50,000 to 90,000 American Women Incorrectly Diagnosed with Breast Cancer Annually– A Story in the NYT That Women and Their Doctors Should Read

As regular Health Beat readers know, over the past three years, Naomi Freundlich, and I have written about the risks as well as the benefits of mammograms more than once on this blog. Nevertheless, I want to call attention to an outstanding New York Times story by Stephanie Saul that appeared on the front page of the Times Wednesday, July 19, taking an in-depth look at what can only be called a medical tragedy.  

Saul tells the tale of women who lost all or part of a breast, and in many cases suffered through radiation, for absolutely no reason.  This could happen to anyone. One of these women was a nurse.


I decided not to write a full post about Saul’s story because Health Beat has covered much (though far from all) of the same ground in the past.  But I do want to call attention to the NYT’s piece.  This is a story that women, and their primary care physicians should read.

The danger that a woman will lose all or part of a breast needlessly is growing. As Saul points out, “Advances in mammography and other advanced imaging technology over the past 30 years  have meant that pathologists must render opinions on ever-smaller breast lesions, some the size of a few gains of salt.”

Too often, pathologists at a small community hospital may not have had enough experience reading the slides. “Concerned about the accuracy of breast cancer pathology, the College of American Pathologists has said that it will set up a voluntary certification program for pathologists who read breast tissue,” Saul reports. “Among its requirements is that pathologists must read 250 breast cases a year.”

Just how often are women misdiagnosed? The Times runs a chart from the National Institute of Health (NIH) showing that diagnosis of non-invasive breast cancer was rare, until the 1980s, when mammograms become popular. Now, 50,000 women are diagnosed with these small lesions (ductal carcinoma in situ, or DCIS) each year. 

The Times also reports that: “In 2006, Susan G. Komen for the Cure, an influential cancer survivors’ organization, released a startling study. It estimated that in 90,000 cases, women who receive a diagnosis of either DCIS or invasive breast cancer did not have the disease or their pathologist made another error that resulted in incorrect treatment.

If you are going to comment on this post, please read the whole New York Times story first. I’m very sorry, but I just don’t have the time to repeat all of the evidence in this excellent, very long story while replying to those who didn’t read it.   (If you don’t have online access to the Times, try Googling the story, using Saul’s name, “New York Times” and the headline: “Prone to Error: Earliest Steps to Find Cancer.”  It should come up.)

21 thoughts on “50,000 to 90,000 American Women Incorrectly Diagnosed with Breast Cancer Annually– A Story in the NYT That Women and Their Doctors Should Read

  1. It’s suggested that unskilled pathologists come across as the problem. Cancer Medicine: Principle of Multidisciplinary Management (6th Edition) tells us that pathologic uncertainty is a shaky foundation on which to build therapeutic strategy. When doubt exists concerning the nature of a neoplasm, additional opinions are always appropriate.
    Pathology is a very visual science. It appeals to people who have a talent for recognizing patterns. This pattern recognition comes after years of practice. Cancers grow in recognizable patterns that allow for their identification. A breast cancer has a certain growth pattern that differs from a carcinoma of the lung. Benign conditions also have patterns.
    It is said that a pathologist will never make a diagnosis unless they are 200% sure of the diagnosis. Having said that, there are situations where a definitive diagnosis cannot be rendered. Sometimes it is because the biopsy sample that was taken by the surgeon is too small, or perhaps taken from an area that is not representative of the patient’s lesion.
    Physicians tend to settle on the smallest amount of tumor tissue possible, often with a fine needle aspirate that collects just a few cells, for biopsy analysis. Larger bore needles (tru-cut) are needed to perform core biopsies or even remove entire lymph nodes, so that they can collect enough “live” tissue to more reliably determine the histologic and molecular features of a cancer.
    Then there comes a time when a pathologist must admit that they do not know. Considering that the rarest of diseases pass under the pathologists’ microscope, this is not surprising. There are several diagnostic tests or special stains (immunohistochemistry) which the pathologist can turn to which may aid in the diagnosis.
    In a statistical analysis, the tentative diagnosis, the interpretation of stains and conclusions drawn from immunohistochemistry are independent factors in reaching a diagnosis. The immunohistochemical (IHC) staining test is performed on microscope slides, with intact cells and looks for proteins themselves.
    The cell-block technique is useful for IHC and can give morphological (structural) details by preserving (iin paraffin wax) the architectural patterns. However, according to cell function analysis, investigators can only measure those analytes (subtance or chemical constituent) in paraffin wax that they know to measure. If you are not aware of and capable of measuring a biologically relevant event, you cannot seek to detect it.
    Cell-blocks are paraffin-embedded and paraffin-embedded tissue can change over time. These proliferating populations of cells are biologically distinct in their behavior from “fresh” live cells that comprise human tumors.
    Because the results of the IHC test can sometimes be ambiguous, many doctors suggest the FISH (fluorescent in situ hybridization) test for a second opinion. However, there has been poor concordance in terms of FISH testing in a central laboratory compared to local laboratories, which the prevalent notion regarding FISH is that it is 100% accurate.
    They have yet to explore all the quality control issues of FISH. Several things can be done to improve performance and reduce variability. One thing is to train the interpreter. Another is to have the laboratory be certified. According to clinicians at the Mayo Clinic, oncologists need to be more aware of which laboratory performs the tests and who interprets the results, because it can make a huge difference.

  2. It’s alarming that so many doctors are removing large amounts of breast tissue for very minor cancers.
    Thankfully, this isn’t universal. My mother recently underwent surgery and radiation for DCIS, stage 0. However, in her case the doctors seemed to work hard to NOT give her misleading information. Rather than diagnosing cancer from the needle biopsy, they stated that there were abnormal cells and that they’d prefer to take them out just to be safe. Once they took them out (they removed only a very small amount of her breast), they determined cancer was present. Only then did they decide to do a short course of radiation (only 5 days, 2x/day). Later testing showed that the cancer was a fast-growing variety. It’s possible early detection and surgery may have saved her life.

  3. Lisa, if you read the NYTimes article you will see that it was about a women losing a large part of her breast NOT because she had cancer, but because she was misdiagnosed with cancer. Repeat: she never had cancer at all. This is much different than losing a large part of a breast to a small cancer.
    And I’m glad to hear that your Mom is doing well after her cancer treatment-all the best to her. But you should not necessarily conclude that her life was saved. It’s very likely that she never needed treatment in the first place, especially if her cancer was discovered via screening mammography. Because not only are many early cancers misdiagnosed, many early cancers are meaningless healthwise, even if they actually ARE cancers. This is called overdiagnosis. In reality it is 10 times more likely that your Mom never needed treatment than that her life was saved. Please see screening.dk

  4. Gregory & Lisa
    Several things seem clear.
    First laboratories should be certified and secondly, only pathologists who have passed certification should be reading biopsises.
    Secondly, this case it seems the pathologist had failed the certification test more than once. This would suggest that he does not have a special talent in this area and should not be specializing in pathology.
    I’m reminded of a story Kaiser Permanente in Colorado told me when I was writing the book. When they looked at the results of mammograms they realized that some of the Kaiser doctors reading mammograms were getting more false positives and false negatives than others. So they began tracking the doctors and found that that, indeed, results varied widely. Some doctors just weren’t as good as others.
    Kaiser decided that the docs on the left side of the bell curve shouldn’t be reading mammograms. They offered to re-train them to do something else.
    Most took them up on the offer; one doctor was angry at being “judged” and he left.
    This was fine with Kaiser–they put patients ahead of doctors.
    Though I fear that the doctor who left is working at a small community hospital somewhere, reading mammograms . . .
    It seems to me this is a situation where we cannot afford to have less than very good pathologists reading biopsises. They need to be certified (and this should not be voluntary) but someone also needs to be looking over their shoulder, keeping track of their record. This is the advantage of having doctors working for a large accountable care organization.
    Secondly, women would be much better off going to oncologists who work at large hospitals where the pathologists in the lab are seeing hundreds of biopsies and have plenty of practice.
    As I recall, you once mentioned that oncologists at your community hospital failed to diagnose your wife’s cancer at least once if not twice.
    In general, the more practice doctors have doing anything (surgery, reading tests) the better they become. This is not always the case. But it’s a rule of thumb that I, as a woman, would keep in mind if someone diagonosed be with cancer.
    Before I did anything I would go to an oncologist at a large urban hospital for a second opinion.
    Finally, women need to recognize that they have choices. If a tiny lesion is found, women need to realize that as Dr.Juliet K. Mavromatis points out: “DCIS represents twenty percent of malignancy detected by mammography. Ninety percent of women in which this condition is detected are asymptomatic at the time of diagnosis. Longitudinal studies of the natural history of DCIS in untreated women suggest that 15 to 60 percent will develop breast cancer in the affected breast after 10 years. This is a broad range and at this point it is not well-understood what factors cause breast cancer to develop in some women with DCIS, while cancerous changes to regress in others.”
    Mavromatis offers a useful discussion of cancer and “watchful waiting” on her website here http://www.drdialogue.com/2010/03/is-watchful-waiting-too-difficult.html
    She adds: “With medicine’s current focus on early detection and the abundance of information that it may provide, it becomes increasingly important to make sure that our remedies are not worse than our diseases. After all, as much as we may not like to hear it, we are all diseased, and in effect, pre-cancerous.”
    This is why the U.S. Preventive Services Task Force now recommends that, for most average-risk women, mammograms are appropriate only after age 50, when the risk of invasive cancer is higher. Dr K. expands: “In the case of breast cancer, the United States Preventive Services Task Force recently published its revised guidelines for breast cancer screening suggesting that mammography screening be delayed in most women until age 50. These recommendations were in part based on the finding of “adverse effects” resulting from overzealous screening procedures. Although breast cancer screening in women ages 40 to 50 is known to be effective for early detection, its use is associated with the detection of a range of abnormalities of the breast, which lead to further evaluations including follow-up mammograms, MRIs and biopsies. Of course, these procedures are anxiety-provoking and costly. What’s more, pre-cancerous breast disease, as is true with other precancerous conditions, may not always progress to invasive cancer.”
    An NIH panel recently recommended that perhaps we should stop calling DCIS “carcinoma.”
    Lisa–I’m glad your mother survived, and I’m assuming she got a second opinion when she was told that it was a fast-growing cancer.
    AS Gregory’s comment indicates, disgnosing cancer, and figuring out which cancers will progress is not as cut and dried as one might think.

  5. If a patient is dx with DCIS, who is going to sit around and see if it invades now or later? I know of only a few invasive breast cancer types that are low grade. One would be tubular carcinoma. Now without a biopsy and a large one at that, how do you know what type of invasive breast cancer you have? Many of the low grade invasive tumors can be mixed with more aggressive components. Again, without a wide excision with clean margins you don’t know. Unlike you Maggis, some women do not want to walk around with an invasive tumor growing in their breast. That goes for you to kitty.
    I don’t know why you guys go on and on about this stuff. Just pay attention to your own breasts. It sounds like this case was a tweener. Let us say she had atypical ductal hyperplasia. ADH is still treated with excision. Pathology is an art as is all medicine. Why don’t you withhold judgement until you become a pathologist.

  6. Part of the problem here is that image analysis and use of MRI involves too much potential for misinterpretation of the results. The MRI machine does not detect tumors, it detects changes in material properties and shows them as an image. Rather than relying so heavily on the use of imaging for diagnosis, of breast cancer we need to promote research and funding for other means of detection that are more accurate and less prone to error. Reports in the last couple years have highlighted promising developments in urine tests that would provide a more accurate diagnosis as to the existence of a tumor. While not an immediately available solution to women, a greater focus on other detection sciences could make it available sooner. Right now, buying better and better imaging equipment is certainly not solving the problem.

  7. Doug–
    If you’re a doctor, I’m sure you know that DCIS is, ,by definition, Not invasive breast cancer.
    This is why the NIH panel is suggesting that we shouldn’t refer to DCIS as “carcinoma” (or cancer.)
    Precisely because pathology is an art, and because there is so much ambiguity in many cases, women should think twice, and get a second opoinion–and a third, if necessary– before letting someone cut into their breast.
    This is not a matter of blaming the pathologist (though this particular pathologoist was not certified, which is troubling).
    Some of these small lesions simply disappear, especially among younger women.
    This is why some would recommend “watchful waiting” and many would recommend that women under 50 shouldn’t go for mammograms.
    If they do there is a real risk that they will be diagnosed with DCIS and then face a dilemma: should I let them remove part or all of my breast?
    You describe that dilemma, somewhat casually, as a “tweener”–might be cancer, might not be. You add: “I don’t know why you guys go on and on about this stuff. Just pay attention to your own breasts.”
    I assume that by “you guys” you are referring to women.
    Doug, I truly hope you are a pathologist who works in a lab and never sees patients. I would hate to think that you are an oncologist who gives the patients the diagnosis: “Sorry to tell you, Lucy, but it’s a ‘tweener.’ Now don’t get all upset. I hate it when women go on and on about something. This is what we’re going to do. We’re going to make a wide excision . . .”
    If you were a woman you might have a clearer idea of how much anguish such an uncertain diagnosis would cause a woman.
    Finally, there are parallels to prostate cancer.
    Ten year ago, my husband was diagnosed with early-stage prostate cancer following a PSA test and a biopsy. The urologist recommended a radiation treatment that could lead to incontience or impotence. My husband decided not to do anything. He chose “watchful waiting,” and was perfectly comfortable with the decison.
    Today, his PSA levels have fallen. It seems clear, to his doctor, that he never had early-stage prostate cancer. Probably the biopsy was mistaken.

  8. Wellescent and Kitty, Kitty
    I agree completely.
    Equipment makers are making a fortune as they design equipment that shows more and more detail–at a higher and higher price.
    But this is only exacerbating the problem of false positive.
    There’s not as much money to be made on something like a urine test (though surely it would be profitable if it worked) but it would be a much better solution to the problem. Hope someone is working on it, in the public sector if not in the private sector . . .
    Kitty Kitty–
    Yes, exactly. That’s the problem.

  9. A bit off-topic but Maggie, your response to Doug was phenomenally witty and amusing. However, anyone who reads your blog knows that you take this issue VERY seriously (unlike Doug). But his casual devil-may- care, rather paternalistic attitude (Hey it’s an ART, not a SCIENCE, you dumb girl) seems to pervade the medical profession.
    I know that something that seems catastrophic to the patient – like hearing that they have cancer and having surgery, radiation and chemo – is not such a big deal to them because they see it every day. But to put a woman through this unnecessarily still seems to me like an absolute tragedy.
    Overdiagnosis and misdiagnosis ruin people’s lives. They also shorten people’s lives. Women endure mutilation and the absolute torments of hell when they are being treated for breast cancer. It just amazes me that so few of those who work in the medical profession even care that women are put through this unnecessarily.
    Your theme – Money Driven Medicine – is absolutely on the mark. It appears that the medical profession is quite corrupted. Everyone in medicine who has paid any attention to the issues of over-and-misdiagnosis knows that they occur (and too often…). But they don’t seem to care…
    For some lovely, literate commentary on the issue of overdiagnosis, please see Miriam Pryke’s posts at: http://www.bmj.com/cgi/eletters/340/mar23_1/c1241#234667 and http://www.bmj.com/cgi/eletters/340/jun24_1/c3106#238780

  10. Maggie occasionally I read lines that completely contradict everything that most readers assume this blog is about.
    You say the more a doctor does (surgery,reading tests) the better they become. It is well known you hate the fact that the more a doctor does the more he or she makes and wish to eliminate it. So on one hand you want someone that does alot of a service because they are “better”, but if they benefit personally from that skill and effort by doing more of that service, that’s wrong. Huh? Do more so you are better at a skill, but don’t do more because that harms patients. Which is it?

  11. Jenga said: ‘Do more so you are better at a skill, but don’t do more because that harms patients. Which is it?’
    This is nothing to do with Maggie’s post, which is about the evidence for carrying out procedures. If the evidence is that a procedure is necessary, it is now becoming mandatory in some countries that hospitals have a certain volume of complex operations to qualify to do them, and that individual surgeons also meet training and volume requirements. A good example is radical prostatectomy, which is not an operation I’d have someone who only does a few a year do on me outside of a major teaching hospital.

  12. KittyKitty7555 wrote:
    Everyone in medicine who has paid any attention to the issues of over-and-misdiagnosis knows that they occur (and too often…). But they don’t seem to care…
    ————-
    This is an astute comment. I believe it too be way too true, and the main reason behind why it works is that medicine has fostered a propaganda that they have almost magical insight and are not to be questioned. It eminates an almost religious aura that the average frightened person just cannot overcome.
    A great gig if you can get it, but someday, sometime (maybe now) average folks may start to ask for some real proof of benefits. Maybe??

  13. Ktty, Kitty, Jenga, Marc, NG
    Kitty, Kity– Thanks very much- and thank you for the links.
    Jenga- See Marc’s reply. Obviously, a doctor gets more practice by working at a large medical center that specializes in the procedure and attracts many patients that need the service–not by persuading patients in his private practice that they need the service when they don’t.
    Marc- Yes, I agree. It makes sense to require that a surgeon has done a certain number of a particular procedure before being the lead surgeon on a team (or being the only surgeon on the team.)
    Atul Gawande (who is a surgeon) explains in his book that becoming a surgeon is not so much about manual dexterity (though no doubt that helps), but having the patience to practice, practice, practice . .
    NG– It’s true people want to believe. They want to believe that “early detection” always works, that if they’re just tested for everything they’ll be safe.
    Of course this isn’t true–sometimes a disease is detected early, and the patient still dies. Sometimes a disease is detected early, and it’s “pseudo-disease” that never would have harmed the patient. Sometimes the disease is detected later–and the patient is treated and lives.
    So we do have to ask questions that begin “What are the odds . . ” “What are the risks . . ” acknowledging
    that in medicine, things are often ambiguous.
    On the whole, the U.S. Preventive Services Task Force recommendations on which preventive servicees are most helpful and when/ how often you should be tested seem to me a sensible guide for average-risk people. The Task Force has no axe to guide, and its recommendations are evidence-based.

  14. One more thing – the figure of 50,000 – 90,000 seems the same sort of fuzzy extrapolated statistical hyperbole used to estimate the number of deaths from medical errors in hospitals. Did anyone conduct an investigation of these 50 – 90K biopsies? Follow the clinical course of these women? Review all slides? I didn’t think so. This sort of hyperbole should be featured on the Cable Show “Scare Tactics.” That’s not to say there’s no element of truth here. However, this sort of hyperbole actually weakens the case.

  15. I think doctors should be more careful with diagnosis and stop playing gods. Most of the cancers are still found too late.

  16. Doc 99–
    Doctors who offer second opinions on mammographies (and often find that the diagnosis is wrong)have reached a consensus on how often women are mistakenly diagnosed with breast cancer.
    See NIH chart in story on the increase in diagnosis of DCIS– up 600%..
    Then, consider this: “The widespread and virtually unchallenged acceptance of screening has resulted in a dramatic increase in the diagnosis of ductal carcinoma-in-situ (DCIS), a pre-invasive cancer, with a current estimated incidence of about 40,000 annually. DCIS is usually recognized as micro-calcifications and generally treated by lumpectomy plus radiation or even mastectomy and chemotherapy . However, some 80 percent of all DCIS never become invasive even if left untreated (18). Furthermore, the breast cancer mortality from DCIS is the same— about 1 percent— both for women diagnosed and treated early and for those diagnosed later following the development of invasive cancer. That early detection of DCIS does not reduce mortality is further confirmed by the 13-year follow-up results of the Canadian National Breast Cancer Screening Study.” Excerpted from Excerpted from “Dangers and Unreliability of Mammography: Breast Examination is a Safe, Effective and Practical Alternative”, by Samuel S. Epstein, Rosalie Bertell, and Barbara Seaman, International Journal of Health Services, Volume 31, Number 3, 2001. (For sources of studies, see footnotes below the excerpt here http://www.preventcancer.com/patients/mammography/unreliability.htm
    The numbers the Times reports are intended to warn women that misdiagnosis is commonplace. This is why mammograms are no longer recommended for average risk women under 50.
    Finally, you are wrong about the most common cause of malpractice suits:
    From a 2010 article in the journal Applied Radiology
    http://www.appliedradiology.com/Issues/2010/01/Articles/Failure-of-radiologic-communication–An-increasing-cause-of-malpractice-litigation-and-harm-to-patients.aspx
    “The most common cause of medical malpractice litigation in the United States is “failure to diagnose,” but data from medical malpractice insurance companies show that the second most common cause is failure to communicate results of radiologic examinations.3 In fact, data disclose that communication problems are at least a causative factor in up to 80% of Medical Malpractice cases.
    Note, in this 80% of cases, teh doctor clearly made a mistake: he never gave the test results to the patient.
    The article continues ” This is not surprising, considering that a survey of family medicine physicians found that errors in communication accounted for 70% of all errors in that specialty, outpacing errors in diagnosis, which accounted for 47%.4 Another study found that physicians failed to acknowledge 36% of abnormal radiologic results; 4% of these, many of which made reference to a possible cancer, were lost to follow-up.”

  17. Clearly the answer lies in further development of pathological tests and procedures that eliminate doubt. In addition, reducing the medical paranoia perpetuated by the cancer industries would also help.