针吸细胞学诊断肝细胞癌的标准探讨

An approach to the cytologic diagnostic criteria of hepatocellular carcinoma by fine needle aspiration biopsy

目的探讨肝细胞癌（ＨＣＣ）的针吸细胞学诊断标准。方法分析６１例ＨＣＣ针吸细胞学涂片，并与１９例肝转移瘤和１６例肝非肿瘤病变对照，找出与ＨＣＣ有关的形态改变；用逐步Ｌｏｇｉｓｔｉｃ回归方法选择区别ＨＣＣ与肝转移瘤、ＨＣＣ与肝非肿瘤病变的最有用的形态特征。结果观察到与ＨＣＣ有关的９种细胞学改变：涂片中组织碎片多（９１．８％），有小梁状排列的细胞团（９１．８％），细胞团内有窦内皮细胞分界（４７．５％），细胞呈多边形，核中位（８２．０％），核浆比例升高（９５．１％），大核仁（５５．７％），胆汁颗粒（３１．１％），大小胞浆内空泡（２７．９％）和背景中有不典型改变的肝细胞裸核（８８．５％）。在ＨＣＣ和转移瘤，选择出多边形细胞中位核、小梁状排列和胞浆内胆汁３个因素，它们能明确９８．４％的ＨＣＣ的肝细胞起源和９４．７％转移瘤的非肝细胞来源。在ＨＣＣ和非肿瘤病变，选择出丰富的组织碎片，小梁状排列，核浆比例升高，不典型改变的肝细胞裸核４种因素，它们诊断ＨＣＣ的敏感性为１００％，特异性为９３．８％。结论丰富的组织碎片，多边形细胞中位核，核浆比例升高，小梁状排列，不典型改变的肝细胞裸核和胞浆内胆汁是细胞学识别ＨＣＣ的有用特征。

Objective To identify useful cytologic features for diagnosis of hepatocellular carcinoma (HCC). Methods Fine needle aspiration(FNA )smears from 61 patients with proven HCC, 19 patients with metastatic tumors in the liver and 16 patients with non neoplastic lesion of the liver were reviewed in order to detect the relevant cytologic changes associated with HCC. Step wise logistic regression analysis was done to select useful cytologic features in discriminating HCC from metastatic tumors and non neoplastic lesions of liver. Results Nine cytologic features observed were significantly associated with HCC: abundant tissue fragments(91.8%); trabecular pattern(91.8%);sinusoidal endothelial cells (47.5%);polygonal cells with centrally placed nuclei(82.0%);increased nuclear to cytoplasmic ratio(95.1%); large nucleoli(55.7%);bile granules(31.1%); cytoplasmic vacuoles(27.9%) and atypical naked hepatocyte nuclei(88.5%). In distinquishing HCC from metastatic tumors, polygonal cells with centrally placed nuclei, bile granules and trabecular pattern were useful cytologic features. Of the 61 patients with HCC, the logistic model correctly predicted 60(98.4%) as having HCC ,while of the 19 patients with metastatic tumors,18(94.7%) were correctly predicted. In distinquishing HCC from non neoplastic liver disease, abundant tissue fragments, trebecular pattern, increased nuclear to cytoplasmic ratio and atypical naked hepatocyte nuclei were selected. When the four criteria were used, the sensitivity of diagnosing HCC was 100% and the specificity was 93.8%. Conclusion Abundant tissue fragments, polygonal cells with centrally placed nuclei, increased nuclear to cytoplasmic ratio, trebecular pattern, atypical naked hepatocyte nuclei and bile granules are useful features for identifying HCC by FNA biopsy.