Interobserver Reliability of Milan and Modified Milan System for Reporting Salivary Gland Cytopathology: An Institutional Experience.

Abstract

In 2018, the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC) was published, defining a diagnostic categorization scheme. However, this system of classification was criticized due to its suboptimal diagnostic accuracy and low interobserver reliability. For these reasons, the modified Milan system was recently proposed by a few authors claiming it to be more clinically reliable. The present study aimed to analyze the interobserver reliability of MSRSGC and modified MSRSGC. A total of 100 salivary gland fine-needle aspirations were reviewed over a 1-year period and classified by two independent observers according to MSRSGC and modified MSRSGC. Interobserver reproducibility was estimated using observed agreement and chance-corrected agreement (Cohen's kappa). An overall chance corrected agreement of 0.37 (fair) was obtained for the original Milan system and 0.41 (moderate) for the modified one. In addition, subcategories of modified MSRSGC performed better than the MSRSGC in terms of interobserver reliability. The present study suggests that modified MSRSGC should be accepted worldwide as the clinical relevance of any categorization scheme requires diagnostic accuracy along with excellent interobserver reproducibility.