Abstract:

Objective To investigate the microscopy rule for nucleated red blood cell （NRBC） by SYSMEX XN-9000 automatic hematology analyzer,and to evaluate the correction formula for white blood cell （WBC） count and NRBC. Methods Samples were determined by SYSMEX XN-9000. The results must be >0 NRBC/100 WBC. Totally,716 samples were collected. Every sample was observed by microscopy after Wright-Giemsa staining. The accuracy of SYSMEX XN-9000 was evaluated,and the results were compared with those of microscopy （gold standard）. The correction formula was investigated. Results The correlation coefficient （r²）between SYSMEX XN-9000 and microscopy was 0.981. Bland-Altman plot showed a good consistency between the 2 methods. The 95% agreement limit of difference between the 2 methods was（-0.59±2.73） NRBC/100 WBC,and 23 samples （3.21%） was beyond the 95% agreement limit of difference. Receiver operating characteristic （ROC）curves were used to identify the optimal smear threshold value. The sensitivity was 87.3%,and the specificity was 76.5% at ≥1.1 NRBC/100 WBC,while the area under ROC curve was 0.88. However,the false negative rate was 23.5%,which was against the criterion of ISO 15189 （≤5%）. At ≥0.3 NRBC/100 WBC,the false positive rate was 49.6%,and the false negative rate was 2.3%. If WBC count error between SYSMEX XN-9000 and microscopy was >5% [1/3 allowable total error （TEa）],the results of SYSMEX XN-9000 should be recalculated according to those of microscopy. The correction formula was WBC=（100×WBC_{SYSMEX XN-9000}+NRBC_{SYSMEX XN-9000}×WBC_{SYSMEX XN-9000}）/（100+NRBC_microscopy）. Conclusions SYSMEX XN-9000 could determine NRBC in peripheral blood rapidly,accurately and effectively. WBC count should be recalculated,when there is difference between SYSMEX XN-9000 and microscopy.

Key words: Nucleated red blood cell, Microscopy rule, SYSMEX XN-9000 automatic hematology analyzer