Interobserver variability between cytopathologists and cytotechnologists upon application and characterization of the indeterminate category in the Milan System for Reporting Salivary Gland Cytopathology

• Published in: Cancer cytopathology Vol. 128; no. 11; p. 828
• Main Authors: Viswanathan, Kartik, Patel, Ami, Abdelsayed, Mary, Rosado, Lucelina, Soong, Lauren, Margolskee, Elizabeth, Heymann, Jonas J, Goyal, Abha, Rao, Rema A
• Format: Journal Article
• Language: English
• Published: United States 01.11.2020
• Subjects: Adolescent; Adult; Aged; Aged, 80 and over; Biopsy, Fine-Needle; Child; Child, Preschool; Cytodiagnosis - standards; Female; Follow-Up Studies; Humans; Infant; Male; Middle Aged; Observer Variation; Prognosis; Reproducibility of Results; Retrospective Studies; Salivary Gland Neoplasms - classification; Salivary Gland Neoplasms - diagnosis; Salivary Glands - pathology; Young Adult; salivary gland; cytopathologist; indeterminate; Milan system; cytotechnologist; fine-needle aspiration; interobserver variability
• ISSN: 1934-6638
• DOI: 10.1002/cncy.22312

The indeterminate categories in the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC) are diagnostically challenging because of inherent heterogeneity and complexity, with wide interobserver variability (IOV). Herein, the authors explore the concordance rate (CR) between cytopathologists (CPs) and cytotechnologists (CTs) in interpreting indeterminate salivary gland lesions using the MSRSGC. Between 2011 and 2016, 86 indeterminate fine-needle aspirations had slides available for review, of which 48 had follow-up. Four CPs and 2 CTs performed an independent, blinded review of these slides and categorized them according to the MSRSGC. The CRs between CTs and CPs with the final sign-out cytopathologist (FCP) were assessed, and interobserver agreement was categorized into uniform, majority, divided, minimal, or no agreement. The overall CR with the FCP ranged from 48.8% to 60.5% for CPs and from 22.1% to 36% for CTs. IOV κ scores for the entire group were 0.314 and, with the FCP as the reference, ranged from 0.403 to 0.539 for CPs and from 0.091 to 0.254 for CTs. Uniform, majority, divided, minimal, and no agreement was noted in 12.8%, 31.4%, 38.4%, 10.5%, and 6.9%, respectively, of all cases and in 16.7%, 35.4%, 31.3%, 8.3%, and 6.3%, respectively, of the cases with follow-up. Diagnostic challenges included distinguishing lymphoma from a reactive process and distinguishing mucin from mucin-like material. CPs had modestly higher CRs compared with CTs; and, although the variable CRs highlight indeterminate IOV, the MSRSGC enables reproducibility. Characterizing larger cohorts in the indeterminate categories will further improve MSRSGC criteria. Moreover, education on the MSRSGC should include CTs and CPs to improve overall diagnostic accuracy.