OBJECTIVE: To evaluate effusion cytology on routine smears in patients with hepatocellular carcinoma. STUDY DESIGN: Filed smears from 106 patients with autopsy-proven hepatocellular carcinoma were retrospectively reviewed for the presence of malignant cells. Morphologic patterns, as well as immunochemical reactivity for a panel of antibodies, were analyzed when feasible. RESULTS: Malignant cells were identified in 10 cases of ascites and 1 of pleural fluid. Positive smears were variably cellular, with paucicellular, round or linear aggregates of polygonal cells. The cytoplasm was usually evident and very stainable. The nuclei were hyperchromatic or vesicular, with inconspicuous nucleoli. Reactive changes were frequent. Malignant cells were positive for keratins (7/8) and erythropoietin (4/8) and negative for carcinoembryonic antigen (0/8). Concomitant cirrhosis was present in 103 cases (97.1%); all the noncirrhotic cases had malignant cells identifiable in the effusion. Antemortem clinical diagnosis of malignancy was made in 59 cases (55.7%); cytology was the only source of the morphologic diagnosis in most of them. CONCLUSION: These results indicate that although the yield of hepatocellular carcinoma in effusion cytology is limited, it may be important in the initial assessment of the disease, given the ineligibility of most patients for invasive procedures and the equivocal features of instrumental investigations. Immunocytochemistry may further assist in differentiating doubtful cases.
OBJECTIVE: To evaluate effusion cytology on routine smears in patients with hepatocellular carcinoma. STUDY DESIGN: Filed smears from 106 patients with autopsy-proven hepatocellular carcinoma were retrospectively reviewed for the presence of malignant cells. Morphologic patterns, as well as immunochemical reactivity for a panel of antibodies, were analyzed when feasible. RESULTS: Malignant cells were identified in 10 cases of ascites and 1 of pleural fluid. Positive smears were variably cellular, with paucicellular, round or linear aggregates of polygonal cells. The cytoplasm was usually evident and very stainable. The nuclei were hyperchromatic or vesicular, with inconspicuous nucleoli. Reactive changes were frequent. Malignant cells were positive for keratins (7/8) and erythropoietin (4/8) and negative for carcinoembryonic antigen (0/8). Concomitant cirrhosis was present in 103 cases (97.1%); all the noncirrhotic cases had malignant cells identifiable in the effusion. Antemortem clinical diagnosis of malignancy was made in 59 cases (55.7%); cytology was the only source of the morphologic diagnosis in most of them. CONCLUSION: These results indicate that although the yield of hepatocellular carcinoma in effusion cytology is limited, it may be important in the initial assessment of the disease, given the ineligibility of most patients for invasive procedures and the equivocal features of instrumental investigations. Immunocytochemistry may further assist in differentiating doubtful cases.