A Peek into the Future

Posted on: December 2nd, 2022 by admin 1 Comment
Print Page

Eleven years ago, we reviewed the Pulitzer Prize-winning book, The Emperor of All Maladies: A Biography of Cancer, by oncologist Dr Siddhartha Mukherjee (© Getty Images). As described on the dust cover, the author “examines cancer with a cellular biologist’s precision, a historian’s perspective, and a biographer’s passion. The result is an astonishingly lucid and eloquent chronicle of a disease humans have lived with – and perished from – for more than five thousand years.” Earlier this week, Mukherjee delivered a keynote lecture entitled “A Peek into the Future of Biomedical Transformation” at the annual meeting of the Radiological Society of North America (RSNA) in Chicago.

In his introduction, Mukherjee said, “To start with, I want to talk about deep learning which means deploying learning algorithms that imitate human learning.” He asked, “Can machines learn like us? Can machines learn medicine?” Mukherjee then referred to an article he’d written in 2017 entitled “AI versus MD” that was published in The New Yorker. The editors had provided the title which he felt should rather have been “AI with MD.” A pertinent question was whether AI could mimic how human intelligence works, and the answer will be important for the future of medicine.

Mukherjee next described a series of clinical examples that involved the detection of cancer using AI algorithms as well as leveraging AI to support cancer research. The areas that he believed held the greatest promise were detecting breast cancer using mammography, identifying pancreatic cancer in high-risk patients, and evaluating the likelihood of recurrence of prior cancer. He also believed that AI algorithms had a pivotal role to play in providing a second opinion.

“You could imagine yourself making a diagnosis while the AI algorithm makes a diagnosis, and then you compare notes,” he said and continued, “And so the algorithm could ask you whether you’re sure of your diagnosis.” Mukherjee stated that the most important question that needs to be asked is, “What is the endpoint?” He argued that the best endpoint was lives saved, although he acknowledged this was a measure that required decades of research to answer.

Mukherjee was optimistic about the future of AI in transforming the practice of medicine, particularly in the next few years. This year’s RSNA meeting has certainly provided numerous examples of the utility of AI algorithms, one of which came from Lunit, a Korean company, which showed that AI models using digital breast tomosynthesis performed better than AI models using digital mammography (seen above right, © Su-A Yang).

One Response

  1. Daniel Kopans says:

    I have not read his latest book. Having watched computer aided detection (CAD) using mammograms fail, I am optimistic, but skeptical, about AI. It seems to me that since computers “never forget a case”, and are not distracted, they should be better than radiologists and this is not yet the case. When I find something on a study that is of concern, I can explain why I am concerned. AI algorithms are truly “black boxes”. The computer cannot tell us why it is “concerned” about a portion of the breast which is a major problem.

    I raised concerns about the author’s previous book and wrote to him in 2017 as below:

    Dear Siddhartha:

    Thank you for asking me to comment on your article. You certainly don’t need me to tell you how eloquently you write. However, I suspect that you may regret asking for my input!

    Just out of residency and having secured a 2 year junior staff position (predating fellowships) at the MGH, I agreed to read the mammograms when everyone else in the Department of Radiology refused. I thought I would do that for a few years and then someone would figure out how to cure breast cancer and I would go off and practice general radiology. More than 40 years later, we are no closer to a cure and I (and I suspect you too) have learned that we are far more complex organisms than anyone thought. Two million (??) years of evolution have selected our cells to be very tough allowing humans to survive. Since our cancer cells are simply normal cells “gone bad”, they have all the tools (and more) to protect themselves which is why cancer is so hard to kill. It is pretty clear to me (and supported by your piece) that for many cancers, including breast cancer, “cure” requires treatment before “successful” metastatic spread has occurred. I suspect that systemic treatment may be able to destroy small metastases, but as you well know, if a woman has clinically evident metastatic breast cancer she will never be cured and will likely die from her metastatic disease. My more than 40 years of learning has taught me that the key to saving women from breast cancer is finding it before it has become “successfully” metastatic (explained below).

    Unfortunately, although beautifully written, I was disappointed in “Emperor” in that you provided inaccurate information about breast cancer screening. Your reference to Don Berry’s report on the CISNET models in which he asked what role screening played in the decline in breast cancer deaths in the U.S. provided a range of 28-65% did not go far enough. Clearly the range was large because each model’s results were determined by the assumptions programmed into their algorithm and they each used different assumptions. I’m not certain that averaging these estimates has any validity.

    The randomized, controlled trials proved that earlier detection reduces deaths. This has been confirmed by numerous observational studies where women, despite having access to modern therapies, have better survivals when they have been participating in screening [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17]. These observational studies are not proof that screening saves lives, but they all reinforce the “proof” provided by the RCT’s. With regard to what is responsible for the major decline in breast cancer deaths, that began in 1990. The observational studies, noted above, clearly support the fact that early detection (screening) is the main reason for the decline. Presumably all of the women in these analyses had access to modern therapy, yet deaths were reduced for women who participated in screening. The final corroboration (there are no other ways that I know of to assess the benefit of screening) was a major review that we undertook at the MGH and Brigham in which we found in a “failure analysis” that more than 70% of the women who died from breast cancer, despite having access to modern therapies, were among the 20% of women who were not participating in screening [18].

    As you know better than I, we don’t “cure” metastatic breast cancer. Cure is achieved when these cancers are treated before successful spread which is why screening reduces deaths. I would point out that all of the CISNET models agree that the most lives are saved by annual screening beginning at the age of 40 [19]. It is both astonishing and telling, that those who seek to reduce access to screening by emphasizing the “harms” of screening (basically recalls from screening since “overdiagnosis”, if it even exists, is not altered by delaying screening until age 50 or biennial participation) fail to inform women who are now in their thirties that, based on the CISNET models, the major “harm” is that if they wait until the age of 50 (there is no scientific support for using age 50 as a threshold for screening) and then are screened every two years, as many as 100,000 will die unnecessarily. These are lives that could be saved by annual screening starting at the age of 40 [20].

    This was just reinforced in an updated analysis using the CISNET models [21] confirming that tens of thousands of lives of women now age 40 will be lost if the USPSFT and even the ACS guidelines are followed instead of annual screening starting at the age of 40. There are actually no data (ZERO) that support the use of the age of 50 as a threshold for screening. The use of the age of 50 is simply due to inappropriately grouping data to make it appear as if numbers change suddenly at the age of 50 when, in fact, none of the parameters of screening change abruptly at that age or any other age [22].

    The only valid argument supporting a delay until the age of 50 is that it reduces the “false positive” rate. Unfortunately, this term “false positive” has been chosen to be purposely ambiguous and pejorative. In fact, this does not connote women who are told that they have breast cancer when they do not (left unclear by those trying to reduce access to screening), but it merely represents the number of women who are recalled for something seen on their screening study that is resolved by a few extra mammograms or an ultrasound. In fact, the vast majority are told that everything is fine. What is stunning is that the effort to reduce access to screening, that will result in tens of thousands of unnecessary deaths, is to “protect” women from the anxiety and inconvenience of being recalled for a few extra pictures or an ultrasound. This height of paternalism is the result of “panels” of inexperienced individuals (none of the members of the USPSTF or the ACS panels provided care for women with breast cancer) projecting their own concerns in an effort to deny women access to screening (the USPSTF guidelines set insurance coverage). They claim to have “weighed” the benefits and harms, but never explain their scale – how many fewer recalls for a few extra pictures or an ultrasound “outweigh” a life lost unnecessarily?

    A few more pre-comment comments.

    I think that in “Emperor” you missed the fact that HIP included a clinical breast examination in the screening arm. This was the main question that led to the other trials – how much of the mortality reduction in the HIP came from the CBE vs. the very poor quality industrial film mammograms of the 1960’s? The other trials (with the exception of Canada and Edinburgh) were undertaken to determine the value of screening using “modern” mammography alone and did not include a CBE. Edinburgh has been excluded by guidelines panels due to major socioeconomic imbalances in the allocation. The Canadian trials were designed and executed because the PI, Dr. Miller, was convinced that all that was needed for screening was a CBE. The nurses who performed the CBE in the CNBSS were highly trained whereas the technologists doing the mammograms and the radiologists who interpreted the mammograms had no specific training and this resulted in the poor quality of their mammograms.

    I think you did a nice job of capturing the allocation issue in Canada which, long ago, should have disqualified it from influencing screening guidelines [23]. In addition, having been invited by the Canadians to review their mammograms, I can personally attest to their poor quality (small cancers were missed) as has also been attested to by their own reference physicist [24] as well as a review that they organized [25, 26]. Not a single radiologist who participated in the CNBSS has defended the quality of their images. You were correct to have ignored the Canadian results, but the other RCT’s clearly proved that earlier detection reduces deaths.

    I don’t recall whether or not you explained to the reader that the RCT’s (with the exception of Canada) were by ”invitation to be screened”. Many women who were invited did not participate (non-compliance), yet if they died from breast cancer they were still counted as having been screened. Conversely, if a woman was assigned to the unscreened control arm and had a mammogram on her own outside the trial (contamination) that saved her life, she was still counted as an unscreened control. In many of the trials as many as 20% or more of women refused to be screened while a similar number had mammograms outside the trials so that the RCT’s underestimate the benefit from screening.

    The BCDDP was undertaken because those who wanted to prevent screening in the U.S. (an effort dating back to the 1960’s) claimed that it would not be possible to screen large numbers of women. This was why the BCDDP was not an RCT. It did what it was designed to do and showed women could be screened efficiently and effectively.

    I would think that as an oncologist you would be very aware of the fact that therapy for breast cancer only saves lives when breast cancers are treated earlier. As you imply in the material that you sent to me (and in “Emperor”) therapy can delay death, but women with metastatic breast cancer will die from their cancers. You can delay many deaths, but the only way to “cure” breast cancer is to find it before it is successfully metastatic.

    This brings me to your paper. I completely agree that we do not understand why breast cancers tend to spread to certain organs and not to others, but I do not think that we yet have a clue as to why this happens. You describe the shower of cancer cells in the rat model as suggesting that cancers cells are floating in the vascular system very early in the tumor’s growth. I don’t think human data support this for breast cancer. Not only were these implanted tumors (an artificial situation), but, as you know, many things that happen in rodents do not happen in humans. The data on breast cancer (including the fact that early detection can lead to cure) suggest that metastatic spread actually occurs late in the development of these malignancies. A recent review of the genetics of metastatic breast cancers [27, 28] showed that metastatic lesions contained the same mutations as the primary tumors. Since mutations continue as the primary cancer grows, the fact that the genetics are similar suggests that dissemination occurs later, reinforcing the importance of early detection.

    Our own modeling showed that, with exceptions, successful metastatic spread is directly linked to tumor volume (the number of cancer cells) [29]. With obvious exceptions (there are some women who present with metastatic disease with very small, or even undetectable cancers in their breast), the data suggest that metastatic spread is stochastic. The larger the number of cancer cells in the primary tumor the greater the chance that a cell or cells will break away and spread to other organs. I use the term “successfully metastatic” because (and this integrates with your thesis) even when cancer cells can reach the vascular system, they are not necessarily able to establish themselves in another organ. I have no idea why and when this occurs. However, the fact that most invasive breast cancers are cured when they are detected and treated before growing larger than 1 cm [30] explains that they are rarely successfully metastatic until larger than 1 cm, explaining why screening saves lives. Although I have no doubt it is a bell shaped curve, for most breast cancers there needs to be a large enough number of cells and perhaps sufficient neovascularity to permit cells to begin to shed into the vascular and lymphatic system. It may also mean that the distant tissues have to be “prepared” to allow the cells to take hold and grow (as you are describing). Metastatic spread is “successful” when the circulating cells have developed the capability to live in tissues other than breast, and presumably when the “soil” in other organs has been altered to foster this growth. Perhaps there are differences in the capillary beds in which the cancer cells lodge that determines “successful” spread? I am not aware of anyone who really has a handle on how all this works and I have learned to be very skeptical. I remember when Judah Folkman seemed to have found the key to stopping cancer by blocking angiogenesis. Would it could be so simple, but we are clearly not simple organisms.

    With regard to Dr. Welch:

    Unfortunately, I could write a book (happy to work on it with you??!!) [31] detailing all of the misinformation that has been promulgated about breast cancer screening, much of it written and promoted by Dr. Welch.

    When I was in junior high school learning first year algebra, our teacher assigned us to write a formula using only x’s and y’s; a’s, and b’s. I came up with a formula that proved that 1=2. It took my teacher a half hour to figure out where I had divided by zero. I have learned to read papers in the medical literature very carefully to see if the authors “divided by zero” (obviously not literally).

    It is astonishing and of great concern that, at least with regard to breast cancer screening, there has been a great effort to establish “alternative facts”. These have been published and the “big lie” (now called “post truth”) has been fostered by some at even the “best” medical journals [32, 33, 34]. The concepts of “overdiagnosis” and “overtreatment” have been adapted to reduce access to screening. Although, through a major stretch, screening does lead to some of both, stopping screening to stop these “overs..” is like removing the engine from our cars to prevent automobile accidents. We all know that there are cancers that do not kill women, and we all know (as you write) that many/most breast cancers are overtreated [35]. Long term follow-up of women diagnosed with breast cancer in the 1940’s, long before there was any screening, in which all were treated by mastectomy alone (there were no systemic treatments) showed that 30% were alive 30 years later [36]. These were not small cancers and today all would have been treated systemically and, clearly, unnecessarily.

    Unfortunately, no one knows how to identify these cancers. But also, unfortunately, an entire mythology has been developed around them. In an effort to reduce access to screening it has been suggested and promoted by Dr. Welch, that mammography finds breast cancers that will disappear if left alone. Authors have pointed to the handful of reported cases that have been “documented” in which breast cancers “disappeared”. Ironically (!!??) these reports describe breast cancers that were palpable. Several “disappeared” in the breast, yet no one is saying we should stop treating palpable cancers! Instead these extremely rare “miracles” (actually several women still died from metastatic disease) have been used to reduce access to screening and to deprive women of the opportunity to be cured. Welch and others support reducing access to screening as a way to prevent “overtreatment”. I believe your approach, which makes far more sense, is to understand who will actually benefit from treatment. Screening doesn’t make the diagnosis and screening doesn’t determine therapy.

    NO ONE, including Dr. Welch, who has claimed tens of thousands of mammographically detected cancers that are found each year would disappear if left alone [37], has actually ever seen a mammographically detected breast cancer regress or disappear on its own. A recent survey of Breast Imaging practices identified almost 500 breast cancers that were not treated. NONE of these cancers regressed or disappeared on their own [38].

    One of Welch’s main tenets, that your writing seems to support, is that breast cancer, thyroid cancer and prostate cancers are all similar. He has taken various pieces of information to make a specious case. I think it is important to evaluate the various lesions that have been called cancer in the breast, but simply because there appear to be many indolent thyroid cancers does not mean that there are similarly numerous indolent breast cancers. Welch has argued against prostate cancer screening yet he ignored the facts. The death rate from prostate cancer had been increasing rapidly in the 1980’s. PSA was introduced in the mid 1980’s and soon after the death rate suddenly began to fall. There was no new treatment for prostate cancer and the main explanation for the decline in deaths is clearly screening. I think you know that the RCT for prostate cancer screening was a shambles. As I understand as many (? more) men in the control arm had PSA testing! There are clearly important issues surrounding prostate cancer screening. Therapy can have major functional consequences, but it is clear that earlier detection reduces deaths.

    It is reasonable to ask similar questions about these cancers, but dangerous to assume that these cancers are very similar. They all have differing “attack” rates and I am certain differing biologies. An example of variability among cancers is evident in the fact that the “logical” progression from a normal epithelium to cancer first described by Volgelstein in the development of colon cancer, has not been found in breast cancer. DCIS and their associated invasive breast cancers are genetically the same. As you know, everyone is trying to understand why DCIS becomes invasive with no good explanation at this time.

    Breast cancer is a problem for women at a younger age than prostate cancer is for men. The work that you cited previously, concerning autopsy material, has been disputed by pathologists who claim that many of the lesions were inappropriately called cancers. Furthermore, they, for the most part, included older women. I have little doubt that older women have undiagnosed breast cancers that might never affect them before they are killed by their heart disease, but there are no identifiably “fake” invasive cancers in younger women.

    The best autopsy study was from the four corners region of the U.S. that looked at the breasts of 500 women who had died accidentally [39]. This was essentially a random cross section of the population. It described 11 cancers in 512 women which is 22/1000. Superficially this seems like most had to be “fake” cancers since the annual incidence of breast cancer in the U.S. is approximately 1/1000 at age 40 up to 5/1000 by age 80. What most do not understand and Welch ignores is the fact that, unless breast cancers grow from a single cell to a clinically evident cancer in a single year, there have to be many cancers in the population at various sizes growing gradually to “reach the surface” of clinical detection. In fact, if the average doubling time for a breast cancer is 120 days, it can take 20 years to go from a single cell to a 2 cm cancer. There have to be many more cancers growing in the population (all “real”) at any time to sustain the annual incidence reaching delectability that we see in the U.S. explaining the “four corners” autopsy findings. Years ago I published a very simple model of breast cancer growth that explains this [40].

    Given the depth and thoughtfulness of your analyses, I am surprised that you have included Dr. Welch’s theories in your work. His arguments are the modern equivalent of copper bracelets. He “divides by zero” all the time. Clearly for all medical problems, some people get better on their own, some are harmed by the treatment for the disease, some are cured, and some die. This is true for every medical intervention.

    With regard to Dr. Welch one has to wonder who actually funds the “Dartmouth Institute for Health Policy and Clinical Practice? Why is Dr. Welch concerned about increasing the profits of insurance companies?:

    “The time has come for it [participation in mammography screening] to stop being used as an indicator of the quality of our health care system” [41].

    You should be aware that he does not provide care for women with breast cancers. He makes uninformed statements such as he made in our debate on the Diane Rehm show on NPR: “Debate Over The Benefits Of Routine Mammograms”

    WELCH: “Well, let me first start about what we’re not talking about. We’re not talking about diagnostic mammography, and that’s the test we order when women become aware of a new breast lump. All doctors agree that’s a good test to figure out whether the lump is something to worry about.”
    FACTS: This is simply false. “Diagnostic mammography” rarely can determine whether or not the “lump” is “something to worry about”. Its primary value is in screening the rest of the breast and contralateral breast for unsuspected occult cancer given that most “lumps” are benign.
    Dr. Welch presents himself as an expert, yet he is either a revisionist or simply uniformed. He made completely false statements commenting on the CNBSS on CNN [42]:
    WELCH: “that randomization did exactly what it is supposed to do: It created two identical groups of women. The rate of death in the two groups was exactly the same, every year, for 25 years” [43].

    FACTS: He was either being purposely misleading or, as a supposed expert, was unaware of the fundamental allocation issues in the CNBSS [44]. There had been great controversy in the 1990’s when the Canadians reported that there were more deaths among the screened women in their forties in CNBSS1 [45]. They claimed that mammography was leading to earlier deaths among screened women in their forties [46]. The rate of death was not “exactly the same, every year, for 25 years.” By combining the results from their two separate trials (CNBSS1 and CNBSS2), the Canadians were covering up the problems in their trial, and Welch was either buying the cover-up, or was simply uniformed of the fundamental issues.

    In fact, Welch frequently “divides by zero”. Every one of his papers on breast cancer and “overdiagnosis” [47] has relied on his “guess” at what the underlying incidence of breast cancer would have been over the years had there not been any screening beginning in the mid 1980’s. In his 2012 paper with Dr. Bleyer in the NEJM [48], he claimed that SEER data showed an Annual Percent Change (APC) of 0.25-0.5% per year. This led them to claim that mammography screening found 70,000 fake breast cancers in 2008 alone since the actual rate in 2008 was considerably higher than his extrapolation! Not only did they admit that 0.25 was their best “guess”, but one has to ask how analysts who had no data on which women had breast cancer and which cancers were found by mammography could conclude that mammography was leading to massive overdiagnosis! It is impressive to blame an intervention with no data on the intervention! Subsequently, three separate analyses have now shown that the paper was scientifically unsupported [49, 50, 51], but the NEJM would not withdraw it and would not even publish a letter to the editor signed by more than 40 experts in breast cancer care (not just radiologists) calling for the paper to be withdrawn.

    Welch wrote a second paper in the NEJM using the same SEER data [52], but this time claimed that the incidence would have not increased at all (APC = 0.0%) had there not been any screening. That peer reviewers allowed him to use the same data to claim different APC’s is a bit of a scandal. Regardless, both guesses are incorrect. The APC had been changing steadily by 1-1.3% per year for decades before screening began. Welch used the period after the wives of the President and Vice President of the U.S. had been diagnosed with breast cancer (1974) which prompted a major flurry of ad hoc screening that ended soon after, to determine what he “guessed” would have been the expected annual baseline increase in breast cancers. He was clearly unaware, or does not understand that when a population is screened there is a jump in “incidence” (actually the annual detection rate called a “prevalence” hump/bump) of cancers that had been building up in the population as well as the cancers detected earlier. When screening stops (as happened soon after the flurry), the bump was followed by a compensatory drop in incidence since screening had already removed some of the incident cancers one or more years in advance. This means that the early years of the SEER database began, the years used by Welch to establish his baseline, are the most unreliable in the database for determining the baseline, prescreening, incidence. Every analysis of breast cancer incidence that I can find prior to the start of SEER used the Connecticut Tumor registry data that go back to 1940. They all (including Walter Willett) acknowledge the fact that, contrary to Welch, the incidence of invasive breast cancer had been increasing steadily at the rate of 1-1.3% per year [53, 54, 55, 56] before there was any screening, yet Welch has ignored the data. Had he used the correct extrapolation he would have found that there is no “overdiagnosis” of invasive cancers, and that, contrary to his assertion, the rate of advanced cancers has decreased by almost 40% since the start of screening [57]. Peer review has failed, and journal biases have allowed scientific nonsense to be published [58].

    In one of his Youtube talks Welch “borrowed” an image from Tabar and Dean’s atlas of mammography without permission or any attribution (see attached). On top of that, in an effort to make his point that mammography was so good that it was finding “fake” cancers, he falsely claimed that the “borrowed” mammogram, showing a very large breast cancer, was “poor quality” “from the 1960’s” to make the point that real cancers are large and indisputable even on the old bad mammograms. He then showed 3 tiny calcifications on another mammogram claiming (I suspect falsely) that these indicated a tiny breast cancer, as a way to show that technology has permitted us to find tiny unimportant cancers. I suspect the image with the calcifications was falsely labeled since I am unaware of anyone who would biopsy a group of 3 unimpressive calcifications. More egregious was the fact that the Tabar and Dean image was not poor quality from the 1960’s, but was rather a high quality mammogram from the 1980’s. Not only was the image used without permission, but it was misattributed and mislabeled. Dartmouth only required Dr. Welch to remove the image from his talk!

    In summary I would suggest:

    1. As an oncologist I am certain that you have rarely (never) seen a woman with metastatic breast cancer “cured”. She may die from some other cause, but, apropos of your chapter, if she has metastatic disease, she will die from her breast cancer. Treating breast cancers early, before “successful” metastatic spread, is the key to cure.
    2. Most breast cancers are not likely to be successfully metastatic for many years after they begin to grow. It is not until they become larger than 1 cm that mets begin to become more common. This is the reason that screening saves lives. It finds many (not all) breast cancers at a time when there has either not been successful metastatic spread, or the tumor burden is so small that systemic therapy actually eliminates it.
    3. All the major panels, including the USPSTF and the ACS admit that annual screening starting at the age of 40 saves the most lives.
    4. There is no scientific support for using the age of 50 as a threshold for screening.
    5. Delaying screening until the age of 50 will have no effect on overdiagnosis since, if fake cancers exist (and they do not), unless they disappear (and no one has ever seen this happen), they will still be there at age 50. The only “harm” that delaying screening will reduce by waiting until the age of 50 and then screening every two years, is recalls from screening. Panels should not be deciding for women whether or not they can tolerate the inconvenience and anxiety of being recalled. This should be each individual woman’s choice regardless of age.

    Dan Kopans

    [1] Tabar L, Vitak B, Tony HH, Yen MF, Duffy SW, Smith RA. Beyond randomized controlled trials: organized mammographic screening substantially reduces breast carcinoma mortality. Cancer 2001;91:1724-31
    [2] Kopans DB. Beyond Randomized, Controlled Trials: Organized Mammographic Screening Substantially Reduces Breast Cancer Mortality. Cancer 2002;94: 580-581
    [3] Duffy SW, Tabar L, Chen H, Holmqvist M, Yen M, Abdsalah S, Epstein B, Frodis Ewa, Ljungberg E, Hedborg-Melander C, Sundbom A, Tholin M, Wiege M, Akerlund A, Wu H, Tung T, Chiu Y, Chiu Chen, Huang C, Smith RA, Rosen M, Stenbeck M, Holmberg L. The Impact of Organized Mammography Service Screening on Breast Carcinoma Mortality in Seven Swedish Counties. Cancer 2002;95:458-469.
    [4] Otto SJ , Fracheboud J, Looman CWN, Broeders MJM, Boer R, Hendriks JNHCL, Verbeek ALM, de Koning HJ, and the National Evaluation Team for Breast Cancer Screening* Initiation of population-based mammography screening in Dutch municipalities and effect on breast-cancer mortality: a systematic review Lancet 2003;361:411-417.
    [5] Swedish Organised Service Screening Evaluation Group. Reduction in breast cancer mortality from organized service screening with mammography: 1. Further confirmation with extended data. Cancer Epidemiol Biomarkers Prev. 2006;15:45-51
    [6] Coldman A, Phillips N, Warren L, Kan L. Breast cancer mortality after screening mammography in British Columbia women. Int J Cancer. 2007 Mar 1;120(5):1076-80.
    [7] Jonsson H, Bordás P, Wallin H, Nyström L, Lenner P. Service screening with mammography in Northern Sweden: effects on breast cancer mortality – an update. J
    Med Screen. 2007;14(2):87-93.
    [8] Paap E, Holland R, den Heeten GJ, et al. A remarkable reduction of breast cancer deaths in screened versus unscreened women: a case-referent study. Cancer Causes Control 2010; 21: 1569-1573.
    [9] Otto SJ, Fracheboud J, Verbeek ALM, Boer R, Reijerink-Verheij JCIY, Otten JDM,. Broeders MJM, de Koning HJ, and for the National Evaluation Team for Breast Cancer Screening. Mammography Screening and Risk of Breast Cancer Death: A Population-Based Case–Control Study. Cancer Epidemiol Biomarkers Prev. Published OnlineFirst December 6, 2011; doi: 10.1158/1055-9965.EPI-11-0476
    [10] van Schoor G, Moss SM, Otten JD, Donders R, Paap E, den Heeten GJ, Holland R, Broeders MJ, Verbeek AL. Increasingly strong reduction in breast cancer mortality due to screening. Br J Cancer. 2011 Feb 22. [Epub ahead of print]
    [11] Mandelblatt JS, Cronin KA, Bailey S, et.al. Effects of mammography screening under different screening schedules: model estimates of potential benefits and harms. Annals of Internal Medicine, 2009; 151: 738-747; see also http://cisnet.cancer.gov, last accessed 16 April 2011.
    [12] Hellquist BN, Duffy SW, Abdsaleh S, Björneld L, Bordás P, Tabár L, Viták B, Zackrisson S, Nyström L, Jonsson H. Effectiveness of population-based service
    screening with mammography for women ages 40 to 49 years: evaluation of the Swedish Mammography Screening in Young Women (SCRY) cohort. Cancer. 2011 Feb
    [13] Broeders M, Moss S, Nyström L, Njor S, Jonsson H, Paap E, Massat N, Duffy S, Lynge E, Paci E; EUROSCREEN Working Group. The impact of mammographic screening
    on breast cancer mortality in Europe: a review of observational studies. J Med Screen. 2012;19 Suppl 1:14-25. Review
    [14] Hofvind S, Ursin G, Tretli S, Sebuødegård S, Møller B. Breast cancer mortality
    in participants of the Norwegian Breast Cancer Screening Program. Cancer. 2013 Sep 1;119(17):3106-12
    [15] Sigurdsson K, Olafsdóttir EJ. Population-based service mammography screening: the Icelandic experience. Breast Cancer (Dove Med Press). 2013 May 9;5:17-25
    [16] Coldman A, Phillips N, Wilson C, Decker K, Chiarelli AM, Brisson J, Zhang B, Payne J, Doyle G, Ahmad R. Pan-canadian study of mammography screening and
    mortality from breast cancer. J Natl Cancer Inst. 2014 Oct 1;106(11).
    [17] Puliti D, Bucchi L, Mancini S, Paci E, Baracco S, Campari C, Canuti D, Cirilli C, Collina N, Conti GM, Di Felice E, Falcini F, Michiara M, Negri R, Ravaioli A, Sassoli De’ Bianchi P, Serafini M, Zorzi M, Caldarella A, Cataliotti L, Zappa M; IMPACT COHORT Working Group.. Advanced breast cancer rates in the epoch of service screening: The 400,000 women cohort study from Italy. Eur J Cancer. 2017 Feb 18;75:109-116.
    [18] Webb ML, Cady B, Michaelson JS, Bush DM, Calvillo KZ, Kopans DB, Smith BL. A failure analysis of invasive breast cancer: most deaths from disease occur in women not regularly screened. Cancer. 2014 Sep 15;120(18):2839-46.
    [19] Mandelblatt JS, Cronin KA, Bailey S, Berry DA, de Koning HJ, Draisma G, Huang H, Lee SJ, Munsell M, Plevritis SK, Ravdin P, Schechter CB, Sigal B, Stoto MA, Stout NK, van Ravesteyn NT, Venier J, Zelen M, Feuer EJ; Breast Cancer Working Group of the Cancer Intervention and Surveillance Modeling Network. Effects of mammography screening under different screening schedules: model estimates of potential benefits and harms. Ann Intern Med. 2009 Nov 17;151(10):738-47.
    [20] Hendrick RE, Helvie MA. USPSTF Guidelines on Screening Mammography Recommendations: Science Ignored. Am. J. Roentgenology 2011; 196: W112 – W116.
    [21] Arleo EK, Hendrick RE, Helvie MA, Sickles EA. Comparison of recommendations
    for screening mammography using CISNET models. Cancer. 2017 Aug 21. doi: 10.1002/cncr.30842. [Epub ahead of print] PubMed PMID: 28832983.
    [22] Kopans DB, Moore RH, McCarthy KA, Hall DA, Hulka C, Whitman GJ, Slanetz PJ, Halpern EF. Biasing the Interpretation of Mammography Screening Data By Age Grouping: Nothing Changes Abruptly at Age 50. The Breast Journal 1998;4:139-145.
    [23] Kopans DB. The Canadian National Breast Screening Studies are compromised and their results are unreliable. They should not factor into decisions about breast cancer screening. Breast Cancer Res Treat. 2017 Aug;165(1):9-15.
    [24] Yaffe MJ. Correction: Canada Study. Letter to the Editor JNCI 1993;85:94
    [25] Baines CJ, Miller AB, Kopans DB, Moskowitz M, Sanders DE, Sickles EA, To T,Wall C. Canadian National Breast Screening Study: assessment of technical quality by external review. AJR Am J Roentgenol. 1990 Oct;155(4):743-7.
    [26] Kopans DB. The Canadian Screening Program: A Different Perspective. AJR 1990 155:748-749
    [27] Yates LR, Knappskog S, Wedge D, Farmery JHR, Gonzalez S, Martincorena I, Alexandrov LB, Van Loo P, Haugland HK, Lilleng PK, Gundem G, Gerstung M, Pappaemmanuil E, Gazinska P, Bhosle SG, Jones D, Raine K, Mudie L, Latimer C, Sawyer E, Desmedt C, Sotiriou C, Stratton MR, Sieuwerts AM, Lynch AG, Martens JW,
    Richardson AL, Tutt A, Lønning PE, Campbell PJ. Genomic Evolution of Breast Cancer Metastasis and Relapse. Cancer Cell. 2017 Aug 14;32(2):169-184
    [28] Zhao X, Powers S. New Views into the Genetic Landscape of Metastatic Breast Cancer. Cancer Cell. 2017 Aug 14;32(2):131-133.
    [29] Michaelson JS, Silverstein M, Wyatt J, Weber G, Moore R, Halpern E, Kopans DB, Hughes K. Predicting the survival of patients with breast carcinoma using tumor size. Cancer. 2002 Aug 15;95(4):713-23
    [30] Tabar L, Fagerberg G, Day N, Duffy SW, Kitchin RM. Breast Cancer Treatment and Natural History: New Insights from Results of Screening. Lancet 1992;339:412-414.
    [31] The Breast Cancer Screening Arcade and the “Whack-a-Mole” Efforts to Reduce Access to Screening. Seminars in Ultrasound, CT, and MRI in Press
    [32] Kopans DB. Bias in the Medical Journals: A Commentary. Am. J. Roentgenol 2005; 185: 176 – 182.
    [33] Kopans DB. Informed decision making: age of 50 is arbitrary and has no demonstrated influence on breast cancer screening in women. Am J Roentgenology 2005;185:177-82
    [34] Kopans DB. More misinformation on breast cancer screening. Gland Surg. 2017 Feb;6(1):125-129.
    [35] Early Breast Cancer Trialists’ Collaborative Group (EBCTCG). Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet. 2005 May 14-20;365(9472):1687-717
    [36] Adair F, Berg J, Joubert L, Robbins GF.Long-term followup of breast cancer patients: the 30-year report. Cancer. 1974;33:1145-50
    [37] Bleyer A, Welch HG. Effect of three decades of screening mammography on breast-cancer incidence. N Engl J Med. 2012 Nov 22;367(21):1998-2005
    [38] Arleo EK, Monticciolo DL, Monsees B, McGinty G, Sickles EA. Persistent untreated screening-detected breast cancer: an argument against delaying screening or increasing the interval between screenings. J Am Coll Radiol 2017; 14:863-867.
    [39] Pollei SR, Mettler FA Jr, Bartow SA, Moradian G, Moskowitz M. Occult breast cancer: prevalence and radiographic detectability. Radiology. 1987 May;163(2):459-62.
    [40] Kopans DB, Rafferty E, Georgian-Smith D, Yeh E, D’Alessandro, H Moore R, Hughes K, Halpern E. A Simple Model Of Breast Cancer Growth May Provide Explanations For Observations Of Apparently Complex Phenomena. Cancer 2003;97:2951-2959.
    [41] Welch HG. Screening mammography–a long run for a short slide? N Engl J Med. 2010 Sep 23;363(13):1276-8
    [42] http://www.cnn.com/2014/02/19/opinion/welch-mammograms-canada/index.html?iref=allsearch accessed 3 3 2014
    [43] http://www.cnn.com/2014/02/19/opinion/welch-mammograms-canada/index.html?iref=allsearch accessed 3 3 2014
    [44] Kopans DB. The Canadian National Breast Screening Studies are compromised and their results are unreliable. They should not factor into decisions about breast cancer screening. Breast Cancer Res Treat. 2017 Aug;165(1):9-15.
    [45] Author of Canadian Breast Cancer Study Retracts Warnings. J Natl Cancer Inst Vol. 84, No. 11, June 3, 1992 pp. 832-834.
    [46] Miller AB, Baines CJ, To T, Wall C. Canadian National Breast Screening Study: 1. Breast cancer detection and death rates among women aged 40 to 49 years. CMAJ. 1992 Nov 15;147(10):1459-76.
    [47] Kopans DB. More misinformation on breast cancer screening. Gland Surg. 2017 Feb;6(1):125-129.
    [48] Bleyer A, Welch HG. Effect of three decades of screening mammography on breast-cancer incidence. N Engl J Med. 2012 Nov 22;367(21):1998-2005
    [49] Kopans DB. Arguments Against Mammography Screening Continue to be Based on Faulty Science. The Oncologist 2014;19:107–112
    [50] Helvie MA, Chang JT, Hendrick RE, Banerjee M. Reduction in late-stage breast cancer incidence in the mammography era: Implications for overdiagnosis of invasive cancer. Cancer. 2014 Sep 1;120(17):2649-56
    [51] Etzioni R, Xia J, Hubbard R, Weiss NS, Gulati R. A reality check for overdiagnosis estimates associated with breast cancer screening. J Natl Cancer Inst. 2014 Oct 31;106(12).
    [52] Welch HG, Prorok PC, O’Malley AJ, Kramer BS. Original Breast-Cancer Tumor Size, Overdiagnosis, and Mammography Screening Effectiveness. N Engl J Med 375;15 nejm.org October 13, 20161438
    [53] Shulman LN, Willett W, Sievers A, Knaul FM. Breast cancer in developing countries: opportunities for improved survival. J Oncol. 2010;2010
    [54] Kessler LG, Feuer EJ, Brown ML. Projections of the breast cancer burden to U.S. women: 1990-2000. Prev Med. 1991 Jan;20(1):170-82
    [55] Miller BA, Feuer EJ, Hankey BF. Recent incidence trends for breast cancer in women and the relevance of early detection: an update. CA Cancer J Clin. 1993 Jan-Feb;43(1):27-41
    [56] Garfinkel L, Boring CC, Heath CW Jr. Changing trends. An overview of breast cancer incidence and mortality. Cancer. 1994 Jul 1;74(1Suppl):222-7
    [57] Helvie MA, Chang JT, Hendrick RE, Banerjee M. Reduction in late-stage breast cancer incidence in the mammography era: Implications for overdiagnosis of invasive cancer. Cancer. 2014 Sep 1;120(17):2649-56.
    [58] Kopans DB. More misinformation on breast cancer screening. Gland Surg. 2017 Feb;6(1):125-129.