Milan Interobserver Reproducibility Study (MIRST): Milan System 2018.

ABSTRACT

INTRODUCTION: Prior to the 2018 publication of the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC), a Web-based interobserver study was performed to evaluate MSRSGC reporting categories, identify cytomorphologic features that represent poor sources of agreement, and establish a baseline for future studies. MATERIAL AND METHODS: Study participants evaluated 75 images chosen from the MSRSGC image set, prior to the release of the Milan Atlas. Images spanned all diagnostic categories including typical and borderline cytomorphology. Participant demographics were collected on level of training, practice patterns, and experience. RESULTS: A total of 647 persons attempted access to the survey. Of these, 555 correctly answered the qualifying questions. Participants included 16.5% ASCP Certified Cytotechnologists, 2.8% Specialist Cytotechnologists, 5.8% IAC Certified individuals, 14.3% Anatomic (AP) Certified Pathologists, 38.9% AP and Cytopathology Certified Pathologists, and 15.3% pathology trainees. Length of participant practice varied from 0 to 54 years. In our sample, 43.4% of participants came from academic centers, 17.6% from private hospitals; and 13.3% from commercial/private laboratories. Overall, 42% of respondents agreed with the reference interpretations of salivary gland lesions. The best agreement was seen in cytopathology certified pathologists. Among the MSRSGC categories, best agreement was found in Neoplasm-Benign (58.9%) and Non-Diagnostic (49.2%) categories, followed by Malignant (48.4%). The agreement rates for Salivary Gland Lesion of Uncertain Malignant Potential (SUMP) and Suspicious For Malignancy (SFM) were 23.6% and 22.7%, respectively. CONCLUSIONS: Similar to the reproducibility studies conducted for gynecologic and urinary cytopathology, the most important factor in diagnostic reproducibility was a priori classification of image difficulty, although people with higher certifications performed better.