engleski

The Verification Of Differential Count And Flagging Accuracy Of 6-Part Differential Hematology Analyzer Siemens Advia 2120i

Aim: The complete blood count (CBC) is one of the most frequently ordered laboratory tests in medicine. One of the components of CBC is the differential count. Performance assessment of the differential count is an indispensable part of the evaluation of the blood cells analysers (2014 ICSH Guidelines), prior to its routine implementation. Therefore, while assessing the performance of the back-up hematology analyser (HA) in our Department (Siemens Advia 2120i (Erlangen, Germany)) we aimed to investigate: i) the comparability of differential count of 6-part differential HA with the reference HA Siemens and ii) the flagging accuracy of a back-up HA in evaluation. Materials and methods: We divided our study into two parts, according to described goals. To investigate the comparability of differential count of 6-part differential HA with our routine HA Siemens Advia 2120i (reference analyser), a total of 190 K2EDTA whole blood samples (Vacutest KIMA, Italia) were analyzed for 6-part differential blood count (neutrophils, eosinophils, basophils, lymphocytes, monocytes, large unstained cells (LUC)) on both Siemens Advia 2120i HAs. For every measured parameter the mean bias (%) was calculated and compared to predefined 2016 state-of-the-art (SOTA) quality criteria by Vis and Huisman. Additionally, flagging accuracy of the back-up HA was assessed on 190 consecutive blood samples. Peripheral blood smears (stained with May-Grünwald Giemsa) were examined for all samples with manual microscopy using Motic BA410 microscope, as a reference method. Two independent experienced pathologists examined all blood smears by 100-cell differential count. The manual differential blood count was recorded, as well as the presence (or absence) of the morphology flag from the HA. Samples were divided into four categories (true positive, false positive, true negative, false negative) according to smear findings and morphology flags from the HA in evaluation. Overall flagging specificity and sensitivity was calculated, as well as the flagging accuracy for two specific flag categories: variant lymphocytes (ATYPS) and blasts (BLASTS). Calculated accuracy measures were compared to 2016 SOTA quality criteria by Vis and Huisman. Results: The observed biases for all measured parameters of the differential blood count were within the predefined quality criteria in comparison with the reference HA, ranging from -0.1% for neutrophils to 18.8% for basophils. Flagging accuracy was as follows: specificity 86.4% (95% CI 79.9 – 91.4), sensitivity 91.7% (95% CI 80.0 – 97.7), NPV 97.1% (95% CI 92.7 – 99.2), PPV 67.7% (95% CI 54.9 – 78.8). The overall flagging accuracy was 87.6%. All values were within the desirable quality criteria. The acceptable performance was also observed for specific morphology flags: sensitivity and specificity of Advia 2120i in detecting blasts (100.0% (95% CI 59.0 – 100.0) and 98.5% (95% CI 94.8 – 98.8), respectively) and variant lymphocytes (85.7 (95% CI 42.1 – 99.6)% and 95.7% (95% CI 90.8 – 98.4), respectively) were excellent. Conclusion: Differential blood count on our new back-up and reference Siemens Advia 2120i HA are completely comparable. Flagging accuracy of the new Siemens Advia 2120i HA meets all the predefined performance criteria. Siemens Advia 2120i is a reliable HA which may lead to the substantial reduction of the manual blood smear reviews in everyday routine.

differential blood count, blood morphology, hematology analyser, flagging efficiency

nije evidentirano