Can We Trust Breast Biopsy Results?
Each year approximately 1.6 million women in the United States have a breast biopsy. Although there are well-established diagnostic criteria to classify biopsy tissue, pathologists are misinterpreting biopsy tissue at a significantly high rate according to a study published in the March 17, 2015 edition of The Journal of the American Medical Association (JAMA). As a result, many women are either not receiving the treatment they need or are receiving treatment they do not need.
In the study, "Diagnostic Concordance Among Pathologists Interpreting Breast Biopsy Specimens," researchers assembled a panel of experts interpret 240 cases and reach a consensus diagnosis in each case. The researchers then sent the biopsy slides to over 600 pathologists and asked them to interpret the slides without knowing the diagnoses of the consensus panel or of the other pathologists participating in the study.
Invasive Breast Cancer: The consensus panel found 23 cases of invasive breast cancer, and 96% of the blind reviewers reached the same diagnosis but 4% missed it (i.e. "underinterpreted" the case).
Ductal Carcinoma in Situ (DCIS): The consensus panel found 73 cases of DCIS, and 84% of the blind reviewers reached the same diagnosis, but 3% overinterpreted the case while 13% underinterpreted it.
Atypical Hyperplasia (Atypia): The consensus panel found 72 cases of atypia, and 48% of the blind reviewers reached the same diagnosis, but 17% overinterpreted the case and 35% underinterpreted it.
Benign: The consensus panel found 72 cases of benign tissue, and 87% of the blind reviewers reached the same diagnosis but 13% of the reviewers overinterpreted the case.
WHY ERRORS OCCURRED
Discrepancies in diagnoses were attributed to three primary factors:
1. Pathologists in smaller, non-academic practices with a lower volume of breast biopsy cases were more likely to misdiagnose the case.
2. Misdiagnosis was more common in women with denser breast tissue.
3. The study design permitted the pathologist to review only one slide whereas pathologists in everyday practice review several slides and order additional tests as needed to confirm a diagnosis. Researchers felt that this design flaw was necessary in order to recruit pathologists to participate as review of only one slide would minimize the time needed to participate.
Most concerning to the researchers was the finding that only 48% of the blind reviewers agreed with the consensus panel's diagnosis of atypia, with 17% of the blind reviewers overinterpreting the cases and 35% underinterpreting them. When atypia is overinterpreted as DCIS, a woman may undergo unnecessary surgery, radiation, or hormonal therapy. When atypia is underinterpreted as benign, women may not be advised of the benefits of increased surveillance, MRI exams, and chemoprevention.
Although the researchers advise that women receiving a diagnosis of atypia get a second opinion, that same advice would seem prudent for women receiving a diagnosis of DCIS or benign tissue since the researchers found that 52% of the women with atypia received the wrong diagnosis (17% of women with atypia received a diagnosis of DCIS and 35% with atypia received a diagnosis of benign tissue).
The study ultimately leaves us with more questions than answers. Can we trust breast biopsy results? Is the level of misdiagnosis as high as suggested by the study or are the results skewed due to design flaw (i.e. having the blind reviewers look at only one slide)? If there are well-established diagnostic criteria for classifying breast biopsy tissue, why is there such a high level of difference of opinion on what the tissue reveals? Further study is needed to answer these questions to identify the root causes for the presumed high level of misdiagnosis of breast biopsy tissue so that women can receive the correct diagnosis and the correct treatment recommendations.