Reliability of assessing morphologic features with prognostic significance in cytology specimens of epithelioid diffuse pleural mesothelioma and implications for cytopathology reporting.

Background: The World Health Organization incorporates morphologic features with prognostic significance in the 2021 classification of epithelioid diffuse pleural mesothelioma (E-DPM). Although cytology specimens are often the first and occasionally the only specimen available for patients with DPM, these features have not yet been investigated in cytology.

Methods: Nuclear atypia, pleomorphic features, necrosis, and architectural patterns were retrospectively assessed in 35 paired cytology and concurrent/consecutive surgical pathology specimens of E-DPM. Agreement between pairs was determined via unweighted κ scores. Discordant cases were re-reviewed to determine the reasons for disagreement.

Results: Interpretation of nuclear atypia in cytology was concordant with histology in all cases (κ = 1.000; p < .001). The presence of pleomorphic features and necrosis was concordant in 97.1% (κ = 0.842; p < .001) and 85.7% (κ = 0.481; p = .001) of paired cases, respectively. Assessment of architectural patterns in cytology showed only slight agreement with histology (κ = 0.127; p = .037). In cytology cases (n = 23) with cell block material available, assessment of nuclear atypia and the presence of pleomorphic features showed perfect agreement (κ = 1.000; p < .001, each), the presence of necrosis showed moderate agreement (κ = 0.465; p = .008), and assessment of architectural patterns showed slight agreement (κ = 0.162; p = .15) in paired specimens. Most disagreements were due to sampling differences between cytology and histology specimens.

Conclusions: Although complete nuclear grading of E-DPM is not possible given the unreliability of mitotic counts in cytology, assessment of nuclear atypia in cytology specimens is shown to be reliable. Identification of pleomorphic features and necrosis is also reliable despite occasional sampling issues. Assessment of architectural patterns is more limited in cytology.