Application of algorithmic cytomorphological assessment and immunocytochemistry with the international system for reporting serous fluid cytopathology on pericardial fluid cytology

Abstract

Aims: The international system for reporting serous fluid cytopathology (ISRSFC) set forth a five-tiered reporting system with comprehensive validation on pleural and peritoneal fluid cytology. An algorithmic approach for cytomorphological assessment and immunocytochemistry was also described in ISRSFC. Limited data on pericardial fluid are supportive but would benefit from further investigation. Methods: Consecutive pericardial fluid cytology over a 4-year period was reviewed by multiple board-certified pathologists according to the ISRSFC. Cytomorphology and immunocytochemistry were assessed sequentially, with respective diagnostic performances computed and compared. Literature review was performed. Results: In total 358 specimens, including 53 with immunocytochemistry available, were reviewed. There were 137 benign and 221 malignant (MAL) cases. The risks of malignancy were 23.5% non-diagnostic (ND), 29.2% negative for malignancy (NFM), 56.0% atypia of undetermined significance (AUS), 82.6% suspicious for malignancy (SFM) and 99.2% (MAL) for cytomorphological assessment, improving to 23.5% (ND), 29.1% (NFM), 56.8% (AUS), 78.9% (SFM) and 99.3% (MAL) incorporating immunocytochemistry. Ten cases (2.8%) received a change in diagnosis after review of immunocytochemistry. All revisions of diagnostic category were appropriate upgrades/downgrades referenced against clinical information. Cytomorphological typing was accurate for adenocarcinoma (n=81/83, 97.6%), while other carcinomas and lymphomas required immunocytochemistry. Certain subcategories within AUS and SFM pertaining to bland indeterminate epithelial cells or mucinous material were not seen for pericardial fluid. Conclusions: The ISRSFC shows robust diagnostic performance for pericardial fluid cytology. For pericardial effusion, disease composition and applicable cytological subcategories differ from its peritoneal and pleural counterparts. Incorporating immunocytochemistry by an algorithmic approach improves diagnostic accuracy. Cytomorphology is accurate for identifying adenocarcinomas, but further typing necessitates immunocytochemistry is necessary.