► Click here to download the original poster (PDF)
Mathias CHEA1, Marion EVEILLARD2, Michelle HEGE3, Andrea TORRES3, Desrick J. PINKNEY3, Amin MOHAMMAD3, Eldad HOD4
1Nîmes University Hospital (FRANCE), 2Nantes University Hospital (FRANCE), 3Baylor Scott & White Healthcare (USA), 4Columbia University Irving Medical Center (USA)
The progress of hematology automation and the achieved good levels of precision and the accuracy in cell counting are incontestable.
The examination and identification of thousands of cells in each sample, the improvement in technologies and the incorporation of new parameters provide reliable and applicable information for diagnosis in several clinical conditions.
We evaluated the accuracy and clinical performance of HORIBA Yumizen H1500 and H2500 hematology analyzers as compared to SYSMEX XN-10 hematology analyzers and to manual slide microscopy examination as a reference method for differential analysis with whole blood samples, as per CLSI H20-A2 recommendations.
For whole blood comparison Yumizen H2500 versus XN-10, the study was performed in 4 sites (2 in USA and 2 in France) collecting 972 samples (K2 EDTA): 270 pediatric samples and 702 adult samples. On both analyzers, Complete Blood Count, Differential, Reticulocyte and Optical Platelet analysis were performed.
For whole blood method comparison versus reference method (manual microscopy), the study was performed in the same 4 sites collecting 421 samples.
Two qualified morphology examiners counted 200 cells each.
Regarding correlation in-between Yumizen H1500/2500 and XN-10 analyzers, the results demonstrated a very good correlation (table 1).
Regarding correlation in-between Yumizen H1500/2500, XN-10 and manual microscopy (table 2), all the correlation results (r²) showed a very good correlation.
| H2500 vs XN (CBC) | H2500 vs XN (DIF) | H2500 vs XN (RET) | H2500 vs XN (PLT-Ox) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Parameter | N | r2 | Parameter | N | r2 | Parameter | N | r2 | Parameter | N | r2 | |||
| WBC | 943 | 0.994 | NRBC% | 932 | 0.857 | RET% | 940 | 0.813 | PLT-Ox | 810 | 0.967 | |||
| TNC | 944 | 0.994 | NRCB# | 932 | 0.904 | RET# | 940 | 0.86 | LPF/IPF | 815 | 0.7 | |||
| RBC | 944 | 0.995 | LYM% | 823 | 0.971 | CRC | 922 | 0.879 | ||||||
| HGB | 943 | 0.996 | LYM# | 823 | 0.987 | IRF | 926 | 0.65 | ||||||
| HCT | 944 | 0.967 | MON% | 845 | 0.884 | RHCC | 925 | 0.53 | ||||||
| MCV | 951 | 0.8 | MON# | 845 | 0.912 | |||||||||
| MCH | 943 | 0.954 | NEU% | 834 | 0.969 | |||||||||
| RDW-CV | 927 | 0.77 | NEU# | 834 | 0.936 | |||||||||
| RDW-SD | 935 | 0.698 | EOS% | 931 | 0.969 | |||||||||
| PLT | 944 | 0.991 | EOS# | 931 | 0.952 | |||||||||
| MPV | 879 | 0.811 | BAS# | 900 | 0.814 | |||||||||
| BAS% | 900 | 0.508 | ||||||||||||
| IMG% | 837 | 0.753 | ||||||||||||
| IMG# | 837 | 0.964 | ||||||||||||
Table 1: Correlation H2500 versus XN-10 for CBC, DIF, Reticulocyte, Optical Platelet
| Correlation (R2) with manual microscopy | XN-10 | H2500 |
| LYM% | 0.94 | 0.96 |
| MON% | 0.67 | 0.8 |
| NEU% | 0.94 | 0.95 |
| EOS% | 0.97 | 0.96 |
| BAS% | 0.71 | 0.72 |
| IMG% | 0.54 | 0.86 |
| NRBC% | 0.85 | 0.90 |
Table 2: Comparison correlation of H2500 and XN-10 with manual microscopy for DIF
The comparison of the results analyzed in-between Yumizen H2500 with XN-10 and in-between Yumizen H2500 with manual microscopy does not show any significant difference.
The correlation of Yumizen H2500 with manual microscopy are even showed a better statistical correlation compared to XN-10 for Lym% (0.96/0.94), Mon% (0.8/0.67), Neu% (0.95/0.94), Bas% (0.72/0.71), IMG% (0.82/0.54) and NRBC% (0.90/0.85), which are useful parameters in deciding specific pathologies like leukemia’s, infections.
Do you have any questions or requests? Use this form to contact our specialists.
