B-014 Performance evaluation of the automated cell image analyzer DI-60 for leukocyte differential count

Abstract

Background The Sysmex DI-60 system (Sysmex, kobe, Japan) is an automated cell image analyzer that captures cell images and analyzes them. The aim of this study was to evaluate the performance of DI-60 for leukocyte differential count in comparison with manual differential count or XN-20. Methods A total of 205 samples were analyzed. The agreement between DI-60 pre-classification and post-verification by medical technicians was determined. The correlation and ability to identify clinically important abnormal cells of DI-60 post-verification were evaluated. Results The overall agreement of DI-60 pre-classification was 84.7% (Table 1). The correlation between DI-60 post-verification and manual differential counts were exellent (r2 > 0.85) for neutrophil, lymphocyte, monocyte and eosinophil, but was very low for basophil (r2 = 0.3759). The correlation between DI-60 post-verification and leukocyte differential counts by XN-20 were exellent (r2 > 0.90) for neutrophil, lymphocyte and eosinophil. However, monocyte and basophil showed a relatively low correlation of r2 = 0.6343 and r2 = 0.3118, respectively but was very low for basophil (r2 0.3759). The ability of DI-60 post-verification to identify clinically important abnormal cells including blast, promyelocyte, myelocyte, metamyelocyte and NRBC demonstrated excellent efficiency (80.8 to 96.3%) except for NRBC (51.0%). Overall sensitivity was 63.7-88.2%, and specificity was 84.3-95.5% excluding NRBC (43.7%). Conclusions DI-60 showed excellent pre-classification and generally high correlation compared to manual leukocyte differential counts. Although additional verification with PBS review by experienced medical technician may still be required, the performance DI-60 makes it an efficient screening tool in clinical laboratories.