This study of 160 consecutive fine-needle aspirates of localized hepatic masses analyzed the efficacy of the technique when extended into a multicenter setting. The aspirates were processed and interpreted by a single cytology service, but they were performed, under radiographic imaging, in eight different hospitals by at least 20 radiologists, few of whom had had the opportunity to acquire long-term experience. Of 108 malignancies, 101 were correctly sampled (sensitivity, 93.5%), including 30 primary hepatocellular carcinomas, 70 metastatic tumors, and 1 lymphoma. Specific benign pathologic lesions in 12 cases were amenable to cytodiagnosis. The problem of the remaining 47 samples, which contained exclusively blood, necrotic material, or normal and reactive hepatic constituents, was addressed at length. These were considered "nondiagnostic" samples in the context of aspirating focal masses. The majority had benign lesions on further study, but tumors were found in seven, all of which had been sampling errors, resulting in a negative predictive value of 82.5%. Fine-needle sampling of localized lesions, particularly neoplasms, can thus perform well in transition from single-center studies to more widespread usage in daily practice. Confident reliance on the procedure can be maintained if appropriate radiologic and cytologic facilities exist, provided that the limitations of nonneoplastic aspirates are fully understood.