Substantial Pathologist Diagnostic Disagreement for Breast Biopsy Specimens
In a study reported in JAMA, Elmore et al found a substantial rate of disagreement between pathologist diagnosis based on single breast biopsy slides and the consensus-derived diagnosis of a panel of pathologists. Disagreement rates were highest for ductal carcinoma in situ (DCIS) and atypical hyperplasia.
Study Details
The study involved data from 165 pathologists who independently interpreted breast biopsy slides between November 2011 and May 2014 from four test sets of 60 breast biopsies (240 total cases, 1 slide per case). The pathologists represented 91% of those responding to an invitation to participate in the study; the response rate was 65%. The cases consisted of 23 cases of invasive breast cancer, 73 cases of DCIS, 72 cases of atypia, and 72 benign cases without atypia. Cases with atypia and DCIS, women aged 40 to 49 years, and women with mammographically dense breast tissue were oversampled to provide increased statistical precision of estimates of agreement/disagreement.
Participant pathologist diagnoses were compared with diagnoses of a consensus panel of three pathologists, who reviewed all 240 cases. Reference diagnosis was arrived at by discussion among panel members in cases in which they did not have unanimous independent agreement. The three reference pathologists agreed unanimously on diagnosis for 75% (180 of 240) of the cases on initial independent evaluation. Compared with the final consensus-derived reference diagnoses, the concordance of the initial independent diagnoses of the panel members was 90.3% (650 of 720 interpretations), including concordance rates of 97% for invasive carcinoma, 97% for DCIS, 80% for atypia, and 91% for benign cases without atypia.
Agreement With Consensus-Derived Diagnosis
Compared with the consensus-derived reference diagnosis, the overall concordance rate of diagnoses of pathologists was 75.3% (5,194 of 6,900 interpretations), including rates of 96% for invasive carcinoma, 84% for DCIS, 48% for atypia, and 87% for benign cases without atypia (P < .001 for trend). Overinterpretation rates were 3% for DCIS, 17% for atypia, and 13% for benign cases without atypia (P < .001 for trend), and underinterpretation rates were 4% for invasive carcinoma, 13% for DCIS, and 35% for atypia (P < .001 for trend).
Disagreement with the consensus-derived reference diagnosis was significantly greater for biopsies from women with higher (n = 122) vs lower (n = 118) mammographic breast density (concordance rate = 73% vs 77%, P < .001). With regard to pathologist characteristics, the agreement rate was significantly lower for those who interpreted lower weekly case volumes (eg, 72% for < 5 and 75% for 5–9 cases/week, P < .001 for trend), worked in smaller practices (74% for 1–9 pathologists vs 78% for ≥ 10, P = .034), or worked in nonacademic settings (eg, 74% for no academic affiliation vs 81% for primary affiliation, P = .007).
The investigators concluded: “In this study of pathologists, in which diagnostic interpretation was based on a single breast biopsy slide, overall agreement between the individual pathologists’ interpretations and the expert consensus-derived reference diagnoses was 75.3%, with the highest level of concordance for invasive carcinoma and lower levels of concordance for DCIS and atypia. Further research is needed to understand the relationship of these findings with patient management.”
Joann G. Elmore, MD, MPH, of the University of Washington, is the corresponding author of the JAMA article.
The study was supported by the National Cancer Institute and by the National Cancer Institute–funded Breast Cancer Surveillance Consortium.
Dr. Elmore reported serving as a medical editor for the nonprofit Informed Medical Decisions Foundation. Kimberly H. Allison, MD, reported personal fees from Genentech.
The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.
