely to die, even after surgery, radiation and chemotherapy. Finding breast tumors before they turn deadly is a challenge and one that medical technology has so far failed to master.

“We desperately need better breast cancer screening tools,” says Otis Webb Brawley, chief medical officer at the American Cancer Society. His organization promotes mammography in an effort to reduce the 40,000 deaths from breast cancer every year in the U.S. But that emphasis, Brawley fears, leads engineers and medical device manufacturers to presume that the problem of breast cancer detection is not worth their attention, because it has been solved. It has not. Mammograms miss up to 20 percent of tumors, and an average of one out of 10 readings mistakenly identifies healthy breast tissue as possibly malignant. Those false positives mean that women who try to do the right thing by going in for routine cancer screening face a substantial risk of needless biopsies (which can themselves be disfiguring and interfere with treatment later on) and expense, as well as the misplaced fear that they have cancer when they really do not.

Mammography’s shortcomings have spawned controversy and confusion. In 2009 the U.S. Preventive Services Task Force (USPSTF) determined that routine mammograms would save too few lives of women ages 40 to 49 to justify the number of false positives and unnecessary biopsies that would result in that age group. Medical societies and patient advocacy groups attacked the recommendation; the American Cancer Society still advises women in their 40s to undergo mammography every year. Some health experts fret, though, that the USPSTF finding has discouraged more than a few women, not just those in their 40s, from getting tested. “It’s made women more skeptical about the test,” says Sheryl Gabram-Mendola, a surgical oncologist at Emory University’s Winship Cancer Institute. “Women say, ‘I’m just not going to do it, I’m too busy.’”

Even before the 2009 controversy, however, women were forgoing screening mammograms. According to the most recent data from the U.S. Centers for Disease Control and Prevention, the percentage of women who have undergone mammograms within the previous two years has dropped a bit across all age groups. For women 40 to 49, the number fell from 64.3 percent in 2000 to 61.5 percent in 2008—which makes sense if mammograms are unhelpful as screening tests in that age group. For women 50 to 64, it fell from 78.7 percent in 2000 to 74.2 percent in 2008—which is worrisome given that the evidence shows mammograms are clearly beneficial for women 50 and older.

Better tools could help encourage screening and make it more useful for women of all ages. No method currently under study is robust enough yet to supplant mammography. But researchers and clinicians hope that a greater understanding of the physiology and biochemistry of breast cancer, combined with more dexterous technology, will one day result in screening tools that can replace or inexpensively supplement mammograms so that the results will be more trustworthy.

Today’s Options
Mammography, which has been used to detect tumors since the 1970s, misses some cancers and wrongly suggests the presence of others because it is based on low-dose x-rays, which have ­inherent limitations in their ability to resolve tumors. When viewed with x-rays, a malignancy appears lighter in color than does normal breast tissue. X-rays also pick up white specks of calcium deposits that may be generated by a tumor as it grows. But x-rays are not good at spotting tiny tumors, partly because of a lack of contrast and partly because the calcifications associated with tumors are much smaller than benign deposits and therefore easy to overlook. Nor are x-rays good at detecting tumors in dense breast tissue, which also reads as white; many women younger than 50 have dense breasts. Finally, mammograms cannot indicate for sure whether an unusual mass is cancerous.

5 Comments

1. 1. dlhahn 11:42 AM 4/23/11

   Beyond Mammograms. Conventional breast cancer screening tests are far from perfect.
   
   Some, perhaps many, women of any age who are fully informed about the benefits and the harms of mammography may rationally decline mammography screening for breast cancer. The body’s immune system continuously searches out and destroys cancer cells, including breast cancer cells. One third of breast cancers that are detected by mammography would never cause symptoms, let alone death. This over detection by screening is called “over diagnosis.” Over diagnosis, not false positive or false negative mammograms, is the most harmful aspect of all forms of cancer screening, including mammography. If 2000 women are screened by mammography over 10 years, one will have her life prolonged, and 10 others, who would not have otherwise been diagnosed with breast cancer, will have a breast cancer diagnosed and treated unnecessarily. In other words, 1 woman will avoid dying of breast cancer. Ten other healthy women will be treated unnecessarily, 4 of these will have a breast removed, 6 will receive breast conserving therapy, and most will receive radiation therapy.
   
   Detecting breast cancer cells at earlier and earlier stages will not necessarily translate into more overall value for women. Earlier detection will almost certainly increase the amount of over diagnosis. Whether earlier detection will translate into any additional benefit in terms of avoiding death from breast cancer is an open question. A better long term approach to preventing death from breast cancer is more basic research into cancer mechanisms leading to better treatments. After all, if there were a perfect treatment for breast cancer after it became symptomatic, there would be no justification for screening. That would save many women from the harm of over diagnosis. At least half or more of the decline in breast cancer mortality recently can be attributed to better treatments, not to better screening.
   
   For your readers, the bottom line should be: always ask about the harms, as well as the benefits, of any test proffered by the medical profession. Then make an informed choice based on your own personal values. You cannot depend on the medical profession to do this for you, because we (I am a physician) almost always emphasize the benefits and discount the harms of what we are selling. Caveat emptor.
   

   Reply | Report Abuse | Link to this

2. 2. friendofscience11 10:39 PM 4/28/11

   Thanks to the Dr. making the comment below. I decline mammograms because of the reasons she/he states. In addition, I am guessing mammograms are riskier for some women than others because of the potential damage to tissue from "mashing the breasts" into the clamps to get the image. I bruise very easily, and am aware that tissue damage can contribute to cancer. So it doesn't seem like a good idea to use a test that is harsh on the breast tissue, unless there is really good evidence of benefit. The author of this article doesn't mention THERMOGRAMS at all. My understanding is that this method of detection, while not perfect, is overall better than mammograms as a screening tool, if done by a well-trained person. Thermograms are completely non-invasive, since they measure the heat radiating from the body. Cancer tissue is "hotter" since it requires a larger blood supply than normal healthy tissue. My understanding is that thermograms are used more widely in Europe, but not very well known in the US at this time. So I would like to see this method included in the discussion. I get thermograms regularly.

   Reply | Report Abuse | Link to this

3. 3. pshaffer 02:16 PM 5/1/11

   I am a radiologist who spends 90% of his time in breast imaging. There are some points on which I would disagree with the author. She states that biopsies can be disfiguring and can cause problems with later diagnosis. I would agree, if the biopsy were surgical, however, in any modern medical practice the standard is needle biopsy, which is certainly not disfiguring, and cannot cause difficulty in later diagnosis. Further, they are not particularly uncomfortable. Most patients agree that they are not nearly as uncomfortable as a dental procedure.
   MRI scans absolutely can tell cysts from solid masses in the breast, but do have difficulty because of overlap between various benign conditions (such as fibroadenomas, or, at times, simply hormone stimulated normal tissue) and cancers. The principle that MR uses is that of increased blood flow to malignant tumors, and this can be seen with other non-malignant pathologies.
   Ultrasound can distinguish some forms of benign tumors from cancer, but, like mammography and MR, there is overlap, with some benign tumors resembling cancers. Even with this, though, often with ultrasound we can make a determination that a mass is "highly likely" to be benign. These patients have no further work-up.(Note the hedge: "highly likely". We know from experience that sometime, somewhere a cancer will arise that will resemble benign tissue.)
   I think of the imaging work-up of possible cancer as a sorting process: for every 1000 women who come for screening, about 70 will have something sufficiently worrisome to warrant more pictures. 60 of these will be found to be fine, 10 will require biopsy to "be sure". About 3 to 4 of these patients will have cancer, and most of these will be cures. About 1 will not be.
   Strikingly absent from the discussion is the fact that the false positive rate can be controlled. If I wanted, I could tomorrow reduce the number of patients I call back for more imaging. However, the statistics of the process are such that by doing so, I will necessarily increase my false negative rate. Our society does not tolerate a false negative. Plaintiffs attorneys enforce this. I have heard it said by colleagues that the cheapest malpractice insurance that one can purchase is to call back a patient for more images. A sad commentary, to be sure, but deadly accurate.
   I appreciate the inclusion of other methods that are being developed, but would point out that, inevitably, there will be overlape between malignant and benign with these methods, also.

   Reply | Report Abuse | Link to this

4. 4. pedrorandrade 09:37 AM 5/10/11

   I urge the publisher to take a look at Deborah Rhodes presentation at TED http://www.ted.com/talks/lang/eng/deborah_rhodes.html
   It is a HUGE breakthrough in visualization techniques for breast cancer discovery and should be regarded in this article.

   Reply | Report Abuse | Link to this

5. 5. BLBrian 10:56 AM 5/23/11

   In the 1980's I tried to develop with others an AI solution to reading mammograms. It was pretty hopeless, due to a number of reasons. The obvious ones that computers were pretty basic and graphics terrible. But I always had a the overwhelming problem that the exceedingly experienced radiographer who worked with us seemed able to see things that we, as lay people, just could not.(Something I have found in other industries). I just gave up disheartened and moved onto other things.
   But at a much later (circa 2000) a large amount of reported literature pointed out that London Taxi Drivers brains actually altered physically when learning the complex task of remembering roads in a big city.
   Similarly sportsmen practice for days on eye hand co-ordination tasks, presumably not just to muscle hone but to establish learning correction patterns.So that the brain knows how to hit the ball/target.
   A quick scout through the literature on mammograms shows some work on people comparing gold standard radiographers with normal clinicians, some of these showed minimal differences, but the samples were all very small and the images did not have artificially added problems.
   Is it possible that better training on a large number of high precision computer modified mammograms, each with subtly induced artificial deposits could lead to better diagnosis?
   PS I am no longer in this field.
   
   

   Reply | Report Abuse | Link to this