Among the more important medical studies reported in recent weeks, even resulting in that rarefied "front page, above the crease" coverage by the New York Times, was a paper in JAMA indicating that the interpretation of breast biopsies is not the infallible gold standard we had hoped. The investigators found that expert pathologists often reached differing conclusions about the same biopsy sample when they reviewed it independently.
This finding has the potential to add consternation to controversy, if not quite insult to injury, as breast cancer screening succumbs to all manner of reconsideration. There has long been debate, and conflicting evidence, about the utility, impact, optimal timing, and optimal frequency of mammography. There have been concerns about the inadequacies of standard mammography when breast tissue is dense. Self-breast examination, long recommended, no longer is- although I, along with many of my colleagues, know patients who ascribe early breast cancer detection to the practice.
The lead author of the paper in JAMA, Dr. Joann Elmore, is a leading expert on breast cancer screening, publishing numerous papers that have challenged, and thereby improved, the standard of practice for more than two decades. She also happens to be a friend and colleague; we have co-authored four editions of an epidemiology/preventive medicine textbook together.
So while I generally sort out the implications of new studies on my own, I don't mind calling in the cavalry when I have that opportunity. I asked Dr. Elmore where her new study leaves us, and here is our exchange. The questions, of course, are mine, the answers are hers -- with the occasional editorial comment from me in brackets.
Q1: What are the key findings and implications of your new study?
Every year millions of women undergo breast biopsies. As a primary care physician, I depend on the pathologist to look at the tissue under a microscope and provide a diagnosis so that I know how to advise and care for the woman. In our study, we turned the microscope back on ourselves to see how well we are doing in classifying these biopsies as normal, abnormal, or cancerous.
We invited 115 doctors to interpret the same biopsy cases and compared their interpretations to a diagnosis decided upon by three very experienced breast pathologists. We had 240 biopsy cases and collected 6,900 individual interpretations.
It was reassuring that the pathologists had very high agreement on the breast cancer cases. Unfortunately, they were much less likely to agree with the expert panels the rest of the time. For cases of ductal carcinoma in situ (DCIS), a very early stage of localized cancer, they disagreed one out of five times. For cases of atypia (or atypical ductal hyperplasia), meaning cells were abnormal but not cancerous per se, they disagreed on roughly half of the cases.
Women who receive a diagnosis of DCIS usually undergo the same types of treatment as women with early stage breast cancer (mastectomy, or lumpectomy and radiation), so these are important diagnoses. Women who receive a diagnosis of atypia are told that they are at increased risk of a subsequent breast cancer diagnosis and often undergo additional biopsies and sometimes more intense screening.
It is important to note that even the experienced pathologists disagreed on some of the biopsy cases. This problem may have more to do with the biology of the disease than the doctor making the diagnosis. Some biopsy specimens are just really challenging to characterize.
Q2: You have published prior papers that questioned the reliability of breast cancer screening with mammograms. Where does that research leave us?
Variability among doctors is not confined to the interpretation of biopsy specimens. We noted extensive disagreement of radiologists who interpreted mammograms many years ago. After our research was published, legislation was passed in the U.S. to address this concern (the Mammography Quality Standards Act). The American College of Radiology also developed a standardized classification system for mammography reports and NIH funded a national Breast Cancer Surveillance Consortium that has gathered and evaluated data on millions of mammogram exams.
[Editorial note: The breast cancer screening recommendations of the United States Preventive Services Task Force are currently being updated; the report-in-progress may be found here.]
Our work in breast pathology is an important benchmark as it identifies another area where we need to do better.
Q3: What would you consider the 'ideal' breast cancer screening protocol?
An "ideal" breast cancer screening protocol is one that starts with informing women about the benefits and risks of screening. Every woman should make a decision that is right for her.
Many women incorrectly think that their personal risk of a breast cancer diagnosis is higher than it really is - thus I encourage them to go on-line and enter information into a "risk calculator." I just entered my own data and the website says that my risk is 2.5% in the next 5 years (this also means that I have a 97.5% chance of NOT getting breast cancer in the next 5 years).
Mammography is the only breast screening examination with data showing a potential mortality benefit when applied to the general population of women. We do not recommend that the average woman get MRI scans or ultrasound exams for screening - we do not have data to support this.
Risks of mammography include over-diagnosing lesions that would never have harmed the woman, yet wind up being treated; in such cases, the "cure" is worse than the disease was ever going to be. Other risks include false alarms where an abnormality is noted, and the woman has to return for additional testing (usually more imaging, but sometimes a biopsy). These false alarms occur in half the women who undergo annual screening for a decade.
Some women are learning that they have dense breast tissue [Editorial note: I have written a prior column on that topic.] State law mandates that women be advised about their breast density in ~20 states and a few states even mandate that women be advised to have ultrasound testing (even though this recommendation is not clearly evidence-based at this point).
Q4: What is your standard recommendation to patients about breast cancer screening?
Get the facts and make a decision that is right for you. Every woman has different values and concerns and priorities.
Now that we have the results of our study on breast biopsies I am encouraging women to think twice before undergoing a breast biopsy. When a very subtle abnormality is noted on a screening mammogram women are often presented with the option of a biopsy vs. waiting 6 to 12 months and just repeating the mammogram to see if the abnormality changes. Many women assume that the biopsy will provide an immediate answer; our study shows that may not be the case, as these biopsies are difficult to evaluate.
Q5: When I talk to patients about prostate cancer screening, I can't ignore the fact that I am a >50 year old guy with a prostate of his own; so how I apply what I know to myself is relevant. I do not get screened for prostate cancer (although of course I do for colon cancer). What is your personal breast cancer screening approach?
A personal decision that I make about breast cancer screening may not be relevant to others.
Q6: Slightly off topic: Angelina Jolie just followed up her prophylactic mastectomy with a prophylactic salpingo-oophorectomy (i.e., surgical removal of ovaries and fallopian tubes). As an epidemiologist, a cancer screening expert, and a woman- would you have done the same in her position? If not, what would you have done?
We need to clarify that Angelina Jolie's position was very unusual. She found out that she is a carrier for a genetic mutation, the BRCA-1 mutation. The majority of women in the U.S. do NOT have this genetic mutation. The majority of women with breast cancer do NOT have this genetic mutation.
Ms. Jolie gathered data, reflected on her own values and she made an informed decision. I would do the same.
There you have it, straight from the source. I can tell you that my wife does undergo mammography at the standard intervals (and, by the way, she tells me it is VERY uncomfortable, which is something that isn't discussed much). We have had to deal with minor abnormalities requiring additional imaging (and generating anxiety, of course)- but so far, have not confronted the biopsy issue.
Cancer screening is, in the most literal sense, looking for trouble. Significant declines in cancer mortality over recent years are a product of both improved treatment, and early detection- so we are clearly doing something right. On the other hand, when our expectations and actual experience diverge, trouble may be exactly what we find. I suppose April 1 is a particularly good day for a reality check, and making sure we don't set off on a fool's errand.
Dr. Elmore's work has improved breast cancer screening standards over the years, for which we all owe her a debt. The new study is likely to do the same in time, but for now- we are forewarned of the fallibility of biopsy assessment, and thus forearmed to make judicious, personalized decisions accordingly. We are also invited to second-guess a biopsy report, and request a second opinion in those cases where interpretation is most apt to vary.
Director, Yale University Prevention Research Center; Griffin Hospital
Editor-in-Chief, Childhood Obesity
President, American College of Lifestyle Medicine
Founder, The GLiMMER Initiative
Author: Disease Proof