CREATINGTHE MYTH OF "OVERDIAGNOSIS" (1)
One of the major concerns that has been raised about breast cancer screening is the belief that it might lead to“Overdiagnosis”. “Overdiagnosed” breast cancers are defined as cancers that look the same to the pathologists but will never become clinically evident and will never harm a woman over the course of her life. Clearly if an “overdiagnosed”breast cancer is treated, then this constitutes “overtreatment” since it would have never harmed the woman. It is claimed that “overdiagnosed” cancers will not cause any problems if they areleft undetected and untreated. If these cancers actually exist (and it is likely that they do not) since they are,histologically, indistinguishable from potentially lethal cancers, they can be termed “fake” cancers.
It is indisputable that there are a number of women who are diagnosed with breast cancer and treated each year who die from some other cause before their breast cancer becomes fatal. For example, a woman who is killed in an automobile accident driving home from her last treatment for breast cancer could be said to have been “overtreated”. It is a fundamental fact that the vast majority of women who are treated with systemic therapy for breast cancer do not benefit from the treatment (1).However, since benefit is far from perfectly predictable, it does not mean thata woman was “overdiagnosed” in the sense that the cancer would have never ledto death had she not died from some other cause.
There are, indeed, breast cancers that are indolent. For example, in over 40 years I have heard of only one case of a “tubular” breast cancer – a well differentiated invasive breast cancer - that was fatal. However, if a woman lives long enough and her breast cancer is not treated, she will likely die from her cancer.
Although the following discussion is primarily about invasive cancers, even some low-grade Ductal Carcinoma In Situ(DCIS), if given enough time, will become invasive and lethal (2).
THE FALSE CLAIM THAT BREAST CANCERS WILLDISAPPEAR ON THEIR OWN
The basic claims of “overdiagnosis” have originated from the false claims that there are thousands of breast cancers that would disappear on their own without therapy. This concept is based on the incredibly rare cases of women diagnosed with breast cancer whose lesions have, reportedly,disappeared without any treatment (3,4). These are so rare that they can be classifiedas true “miracles” although at least one woman, whose breast cancer disappeared from her breast, still died from metastatic disease. Nevertheless, these case reports have resulted in false claims of massive “overdiagnosis” (tens of thousands each year in the U.S. alone).
Not only are “disappearing” cancer sincredibly rare, but they are being used to claim that mammography screening isthe main problem. This ignores the fact,all of the examples of disappearing cancers, that have been reported, have been clinically evident. It is ironic that the arguments of disappearing cancers have been used to limit mammography when no one has ever seen a mammographically detected breast cancer disappear on itsown. There has never been a report of a mammographically detected invasive breast cancer disappearing on its own. I have personally polled thousands of radiologists at meetings, and no one has ever seen a mammographically detected breast cancer disappear on its own. In a study of almost 500 women whose breast cancers were not treated, none disappeared,and none even regressed on follow-up studies (5). The very rare “disappearing” cancers have all been clinically evident, yet no one is arguing that clinically evident cancers are “overdiagnosed” and should be left untreated. Fundamentally, pathologists make the diagnosis and on cologists decide on treatment. The effort to limit mammography screening to reduce “overdiagnosis” is like arguing that we can reduce car accidents by removing the engines from our cars.
THE FALSE CLAIM THAT DELAYING SCREENINGUNTIL THE AGE OF 50 WILL REDUCE “OVERDIAGNOSIS”
Because the arguments to limit screening to reduce “overdiagnosis” are so convoluted, and not supported scientifically, a fundamental fact has been over looked. For example: it has been argued that we should not screen women in their forties and wait until the age of 50 to reduce “overdiagnosis”. What has been unstated but required to justify delaying screening, is that the “overdiagnosed” breast cancers that would have been found by screening women in their forties have to disappear by age 50. Otherwise, they will still be there when women begin screening at the age of 50. Groups like the US Preventive Services Task Force (USPSTF) have made the argument that they want to reduce “overdiagnosis” and advise delaying screening until the age of 50 without explaining that this will have no effect on “overdiagnosis” unless the “fake” cancers disappear. If the “fake” cancers detected by mammography do not disappear, then dealying screening will not reduce “overdiagnosis”. Those who advocate delaying screening fail to explain this incongruity but acknowledge that women will die unnecessarily, by delaying screening.
FALSE CLAIMS
Using specious arguments, it was claimed that a very large number of mammographically detected breast cancers, if left undetected, would disappear on their own (6). As the authors stated “we believe that the most tenable explanation of our findings is that some screen-detected breast cancers spontaneously regress[disappear].” This was simply misinterpretation of data. I have, informally, surveyed more than 4000 radiologists at various meetings and asked for a show of hands from anyone who has seen amammographically detected cancer disappear on its own. Not a single hand has ever been raised. In a study from multiple centers including almost 500 breast cancers detected by mammography, that had not been treated,none regressed or disappeared over time (7). Yet one often cited paper claimed that in 2008 alone there were more than 70,000 “fake” (“overdiagnosed”) cancers (8). It is hard to imagine that if there were 70,000 “fake” cancers detected in one year alone, no one has ever seen one disappear.
THERE ARE MANY MORE CANCERS IN THEPOPULATION THAN ARE DETECTED AND DIAGNOSED EACH YEAR
One of the arguments in support of“overdiagnosis” is based on the fact that, as we get better at early detection,we find more cancers in the population. It is fairly certain that breast cancers take many years to develop before they become large enough to be clinically evident (9). In fact, some may take decades to progress from a single cell to become a 2 cm, clinically evident mass. We should not be surprised to find many more cancers in the population than are clinically detected each year. This does not mean that these are “fake”cancers. Remember, a clinically evident cancer is simply one that has grown sufficiently large that the mass that has formed can be palpated. The “deeper” weare able to “dive” into the population with screening, the more cancers we can expect to find as the grow progressively larger. These are “real” cancers. This is why mammography detects more cancers than clinical breast examination and MRI detects more cancers than mammography. These are not “fake”cancers, but rather cancers that are slowly growing in the population at a rate that supports the annual detection rate of clinically evident cancers. A crude estimate is that, in order for there to be 3 cancers per 1000 women diagnosed each year in a population, and assuming it takes from 10-30 years for a cancer to grow to 2 cm., means that,for every three breast tumors that reach 2 cm, there are likely another 57 tumors (39 smaller invasive carcinomas and 18 DCIS) in various stages of development that would be undetected “below the surface” in the population. These are cancers that have not yet been detected, but given sufficient time will increase in size and, ultimately,become clinically evident lumps. The purpose of screening is to find these cancers before they reach clinical detectability, but more importantly, in an effort to treat them before they become successfully metastatic
EVIDENCE FOR A MUCH HIGHER RATE OF CANCERIN THE POPULATION THAN EXPLAINED BY THE ANNUAL DETECTION RATES
As noted in the model referenced above, breast cancers don’t simply grow from a single cell in one year to a 2 cm., clinically evident cancer. It likely takes years andeven decades for this to happen. Infact, if breast cancers do take years to grow there have to be many more growing in the population than become evident each year. This means that there are many more cancers“beneath the surface” in the population than the annual incidence would suggest. A major study was published years ago in which the breasts of almost 500 women who died, prematurely, from accidents or other traumatic causes were autopsied (10)(breast tissues are not normally evaluated in an autopsy). There were 9 unsuspected cancers that werefound (and additional microscopic cancers might have been overlooked). Given the very young ages of many of the women, this was, essentially, a random sample of the rate of breast cancers growing in the population. Based on this small sample, it would appear that there are approximately 18 cancers per 1000 women in the population of women of all ages at any given time. Since the clinical detection of breast cancers increases from 1/1000 at the age of 40 to 2/1000 by age 50 and to 3/1000 at the age of 60, this study would suggest that there are as many as 6 times as many cancers (and likely more) “below the surface” in the population,as are diagnosed each year. If clinically evident cancers are “real” then these are all “real” cancers that have simply not yet reached a clinically evident size. We find many of them by screening before they become clinically evident. They are“real” cancers. The fact that there are so many occult cancers in the population that are needed to produce the cancers that become clinically evident each year is the reason that MRI screening detects so many more cancers than mammography. These are almost certainly “real” cancers and it is likely that MR screening would further reduce deaths from breast cancer each year by detecting cancers earlier than mammography.
CANCERS THAT ARE NOT CLINICALLY RELEVANT
There are cancers that are not clinically relevant. These occur among older women who die from some other cause and cancer is found in their breasts that did not cause them any harm during their lives. There are limited data from autopsies. In one study of 83 elderly women who died from some unrelated cause, at autopsy, some were found to have breast cancers that had not affected them during their lives (11). Among 77 women who had no known previous cancer, there was one case of invasive breast cancer and 14 cases of what was then called “in situ” cancer (these have been, subsequently, disputed). This, however, is a better measure of overdiagnosis since, had these women been diagnosed with breast cancer and treated while alive they would, indeed, have been “overtreated”.
The disconnect lies in the fact that the effort to reduce access to screening to prevent “overdiagnosis” has been confined to screening younger women when, in fact, the only real evidence of possible “overdiagnosis” would be among the very old women who are under thereal threat of “competing causes of death”.
LIMITING SCREENING TO LIMIT OVERDIAGNOSIS REQUIRES DISAPPEARING CANCERS
As noted earlier, the main problem is that it has been argued that “overdiagnosis” can be reduced by delaying the start of screening from age 40 (the RCT proven threshold) until the age of 50 (an arbitrary threshold) and then screening every two years instead of annually. The only way that this will reduce“overdiagnosis” is if the “fake” cancers that would have been diagnosed among women in their forties, disappear by age 50. Unfortunately, those who claim to want to limit access to screening to reduce “overdiagnosis” have failed to explain to women and their physicians the fact that delaying screening to reduce “overdiagnosis” will only work if the fake cancers disappear. Otherwise, they will still be there at the age of 50. Ignoring the fact that no one has ever seen this happen has resulted in a whole mythology built around the exaggerated concept of “overdiagnosis”. There is no reason to expect that dealyingscreening will have any effect on “overdiagnosis” (if it even exists), but wedo know that limiting screening by dealying the start until the age of 50 andthen every two years will mean that thousands will die whose lives could besaved by annual screening starting at the age of 40 (12).
REFERENCES
1 Early Breast Cancer Trialists'Collaborative Group (EBCTCG). Effects of chemotherapy and hormonal therapy forearly breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet. 2005 May 14-20;365(9472):1687-717
2 Sanders ME,Schuyler PA, Simpson JF, Page DL, Dupont WD. Continued observation of the natural history of low-grade ductal carcinoma in situ reaffirms proclivity forlocal recurrence even after more than 30 years of follow-up. Mod Pathol. 2015May;28(5):662-9.
3 Dussan C, Zubor P, Fernandez M, YabarA, Szunyogh N, Visnovsky J. Spontaneous regression of a breast carcinoma: acase report. Gynecol Obstet Invest. 2008;65(3):206-211.
4 Krutchik AN, Buzdar AU, Blumenschein GR, Lukeman JM. Spontaneous regression of breast carcinoma. Arch Intern Med.1978;138(11):1734-1735.
5 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.
6 Zahl PH, Mæhlen J, Welch HG. The Natural History of Invasive Breast Cancers Detected by Screening Mammography. ArchInternMed. 2008;168(21):2311-2316
7 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.
8 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
9 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.
10 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.
11 Nielsen M, Jensen J, Andersen J. Precancerous and cancerous breast lesions during lifetime and at autopsy. A study of 83 women Cancer 1984;54:612-5
12 Arleo EK, Hendrick RE, Helvie MA, Sickles EA. Comparison of recommendations for screening mammography using CISNET models. Cancer. 2017 Oct 1;123(19):3673-3680.
