Atypical Ductal Hyperplasia at 25 Years – Interobserver and Intraobserver Variability
RK Jain, R Mehta, R Dmitrov, L Larsson, P Musto, K Hodges, TM Ulbright, E Hattab, N Agaram, MT Idrees, S Badve. Indiana University, Indianapolis, IN
Background: Atypical ductal hyperplasia (ADH) was formulated as a diagnostic category in 1985 and is lies in between usual ductal hyperplasia (UDH) and ductal carcinoma in situ (DCIS). These categories are associated with varying levels of risk for development of invasive breast cancer. Both morphological and size criteria have been used to differentiate these entities. Rosai et al have demonstrated poor inter-observer reproducibility, while Schnitt et al have demonstrated marked improvement if standard criteria were used. The aims of the study were to investigate the inter-observer and intra-observer variations, and evaluate the improvement in reproducibility by using high molecular weight keratin (HMWK) expression analysis.
Design: Nine pathologists reviewed 81 cases of breast proliferative lesions in a three stage process. The lesion was marked and the pathologists were instructed to only evaluate the marked area as per their usual diagnostic criteria with the exception of the size criterion and assign diagnoses of UDH, ADH or DCIS to each lesion. In the second stage, the H&E slides were re-labeled and re-evaluated, while in the third stage an immunostain for HMWK was also provided. Concordance was evaluated at each stage of the study.
Results: The overall reproducibility among the 9 pathologists for diagnosing the 81 proliferative breast lesions was fair (Kappa value =0.342), whereas when each of the 9 pathologists was compared to each of the other 8 pathologists, the kappa value ranged from 0.152 to 0.568 (slight to moderate agreement). The intra-observer kappa value ranged from 0.561 to 0.883 (moderate to strong). Complete agreement among 9 pathologists was achieved in only 9 (11.1%) cases, atleast 8 agreed in 20 (24.7%) cases and 7 or more agreed in 38 (47.0%) cases. Out of 81 cases, maximum agreement among pathologists was observed in 34 lesions of UDH, 29 lesions of ADH and 13 lesions of DCIS. Equivocal agreement was obtained for five lesions. Following IHC stain a significant improvement in the inter-observer concordance was observed.
Conclusions: ADH still remains a diagnostic dilemma with wide variation in both inter- and intra-observer reproducibility among pathologists. The addition of IHC stains improves the concordance for the diagnosis of these difficult lesions.
Wednesday, March 24, 2010 9:30 AM
Poster Session V # 34, Wednesday Morning