A Correlation between the Cytology and Histology of Serous Effusions at a Teaching Hospital in South-East Nigeria

Abstract

Background: Cytological evaluation of body fluids is an important diagnostic test for various malignant and benign conditions. Objectives: Our study aims to analyse the various body fluids received in our department over a ten-year period. It also seeks to determine the accuracy and significance of specimen volumes of fluid cytology specimens in diagnosing malignancy in a resource-poor setting. Methodology: This is a retrospective study carried out in the Department of Anatomic Pathology of a teaching hospital in South-East, Nigeria. Histopathologic tissue results, if available were also retrieved from the archives as gold standard. Results: Within the period under review, 358 pleural fluid specimens, 358 ascitic fluid specimens and 2 pericardial effusion specimens were retrieved. Of all effusion cytology specimens, 126 (17.5%) were cytologically malignant. More specifically, 48 (13.4%) of pleural effusion specimens, and 78 (21.8%) of ascitic fluid specimens were malignant. The most common histologically diagnosed cancer observed in patients with malignant pleural effusion was breast cancer, while for ascitic fluid, it was ovarian cancer. For all cytologic diagnoses, the sensitivity was 37.0%, specificity 87.2%, PPV 84.4%, NPV 42.5% and accuracy 54.5%. With respect to the specimen volume of the histologically confirmed malignant cases received for cytological examination, 114 (52.1%) of the specimen were <10mls, only 36 (16.4%) were ≥20mls. Those cytologically positive for malignant cells had a median volume of 10.0ml, while those cytologically negative for malignant cells had a median volume of 8.8ml. In addition, there was an incremental increase in the percentage of cytologically malignant effusions with increased volume of specimen used for the analysis from 34.2% for specimens <10mls to 50% for specimen volumes ≥20mls. However, a Chi-squared test showed there is no statistically significant difference between these (P = 0.213). Conclusion: The sensitivity of effusion cytology in this study is at the lower end of the spectrum. This may be related among other factors to the suboptimal specimen volume received for evaluation and lack of resources for cell block and immunocytochemistry.