Wielders J P, Beunis M H, van Wersch J W
J Clin Chem Clin Biochem. 1986 Jul;24(7):471-80. doi: 10.1515/cclm.1986.24.7.471.
The ability to screen out abnormal leukocyte differential counts by the H 6000TM and the Hematrak 590TM has been evaluated, using the visual 100 cell differential as a reference. Optimization of the discrimination levels for both instruments by the use of linear regression procedures and "Receiver Operating Characteristics (ROC) curves" and its impact on their performance characteristics is discussed. The overall sensitivity for 314 outpatient/188 inpatient samples was 0.84/0.94 for the H 6000 and 0.84/0.88 for the Hematrak, respectively. The overall specificities for out-/inpatient samples were found to be 0.61/0.44 and 0.47/0.30 for the H 6000 and the Hematrak, respectively. Both instruments had very similar screening abilities for the detection of abnormal leukocyte differential counts. However, the H 6000 flags correlated better with the visual findings than the Hematrak flags as demonstrated by the parameter-sensitivity and predictive value calculations and by the ROC curves.
以视觉100细胞分类为参考，评估了H 6000TM和Hematrak 590TM筛选异常白细胞分类计数的能力。讨论了通过使用线性回归程序和“受试者工作特征（ROC）曲线”对两种仪器的判别水平进行优化及其对性能特征的影响。对于314例门诊/188例住院患者样本，H 6000的总体灵敏度分别为0.84/0.94，Hematrak的总体灵敏度分别为0.84/0.88。发现H 6000和Hematrak门诊/住院患者样本的总体特异性分别为0.61/0.44和0.47/0.30。两种仪器在检测异常白细胞分类计数方面具有非常相似的筛选能力。然而，如参数灵敏度、预测值计算和ROC曲线所示，H 6000标记比Hematrak标记与视觉结果的相关性更好。
