Putting morphology to the test: An established classification scheme reliably stratifies salivary gland cytology by risk of malignancy with substantial interobserver agreement

Abstract

Background The Milan System for Reporting Salivary Gland Cytopathology describes several salivary gland fine‐needle aspiration cytology (SGFNAC) morphologies developed by Griffith et al. Basaloid neoplasms are pleomorphic (PB) or monomorphic with fibrillary (MBFib), hyaline (MBHy), or other (MBOther) matrix. Oncocytoid neoplasms can be pleomorphic (PO), demonstrate granular and/or vacuolated cytoplasm (OGV), or be monomorphic with mucinous (MOMuc), cystic (MOCyst), or other (MOOther) background. In the current study, the authors explore interobserver agreement (IOA) and risk of malignancy (ROM) for these subcategories. Methods The study included 169 SGFNAC cases with surgical follow‐up. Four reviewers categorized these cases using the criteria of Griffith et al. with consensus determined by majority. For all morphologic categories, IOA (using the Fleiss kappa) and ROM were calculated. Results ROMs for basaloid categories were: PB: 100% (1 of 1 case); MBHy: 71.4% (5 of 7 cases); MBFib: 50.0% (3 of 6 cases); and MBOther: 47.4% (9 of 19 cases). ROMs for oncocytoid neoplasms were: OGV: 100% (10 of 10 cases); MOMuc: 92.3% (12 of 13 cases); PO: 88.9% (8 of 9 cases); MOOther: 33.3% (5 of 15 cases); and MOCyst: 0 (0 of 1 case). The system demonstrated substantial agreement overall (κ = 0.69). For basaloid neoplasms, the IOA results were: MBHy: κ = 0.59; MBFib: κ = 0.41; MBOther: κ = 0.41; and PB: κ = 0.11. For oncocytoid neoplasms, the IOA results were: MOMuc: κ = 0.88; OGV: κ = 0.67; PO: κ = 0.63; MOOther: κ = 0.57; and MOCyst: κ = 0.18. Conclusions The SGFNAC scheme proposed by Griffith et al. and incorporated into the Milan System for Reporting Salivary Gland Cytopathology demonstrated substantial agreement overall, with particularly high agreement for the MOMuc, OGV, PO, and MBHy categories. The PB and MOCyst categories demonstrated slight agreement and may be improved by revised criteria. The PB, PO, MOMuc, and OGV categories demonstrated high ROM, and the latter 2 categories might best be classified as suspicious for malignancy.

Salivary gland neoplasms with basaloid features in the era of the Milan system for reporting salivary gland cytology: Classification and interobserver agreement

“…Interobserver agreement for categorization of samples of a basaloid morphology with varying amounts of stroma is fair with a Fleiss' κ varying from 0.11 to 0.59 depending on nuclear morphology, type of stroma and study reporting the results. 22 Approximately 20% of cases with a basaloid morphology are placed in the SUMP category but interobserver agreement for this category is only fair with a Cohen's κ of 0.28 in our study and a reported Fleiss' κ of 0.024. 23 These findings suggest that while the SUMP category may be clinically useful, it is a category with only fair to poor reproducibility and ability to predict the type of neoplasm present or the benign or malignant behavior of neoplasms with a basaloid morphology and variable amounts of stroma.…”

“…19 That study demonstrated a Fleiss' κ for overall categorization agreement of 0.69. 22 We thought that interobserver agreement might be poorer for the diagnostically difficult category of basaloid neoplasms. This hypothesis is supported by the study of Lubin, et al 22 where 33 cases classified as basaloid neoplasms with variable types of stroma disclosed Fleiss' κs varying from 0.59 to 0.11.…”

“…22 We thought that interobserver agreement might be poorer for the diagnostically difficult category of basaloid neoplasms. This hypothesis is supported by the study of Lubin, et al 22 where 33 cases classified as basaloid neoplasms with variable types of stroma disclosed Fleiss' κs varying from 0.59 to 0.11. Our study documented a Cohen's κ of 0.27 supporting our hypothesis and confirming the finding of Lubin et al 22 that diagnosis of basaloid neoplasms is difficult and associated with only fair agreement between observers.…”

“…Interobserver agreement for categorization of samples of a basaloid morphology with varying amounts of stroma is fair with a Fleiss' κ varying from 0.11 to 0.59 depending on nuclear morphology, type of stroma and study reporting the results 22 . Approximately 20% of cases with a basaloid morphology are placed in the SUMP category but interobserver agreement for this category is only fair with a Cohen's κ of 0.28 in our study and a reported Fleiss' κ of 0.024 23 .…”

Analysis of the risk of malignancy associated with the basaloid and oncocytic subtypes of the salivary gland neoplasm of unknown malignant potential (SUMP) category in the Milan system

“…Notably, the pleomorphic tumors (basaloid or oncocytic) were excluded from the SUMP category because of their almost 100% association with ROM 5,6 . Subsequently, there have been several studies that have reported the risk of neoplasm (RON) and ROM for SUMP as a category in its entirety, 2,3 but only a limited number of studies have focused primarily on determining the RON and ROM for the SUMP subcategories 7‐11 . This study was aimed at evaluating the RON and ROM for the basaloid and oncocytic subtypes of the SUMP category for primary salivary gland neoplasms.…”