This article is more than 5 years old. Information may no longer be current.
Pathologists concordance rates vary for DCIS, atypia diagnoses
Click Here to Manage Email Alerts
Pathologists’ concordance rates for ductal carcinoma in situ and atypical hyperplasia diagnoses were generally lower than their concordance rates for invasive carcinoma, according to study results.
Disagreement occurred more frequently when interpreting results from women with dense breasts and among pathologists who had lower case volumes and who worked in nonacademic settings, results also showed.
“About 1.6 million breast biopsies are done every year in the U.S., yet in nearly half of the cases labeled atypia, our study indicates that there simply isn’t ‘concordance,’ meaning there is not agreement with a standard reference diagnosis,” study researcher Anna N. A. Tosteson, ScD, of the Dartmouth-Hitchcock Norris Cotton Cancer Center, said in a press release.
Tosteson and colleagues evaluated data from 60 breast biopsy test sets for 240 total cases between November 2011 and May 2014. Forty-nine percent of the women were aged 40 to 49 years. The sample included cases of invasive carcinoma (n = 23), ductal carcinoma in situ (DCIS; n = 73), atypical hyperplasia (atypia; n = 72) and benign cases without atypia (n = 72).
A panel of three independent reference pathologists reviewed all cases and recorded their ratings of case difficulty and diagnoses using a Breast Pathology Assessment Tool and Hierarchy for Diagnosis form. These panelists had unanimous agreement of their independent diagnoses in 75% of the cases. Overall concordance of the initial independent diagnoses compared with the final consensus-derived reference diagnoses was 90.3%.
The analysis included 115 participant pathologists who each interpreted 60 cases. Participating pathologists were not aware of the interpretations of other study participants or consensus panel members.
The researchers observed that participants agreed with the consensus-derived reference diagnosis for 75.3% (95% CI, 73.4-77) of the interpretations.
The overall concordance rate for invasive carcinoma was 96% (95% CI, 94-97). One invasive test case contained primarily DCIS cases with a focus on microinvasion, which was missed by two reference panelists.
The concordance rates between participant pathologists and the consensus-derived reference diagnosis were considerably lower in cases of atypia (48%; 95% CI, 44-52).
Overinterpretation as an invasive carcinoma occurred in more cases of atypia (17%; 95% CI, 15-21) and benign cases without atypia (13%; 95% CI, 11-15) than in DCIS (3%; 95% CI, 2-4).
Rates for underinterpretation of invasive carcinoma were higher among DCIS (13%; 95% CI, 12-15) and atypia (35%; 95% CI, 31-39) than invasive carcinoma (4%; 95% CI, 3-6).
Disagreement with the reference diagnosis occurred significantly more frequently when interpreting results from women with higher vs. lower breast density on prior mammograms (overall concordance rate, 73% vs. 77%; P ˂ .001). Disagreement also was associated with pathologist characteristics, including lower weekly case volumes (P ˂ .001), working in smaller practices (P = .034) and working in nonacademic settings (P = .007).
The researchers acknowledged several limitations to their study such as the uncertainty of how the use of test sets may have influenced interpretive performance and that only a single slide was used to represent each case.
“Our findings show that, particularly for women with dense breasts, the lack of consensus in diagnoses of atypia or DCIS is considerable,” study researcher Tracy Onega, PhD, of the Dartmouth-Hitchcock’s Norris Cotton Cancer Center, said in the release. “The take-home message for women is that it’s not overly cautious to seek a second opinion with one of these diagnoses. In fact, it’s likely to be a good idea.” – by Cameron Kelsall
Disclosure: Tosteson and Onega report no relevant financial disclosures. One researcher reports personal fees from Genentech.
Perspective
Back to Top
Fang Fan, MD, PhD
The criteria for the pathological diagnosis of atypical ductal hyperplasia (ADH) and low-grade ductal carcinoma in situ (DCIS) of the breast have been clearly defined in the leading textbooks of breast pathology. In the daily pathology practice, nevertheless, there are quite a few cases showing overlapping features of both entities. Such "grey zone lesions" or "borderline cases" are a common cause of discordance among general surgical pathologists and experts in the field, and even among specialized breast pathologists. This lack of concordance and the existence of bona fide controversial borderline cases are well known to all pathologists, oncologists and surgeons specializing in breast diseases. However, similar diagnostic disagreements exist in other areas of surgical pathology and accordingly this problem is not unique to breast pathology. A good quality assurance program in the pathology department and close communications among the pathologists, radiologists and surgeons with regular clinical-pathological-radiological correlation could, in my opinion, resolve most diagnostic dilemmas and thus assure high-quality patient care.
In our pathology department, we have a policy that all malignant diagnoses rendered for the first time (including DCIS and invasive carcinoma of the breast) must be co-reviewed by two staff pathologists, and the co-review is documented in the pathology report. The same rule applies for borderline cases. When a borderline atypical breast lesion is encountered, the slides are shown to several pathologists until a consensus is reached. If a consensus diagnosis cannot be reached, the diagnosis is accompanied by a comment relaying the degree of uncertainty to the clinicians. A diagnostic term such as “highly atypical intraductal proliferation, bordering on low-grade DCIS” may be used in some cases. It is to be noted that such diagnoses are only occasionally rendered, accounting for less than 1% of all our breast biopsy diagnoses last year. In addition, we have weekly pathology-radiology correlation conferences where benign breast biopsies are reviewed with the breast images to identify discordant cases with radiology findings. We have weekly breast cancer tumor boards where breast cancer cases are discussed. Patient management decisions are made in these multidisciplinary conferences.
The clinical management of ADH and low-grade DCIS should not be drastically different knowing the lack of clear distinction between these two entities on morphology and therefore lack of concordance among pathologists in some cases. When diagnosed on core needle biopsy, ADH is followed by excisional biopsy to rule out a more significant adjacent lesion, while low-grade DCIS is followed by excision with clear margins. Excessive treatment such as radiation should be avoided in such cases.
Click Here to Manage Email Alerts