Abstract: Objective　To understand the application of microscopic scanners in detecting the false positive and negative results of acid-fast bacilli in clinical smear samples, to optimize the instrument detection parameters, and to continuously improve the detection accuracy. Methods　The data from Guangzhou Chest Hospital were studied from December 2019 to June 2020. For the clinical samples from sputum, bronchoscopy flushing fluid, pleural effusion, cerebrospinal fluid, and joint effusion, direct smears were performed on their glass slides provided with the instrument as required. The dried smears were stained for auramine O fluorescence. Manual retest was performed on all the instrument-reported positive smears, and about 15% of the instrument-reported negative smears were randomly selected. The false positive and negative rates were described as categorical data. For those with false negative results, the position of the manually retested acid-fast bacilli in the smear was marked with a microscope slide clip coordinate ruler, and the locations inside and outside the scanning area of the instrument were estimated. Results　Among the 5 284 smears, 1 290 cases were reported positive by the instrument, and 850 cases were reported positive by manual retest, with a false positive rate of 34.11% (440/1 290). Six hundred and sixty-five negative smear samples reported by the instrument were randomly checked, and 100 cases were reported positive by manual retest, with a false negative rate of 15.04%. The false negative results included 38 cases ±, 50 cases 1+, 6 cases 2+, 4 cases 3+, and 2 cases 4+. Among the false-negative smears, 42 cases of acid-fast bacteria were found in the scanning area of the instrument, and 58 cases outside the area. The false negative rates of the two were 6.32% (42/665) and 8.72% (58/665), respectively. Conclusion　The detection of acid-fast bacilli in clinical samples by microscopic scanners has certain false positive and negative rates. The instrument scanning parameters should be continuously improved, and the laboratory manual retest procedures should be strengthened.

Key words: Acid-fast bacteria, Auramine O fluorescence staining, Smear examination, Microscopic scanning

摘要： 目的　了解应用显微扫描仪检测临床样本中抗酸菌的假阳性与假阴性情况，以优化仪器检测参数，并不断提高检测准确度。方法　研究时间为2019年12月至2020年6月。资料来源于广州市胸科医院。采用试验研究方法对痰、纤维支气管镜冲洗液、胸腔积液、脑脊液、关节腔积液等临床样本在仪器配套的玻片上按要求进行直接涂片，对干燥涂片行金胺O荧光染色，对所有仪器报告阳性涂片、约15%的仪器报告阴性涂片（随机抽取）进行人工复检。采用分类资料统计描述假阳性率和假阴性率，对假阴性者，应用显微镜玻片夹坐标尺标注人工复检到的抗酸菌在涂片中所在位置，估计其在仪器扫描区域内外的情况。结果　在5 284张涂片中，仪器报告阳性1 290例，人工复检报告阳性850例，假阳性率为34.11%（440/1 290）；随机抽查仪器报告阴性涂片665张，人工复报告阳性100例，假阴性率为15.04%，假阴性结果包括38例±、50例1+、6例2+、4例3+、2例4+；在假阴性涂片中，人工复检查见抗酸菌在仪器扫描区域内者42例、区域之外者58例，两类假阴性率分别为6.32%（42/665）和8.72%（58/665）。结论　显微扫描仪检测临床样本中抗酸菌存在较高程度的假阳性与一定程度的假阴性，应不断改进和优化仪器扫描参数，并强化实验室人工复检程序。

关键词:

"> 抗酸菌, 金胺O荧光染色, 涂片检查, 显微扫描