Hye Ju Kang, Sun Young Kwon, Ahrong Kim, Woo Gyeong Kim, Eun Kyung Kim, Ae Ree Kim, Chungyeul Kim, Soo Kee Min, So Young Park, Sun Hee Sung, Hye Kyoung Yoon, Ahwon Lee, Ji Shin Lee, Hyang Im Lee, Ho Chang Lee, Sung Chul Lim, Sun Young Jun, Min Jung Jung, Chang Won Jung, Soo Youn Cho, Eun Yoon Cho, Hye Jeong Choi, So Yeon Park, Jee Yeon Kim, In Ae Park, Youngmee Kwon
J Pathol Transl Med. 2021;55(6):380-387. Published online October 6, 2021
Background Papillary breast lesions (PBLs) comprise diverse entities from benign and atypical lesions to malignant tumors. Although PBLs are characterized by a papillary growth pattern, it is challenging to achieve high diagnostic accuracy and reproducibility. Thus, we investigated the diagnostic reproducibility of PBLs in core needle biopsy (CNB) specimens with World Health Organization (WHO) classification.
Methods Diagnostic reproducibility was assessed using interobserver variability (kappa value, κ) and agreement rate in the pathologic diagnosis of 60 PBL cases on CNB among 20 breast pathologists affiliated with 20 medical institutions in Korea. This analysis was performed using hematoxylin and eosin (H&E) staining and immunohistochemical (IHC) staining for cytokeratin 5 (CK5) and p63. The pathologic diagnosis of PBLs was based on WHO classification, which was used to establish simple classifications (4-tier, 3-tier, and 2-tier).
Results On WHO classification, H&E staining exhibited ‘fair agreement’ (κ = 0.21) with a 47.0% agreement rate. Simple classifications presented improvement in interobserver variability and agreement rate. IHC staining increased the kappa value and agreement rate in all the classifications. Despite IHC staining, the encapsulated/solid papillary carcinoma (EPC/SPC) subgroup (κ = 0.16) exhibited lower agreement compared to the non-EPC/SPC subgroup (κ = 0.35) with WHO classification, which was similar to the results of any other classification systems.
Conclusions Although the use of IHC staining for CK5 and p63 increased the diagnostic agreement of PBLs in CNB specimens, WHO classification exhibited a higher discordance rate compared to any other classifications. Therefore, this result warrants further intensive consensus studies to improve the diagnostic reproducibility of PBLs with WHO classification.
Citations
Citations to this article as recorded by
Invasive papillary carcinoma of the breast Shijing Wang, Qingfu Zhang, Xiaoyun Mao Frontiers in Oncology.2024;[Epub] CrossRef
Recommendations for Performance Evaluation of Machine Learning in Pathology: A Concept Paper From the College of American Pathologists Matthew G. Hanna, Niels H. Olson, Mark Zarella, Rajesh C. Dash, Markus D. Herrmann, Larissa V. Furtado, Michelle N. Stram, Patricia M. Raciti, Lewis Hassell, Alex Mays, Liron Pantanowitz, Joseph S. Sirintrapun, Savitri Krishnamurthy, Anil Parwani, Giovann Archives of Pathology & Laboratory Medicine.2024; 148(10): e335. CrossRef
Encapsulated papillary carcinoma of the breast: A single institution experience Liang Xu, Qixin Mao, Qiuming Liu, Yufeng Gao, Lihua Luo, Chungen Guo, Wei Qu, Ningning Yan, Yali Cao Oncology Letters.2023;[Epub] CrossRef
High-risk and selected benign breast lesions diagnosed on core needle biopsy: Evidence for and against immediate surgical excision Aparna Harbhajanka, Hannah L. Gilmore, Benjamin C. Calhoun Modern Pathology.2022; 35(11): 1500. CrossRef