Diagnostic reliability of histological and cytological fine needle biopsies from focal liver lesions.

In 175 consecutive cases of ultrasonically detected focal liver lesions both cytological and histological fine needle biopsies were performed using two different 0.6 mm needles. The cytological and histological material was evaluated blindly and separately by two examiners, who had no access to clinical data. In 65 cases a reliable final diagnosis as defined by autopsy or surgical biopsy within 6 months could be reached. Eleven cases of benign liver lesions were correctly classified by the two examiners by both cytological and histological fine needle biopsies, resulting in a predictive value of 100% for a malignant diagnosis. The predictive value for a benign diagnosis was 85% for both examiners' evaluation of cytological fine needle biopsies, while it was 73% and 85%, respectively, for histological fine needle biopsies. At reevaluation both sampling error and difficulties in interpretation of biopsies of poor quality were responsible for the false benign diagnosis. The origin of the primary tumour was correctly diagnosed in about 50% of the cases by both techniques. Correct tumour typing could most often be made by both methods in cases of hepatocellular carcinoma, breast carcinoma, colonic carcinoma and small cell carcinoma of the lung.