mammograms, again

the boston globe ran a story today pointing out the marginal benefit of mammography from a recent analysis --see below, but they had a thoughtful summary (see http://www.bostonglobe.com/lifestyle/health-wellness/2014/04/02/doctors-may-oversell-mammograms-benefits-mammograms-may-have-been-oversold-study-suggests/cA0Djh3Xj3uqGMBKxESYQO/story.html )

the article just came out in JAMA (see mammog screening review jama 2014 in dropbox, or doi:10.1001/jama.2014.1398) and reinforces the not-so-great efficacy of mammog screening, confirming what i sent out in several recent blogs. they did literature search finding 8 large RCTs (all done between the 1960s and 1990s, which may be an issue: see below) finding a 15-20% decreased breast cancer mortality. another meta-anal from canadian task force found 19% decrease after 11.4 yrs of followup. the new JAMA data (see tables below) includes the relative risk (RR), absolute risk reduction (ARR), overdiagnosis (detected tumors on screening that would never become clinically evident. mostly DCIS, but some suggestion that may also be some invasive cancer diagnoses). this article is available for free, so i am reproducing their results below.

table 1: pooled results from RCTs on mortality reductions with mammog screening by age group

age        total events in group/total number        RR with mammog    ARR with mammog        number                                                                                                                                             needed to screen

              mammog grp    control grp

39-49    448/152300        625/195919                0.85 (0.75-0.96)    0.0005                            1904

50-59    361/78465        410/69849                    0.86 (0.75-0.99)    0.0007                            1339

60-69    110/19093        155/18377                    0.68 (0.54-0.87)    0.0027                            377

70-74    42/5073            36/4859                       1.12 (0.73-1.72      NA                                NA

table 2: estimated benefits and harms of mammog screening in 10,000 women with annual mammog over 10 year period

age        # diagnosed with invasive        # breast ca deaths        # deaths averted with          overdiagnosis        # with >=1 false pos        # with >=1 unnecessary

            breast ca or DCIS over 10yrs        in next 15 yrs               screen over next 15 yrs      during 10 yrs            during 10 yrs                 biopsy during 10 yrs

40            190                                        27-32                            1-16                                ?-104                    6130                                700

50            302                                        56-64                            3-32                                30-137                  6130                                940

60            438                                        87-97                            5-49                                64-194                  4970                                980

(note: the "# of breast ca deaths in next 15 yrs" column is the number who would die even if they were screened)

they point out that these studies were of women of average risk and the benefit would likely be greater if women at higher risk. for example 4 microsimulation (ie mathematical) models found that women 40-49 with Gail model score twice average and given biennial screen would have same ratio of benefit/harm as woman>50 with average risk.

so, this is a very complex issue. this article confirms the pretty paltry benefits of mammography screening (although average lifetime risk of breast cancer is 12.3%, aggressive mammog screening will avoid breast cancer mortality in only 3-32 of 10,000 women screened in their 50's, with 6130 having false positives, 940 getting biopsies, 30-137 with overdiagnosed breast cancers) and does not include the fact that treatment is much better now than when the studies were done, and does not even mention that all these mammograms are likely (by mathematical modeling) to create some cancers. also important to note that the differences in breast cancer mortality do not translate into differences in all-cause mortality (prob in part because the numbers saved by mammog is really so small). attempts at provider/patient shared-decision-making so far have probably not been adequate (eg, no change in mammog ordering after USPSTF came out with guidelines to decrease mammogs to every other year and only in women 50-74.  also, boston globe comments on 2010 study of 460 women, where >96% reported that MD discussed benefits of screening but <20% discussed risks).  be aware that the suggestion of doing individualized risks, esp those 40-50 yo, (Gail model etc) makes sense mathematically, but there are no real-world data. also that the array of imputed breast cancer risk factors (inc BMI, dec bone density, smoking, alcohol, estrogen/androgen exposure, late age of first pregnancy...) are absent in 60% of women with breast cancer.  which also (again) raises the elephant-in-the-room: the huge numbers of industrial toxins in our environment, food chain, water... which may well be a really major cause of breast cancer (the vast vast majority never adequately tested, but we do know of many so far which have hormonal stimulatory  effects on breast tissue). but in the context of mammog, i think it makes sense to discourage women under 50 from getting mammog, and doing mammog every 2 year in women aged 50-74 (there are pretty good data that every other year testing is just about as sensitive and fewer false positives... and also less radiation)