Mammograms, Breast Biopsies, and Uncertainty

Of the 1.6 million breast biopsies done last year, 20 percent or 320,000 are reported back as invasive carcinoma.  10 percent or 160,000 are reported as atypia and another 60,000 are reported ductal carcinoma in situ (DCIS).  DCIS is not invasive carcinoma, although apparently transformed cancer cells are found within the lactiferous (milk) ducts.  There is no general agreement on whether DCIS really represents cancer because such lesions may never become invasive.  If the apparently cancerous cells never break out of the duct and spread through the breast tissue, much less metastasize to other areas of the body, then the diagnosis of cancer doesn’t really apply due to the lack of actual consequences that we fear from cancer.

A new report in the Journal of the American Medical Association (JAMA) suggests how uncertain the diagnosis can be.  The authors gathered 115 pathologists and had them independently review individual slides from 240 breast biopsies.  Sets of three pathologists were also used to make a consensus diagnosis on each slide.

For invasive carcinomas, 96% agreed with the diagnosis.  For DCIS, there was 84% agreement (3% called it invasive, while 13% called it negative.)  For atypia, only 48% agreed, while 17% called it something worse and 35% called it negative.  For benign cases without atypia, 87% agreed and 13% called it at least atypia.

There was more disagreement for cases in which the women had been described as having “dense” breasts on a previous mammogram.  There was also more disagreement among pathologists who had lower case volumes, in smaller practices, and in nonacademic settings.

The level of disagreement is more significant because there are so many biopsies done each year. (Note, however, that disagreement was greater in pathologists with lower case volume, making for better results all around than these numbers would suggest.)  If only 4% of invasive carcinomas are missed and there are 320,000 cases altogether, that makes 12800 cases a year. If, on the other hand, 3% of DCIS cases are called invasive, out of 60,000 cases, 1800 will be overdiagnosed.  Atypia is the most contentious diagnosis, with only 48% agreement in the study.

This study demonstrates that there can be significant disagreement among pathologists as to the degree of atypia or even invasive carcinoma in breast biopsies.  It suggests that there could be over ten thousand cases of missed invasive carcinoma in breast biopsies every year in the US.  This could be a heavy burden when you consider that most of those invasive cases will turn out to be Stage II or above in a year or two, with probably obvious masses or even metastases.

One would hope that most of the missed cases will be borderline or very early, and thus, not so likely to result in the appearance of advanced cases later on.  One would hope.

The alternative case is those DCIS diagnoses.  Some experts think that many of these cases will never become invasive, and thus do not need to be treated aggressively by excision of the entire breast, or worse, radiation therapy.  These patients can be thought of as suffering an unnecessary burden of treatment, including its direct effects and its side effects.

The entire system of 1.6 million biopsies appears to be grounded in the numerous mammograms performed each year.  Certified facilities report their numbers every three years, at the time of recertification, and these numbers are aggregated into an approximate yearly total of 38.7 million mammograms in the US.  This suggests that about 4 per cent of mammograms result in a biopsy.

Since only 20 per cent of biopsies are positive for invasive carcinoma, this means that four out of five suspicious mammograms (those prompting  a biopsy) are “false positive.”  The remainder of the patients biopsied suffer the anxiety and side effects without result.

How many total cases of invasive breast cancer are reported each year in the US?  According to breastcancer.org, approximately 233,000.  That’s almost a hundred thousand less than the number of positive biopsies quoted above.  You could add the 62,000 cases of “non-invasive breast cancer” quoted as well as the 2,360 cases in men, but you’d still be short.

These discrepancies are related to the conclusion that three out of four suspicious mammograms are “false positives” and result in unnecessary biopsies.  However, as a result of general use of mammography, late diagnoses have decreased by about 37% and early diagnoses have increased by about 48%.  In addition, the total rate of breast cancer has gone down by about 9% “since mammography was introduced.” (per breastcancer.org)  (Some of this decrease may have been due to a decrease in the  use of postmenopausal hormones, although this is controversial.)

So, at a significant cost, we have relieved, to a considerable degree, the burden of breast cancer for American women.  The next step should be a better screening tool, applied to more of the women at risk.   There are (very) approximately 80 million women in the US over the age of 40, so less than half of women eligible by age are getting yearly mammograms.

For better results, we need to increase the number of women screened and reduce the percentage of biopsies.

MRI scans are used on occasion, although their expense prohibits screening use.  The American Cancer Society recommends MRI scans in addition to mammograms in women who are at high risk of breast cancer (for example, those who carry the BRCA-1 or -2 gene, which increases one’s lifetime risk to 40-60%, although these genes only cause perhaps 10% of breast cancers as well as ovarian cancers.)  Perhaps the cost of MRI’s can be reduced, as we know there is a problem with cost in medical advances that is not seen in general technological advances.  More on this problem later.

The abstract of the JAMA report can be found at: http://jama.jamanetwork.com/article.aspx?articleid=2203798