USPSTF breast cancer screening recommendations

the US Preventive Services Task Task Force just published a draft recommendation for breast cancer screening in women after age 50 (see http://www.uspreventiveservicestaskforce.org/Page/Document/RecommendationStatementDraft/breast-cancer-screening1 ), as follows.

for women 50-74 years old, screening mammography every 2 years, grade B recommendation (moderate certainty that there is net benefit):

    --meta-analysis suggests that screening 10,000 women age 50-59 over 10 years will result in 8 fewer breast cancer deaths; screening 10,000 women age 60-69 over 10 years will result in 21 fewer breast cancer deaths. these data are from really old studies. it is likely that current screening may detect more cancers but also that current treatment will decrease the deaths

    ​--harms of screening: most important is overdiagnosis and overtreatment. hard to know for sure what the % is, depends on modeling methods used, but it is likely that the increased sensitivity of mammography screening leads to more overdiagnosis. estimates range from 0% to 54%, but the accepted number is that about 20% (1 in 5 women) are treated for a cancer that would never have been discovered without the mammography and would not have led to health problems.  the other screening harm is false-positives leading to more imaging/biopsy. The data from the Breast Cancer Surveillance Consortium (BCSC, collaborative network of 5 mammography registries plus 2 others with linkage to tumor registries): per 10,000 women screened once: 

            --age 50-59 -- 932 false positives; 60 biopsies performed for each case of invasive cancer; 11 false-negative mammograms (missed cancers). 

            --age 60-69 -- 808 false positives; 30 biopsies for each case of invasive cancer; 12 false-negative mammograms. 

            --age 70-74 -- 696 false positives; 30 biopsies for each case of invasive cancer; 13 false-negative mammograms. 

    --frequency of screening: no direct data from clinical trials, but looking at the current trials with screening intervals o f 12 to 33 months, there was no clear trend to benefit in more frequent screening for different starting ages. observational evidence that there was no difference in breast cancer deaths in women >50 screened biennially vs annually

    ​--when to consider stopping screening: this data is based on modeling, since there are inconclusive data on 70-74 yo women. USPSTF does not recommend screening 70-74 yo women with moderate to severe comorbid conditions (moderate=cardiovascular disease, paralysis, diabetes; severe=AIDS, COPD, liver disease, renal failure, dementia, CHF, MI, ulcer, rheumatologic disease and combo of moderated conditions).

for women 40-49 years old, grade C recommendation (selectively offer screening, based on professional judgment and patient preferences. moderate certainty that net benefit is small vs more common harms). still suggesting biennial screening.

    --breast cancer deaths avoided by repeated screening of 10,000 women over 10 years = 4.

    --harms of mammography for 10,000 women screened once: for age 40-49: 1,212 false positives; 100 biopsies done to find 1 case of invasive cancer; 10 false-negative mammograms. 

    --frequency of screening: as with 50-74 year olds (above)

    ​--consider starting at age 40 especially in women with first-degree relative (parent, child, sibling) with breast cancer (increases risk 2-fold)

comparing starting at age 40 vs age 50:

--life-time benefits for biennial screening mammograms per 1,000 women (note: this is per 1,000 women vs the 10,000 in other data above)

    --age 40-74: reduced breast cancer deaths 8; life-years gained: 152.  harms: false positive test 1,529; unnecessary biopsies 204; overdiagnosed breast cancers 20.

    ​--age 50-74: reduced breast cancer deaths 7; life-years gained: 122.  harms: false positive test 953; unnecessary biopsies 146; overdiagnosed breast cancers 18.

--10-year cumulative probability of false positive​ mammogram or biopsy:

    --begin at age 40: annual mammograms with 61.3% having false positive, 7.0% with false-positive biopsy recommendation

                                    biennial mammogram with 41.6% having false positive, 4.8% with false-positive biopsy recommendation

    --begin at age 50: annual mammograms with 61.3% having false positive, 9.4% with false-positive biopsy recommendation

                                    ​biennial mammogram with 42.0% having false positive, 6.4% with false-positive biopsy recommendation

for women age >75: insufficient evidence, though some models do suggest continued benefit after age 74 (these are all mathematical models, no real clinical data)​

tomosynthesis (3-D digital mammography): insufficient data for use as primary breast cancer screening strategy, though there may be reduced recall rates for false-positives, and does expose women to twice the radiation

breast density: BCSC data suggest that 25 million women (43%) have heterogeneously or extremely dense breasts, and % is highest in those age 40-49. increased breast density is associated with higher risk of breast cancer (though not increased risk of dying from breast cancer), and is associated with lower sensitivity (from 87% to 63%) and specificity (from 96% to 90%) of mammography. so, women with increased breast density are at increased risk of false-positive test, unnecessary biopsy, and false-negative test. for women aged 40-49 (but not other groups), there are data suggesting that those with extremely dense breasts have more benefit from annual vs biennial screening.

other modes of screening, esp in women with dense breasts (eg, breast ultrasound, MRI) -- insufficient evidence to recommend.

One of the big unknowns with breast cancer screening is around overdiagnosis and specifically with DCIS (ductal carcinoma in-situ). The incidence of DCIS increased dramatically now that mammography is routine, from 6 to 37 cases/100K women/yr. DCIS is not necessarily a cancer (there is a movement underfoot to reclassify it as not a cancer)​ in that it is in most cases localized/confined to the mammary ductal-lobular system and does not metastasize, there is no good way to differentiate the majority of regular DCIS from the unusual ones that do metastasize, and the current treatment of DCIS is pretty aggressive (lumpectomy/mastectomy, then maybe tamoxifen) -- ie, this is a big hole in our knowledge base and needs more research. Another big issue is the ethnic/racial disparity: African-American women have a slightly lower incidence of breast cancer (127 cases/100K, vs 133 cases/100K in white women) yet a significantly higher mortality. ??role of socio-economic status/access to care (which is undoubtedly part of the issue in many areas of the country) vs biology -- are there underlying biological differences leading to the finding thatAf-Am women tend to have more aggressive and treatment-resistant tumors (egwith triple-negative phenotypes, and more dysplastic tumors), which, by the way, tend to be less amenable to screening?

So, I think these recommendations overall are pretty appropriate. The efficacy of screening is higher in the older age group (median age of cancer diagnosis is 61, and the above numbers confirm increased utility in the 60-69 yo age group). Based on modeling data a few years ago, I have been advocating biennial exams in order to decrease the risk of radiation exposure, and most of my patients are very open to that. I do offer screening at age 40, though I try to make it clear that this is an individual choice, that though there may be some benefit, there are also clear risks of overdiagnosis/increased likelihood of needing more xrays(additional radiation) and biopsies.

For past blogs, see

http://gmodestmedblogs.blogspot.com/2015/03/discordance-in-interpreting-breast.html  which reviews inconsistencies in breast biopsy interpretation

http://gmodestmedblogs.blogspot.com/2014/01/mammography-concerns-update.html  goes through more about the risks of screening, utility of using meds in high risk women as cancer prevention, and the need to look into environmental toxins to prevent breast cancer from happening in the first place

http://gmodestmedblogs.blogspot.com/2014/04/mammograms-again.html also critiquing the low efficacy of screening (it should be done, but overall it is not having a huge impact on breast cancer survival)