Pathologists disagree about 8% of the time when interpreting a single breast biopsy slide, with more overinterpretation than underinterpretation in discordant cases, according to an analysis combining results from the B-Path (Breast Pathology) study with data on the prevalence of breast cancer pathology diagnoses in women aged 50 to 59 years from the Breast Cancer Surveillance Consortium. Disagreement was more likely in cases of ductal carcinoma in situ and atypia, and agreement was more likely in cases of invasive breast cancer, Joann G. Elmore, MD, MPH, of the University of Washington, Seattle, and colleagues reported in the Annals of Internal Medicine. For most women who have a biopsy in the United States, the diagnosis is benign without atypia, the investigators noted, and for these women, diagnostic agreement was 97.1%.
The B-Path study compared pathologists’ interpretations of a single biopsy slide and a reference consensus interpretation from three experts, “experienced breast pathologists who are internationally recognized for research and continuing medical education on diagnostic pathology,” the authors noted. A total of 115 practicing pathologists participated, resulting in 6,900 interpretations from 240 distinct cases.
The researchers estimated the probability that a pathologist’s interpretation of a single breast biopsy slide would be confirmed by the reference consensus interpretation. For example, if a single slide from a woman’s biopsy was interpreted as ductal carcinoma in situ, how likely is it that her slide would get the same diagnosis from a panel of three expert pathologists? In that case, the researchers found that one in five women with an initial diagnosis of ductal carcinoma in situ would have her biopsy specimen interpreted as atypia or benign without atypia by the reference consensus panel. They also found that more than half of the women with atypia would have their diagnosis downgraded to benign without atypia.
“These noninvasive but potentially high-risk breast lesions represent a grey area with subjective boundaries imposed on a biological continuum; there is not always a ‘right’ or ‘wrong’ diagnosis and, as in many areas of medicine, professional opinions may differ,” the authors wrote. “Our results,” they added, “suggest that overinterpretation of the pathologic findings may contribute to overdiagnosis and overtreatment of ductal carcinoma in situ.”
“Overall, if 1 representative slide were used per case, 92.3% (95% CI, 91.4%–93.1%) of breast biopsy diagnoses would be verified by reference consensus diagnoses, with 4.6% (CI, 3.9%–5.3%) overinterpreted and 3.2% (CI, 2.7%–3.6%) underinterpreted,” the researchers reported. “Verification of invasive breast cancer and benign without atypia diagnoses is highly probable; estimated predictive values were 97.7% (CI, 96.5%–98.7%) and 97.1% (CI, 96.7%–97.4%), respectively. Verification is less probable for atypia (53.6% overinterpreted and 8.6% underinterpreted) and ductal carcinoma in situ (18.5% overinterpreted and 11.8% underinterpreted).”
The researchers pointed out that actual accuracy of interpretations might be higher, because in practice, pathologists often obtain second opinions, evaluate more than one slide, or use special stains. “We should also note that we used diagnostic prevalence rates based on a population of women in their 50s who were screened using film mammography, and results may differ with newer technologies, such as digital mammography, magnetic resonance imaging, and tomosynthesis,” the authors wrote. “In addition, our estimates may not reflect outcomes for women in other age groups, unscreened women, or women from different countries.”
Efforts to reduce diagnostic variability “might include educational programs, improved diagnostic techniques, or second-review policies,” the authors wrote. “Alternatively, women with borderline breast lesions that are difficult to categorize, such as atypical ductal hyperplasia and low-grade ductal carcinoma in situ, may benefit from revised guidelines for clinical treatment and management given the degree of diagnostic variability and biological overlap between these diagnostic categories.”
Accompanying Editorial
“Diagnostic uncertainty implies a need to revise our classification of proliferative lesions to minimize confusion, more appropriately reflect risk, communicate uncertainty, and minimize unnecessary treatment,” Alexander Borowsky, MD, University of California Davis, and Laura Esserman, MD, MBA, UCSF Medical Center, San Francisco, wrote in an accompanying editorial. “Unfortunately, our terminology contributes to inaccuracies by drawing sharp clinical distinctions between lesions that are imprecisely classified in practice and may not reflect an accurate description of risk,” the editorialists added.
For example, “the word ‘atypia’ remains a trigger for excision in almost all clinical settings. Similarly, when patients with low-grade ductal carcinoma in situ hear that they have ‘cancer’ (carcinoma), both they and their physicians feel compelled to choose excision plus radiation or mastectomy,” the editorialists explained. “The B-Path study further emphasizes the need to redefine ‘cancer’ and avoid using the term for lesions that are not destined to kill the patient.”■
Elmore JG, et al: Ann Intern Med. March 22, 2016 (early release online).
Borowsky A, Esserman L: Ann Intern Med. March 22, 2016 (early release online).
