Beschreibung
The point-of-care hematology system scil vCell 5 (hereafter scil POCA), was introduced as a cost-effective alternative to currently available laser-based hematology instruments commonly used in veterinary practice. The laser- and impedance-based analyzer provides a complete blood count, including a 5-part differential leukocyte count, with leukocyte cytograms and instrument flags serving as interpretive aids for numerical results. The objective of the study was to evaluate the scil POCA for canine and feline species. The evaluation included a method comparison study to assess correlation and mean differences (bias) and to determine linearity, carryover, and precision. To assess the acceptability of the scil POCA measurement results, observed total error (TEo) was calculated from the coefficient of variation (CV) and %bias and compared with the allowable total error (TEa) according to the recommendations of the American Society for Veterinary Clinical Pathology (ASVCP).
For the method comparison study, fresh blood samples from 192 healthy and diseased dogs and 159 cats were analyzed, and the results of the scil POCA were compared with those of the corresponding reference methods (i.e., manual differential count, spun hematocrit [PCV], and ADVIA 2120 hematology analyzer [Siemens]). In addition, method comparison was performed considering instrument indications (flags), leukocyte cytograms, and software adaptations of the scil POCA, i.e., data were analyzed before and after a validation process (exclusion of measurement results that were flagged by the instrument or showed an abnormal cytogram) and a software update.
Excellent linearity was found for white blood cells (WBC), red blood cells (RBC), platelets (PLT), and hemoglobin (HGB) over a wide measuring range.
CVs lower than 2.5% for WBC, RBC, hematocrit (HCT), HGB, and neutrophil granulocytes (NEU), and less than 5% (dog) and 8% (cat) for PLT were detected for values within the reference range. Higher CVs (7-31%) were evident for absolute numbers of lymphocytes (LYM), monocytes (MON), and eosinophil granulocytes (EOS) in both species. The recommended quality requirements for CV (CV < 0.25 TEa) were met for all hematological parameters except for LYM in dogs and LYM, MON, EOS, and PLT in cats.
For both species, an excellent correlation (rs = 0.82-1.00) was detected between the methods for WBC and RBC, hematocrit, HGB, and PLT, except for feline PLT (rs = 0.79). Regarding the erythrocyte indices, the correlation of erythrocyte distribution width (RDW, rs = 0.82-0.85), mean corpuscular volume (MCV, dog rs = 0.82), and mean corpuscular hemoglobin (MCH, dog rs = 0.85) was good, whereas the correlation of mean corpuscular hemoglobin concentration (MCHC) was poor (rs = < 0.32). The correlation between the scil POCA and ADVIA 2120 for differential leukocyte counts was good to excellent (rs = 0.81-0.97) for all parameters except for feline MON (rs = 0.21-0.63) and canine MON% (rs = 0.50).
With regards to the Bland-Altman analysis, there were high proportional biases ranging from -6.4% (dog) to -8% (cat) for WBCs and -5.6% for feline RBCs. In dogs, a high constant positive bias for HCT was detected when the scil POCA was compared with the ADVIA 2120 (7.3%) and with the microhematocrit method (12.9%). Smaller biases between the methods could be detected for the feline HCT (-2.8% and 4.9%). Correction of the known systematic error of the ADVIA 2120 for HGB led to minimal bias between the analyzers for both species (dog -0.7%, cat -2.3%). For the erythrocyte indices, except for RDW, moderate (MCV 3.1-5.8%) to marked (MCH18-24.9%, MCHC 12.3-22%) positive biases could be shown for both species. For PLT, there was a bias of -3.4% in canine and of 8% in feline specimens. Regarding the differential leukocyte count, the lowest deviations of bias were seen for NEU (dog: -5.7 to 0.8%; cat: 1.5-9.4%) when comparing the results of the POCA-Diff to the reference methods.
Quality requirements (TEo < TEa) were met for WBC (TEo = 8.6-11.1%; TEa = 20%) and RBC (TEo = 3.5-7%; TEa = 10%), HCT (TEo = 5.7-9.4%; TEa = 10%), PCV (cat TEo = 7.8%; TEa = 10%), MCV (cat TEo = 5.1%; TEa = 7%), and PLT (TEo = 13.1-24.1%; TEa = 25%). After bias correction, both species also showed compliance with the quality goals for HGB (TEo = 2.2-3.8%; TEa = 10%). Higher TEos, exceeding the recommended TEa, were detected for MCV and PCV in dogs and for MCHC in both species. Although not completely satisfactory for the differential count, the scil POCA provides reliable results in compliance with ASVCP quality goals for canine and feline
NEU (TEo = 5.3-10.6%, TEa = 15%) and EOS (TEo = 67.1-83%, TEa = (90)-50%) considering at least one reference method.
Exclusion of samples with potentially invalid differential counts, either flagged by scil POCA or identified on the basis of manual cytogram review, resulted in increased correlation for most parameters but had minimal effect on bias, thus failing to meet quality objectives for LYM and MON. Reanalysis of the method comparison data after software adaptations resulted in a significant reduction of bias for EOS in both species (dog > cat) comparing the POCA-Diff and the M-Diff.
In conclusion, considering the results of the present evaluation study, the point-of-care hematology system scil vCell 5 can be considered as suitable for routine use in veterinary practices. Overall, the performance of the scil POCA is comparable to other laser-based in-house hematology systems available on the veterinary market, which provide a complete blood count including a 5-part differential count. This study demonstrated that the scil POCA was able to meet the quality requirements recommended by the ASVCP for the main variables of the CBC in dogs and cats and for values within the reference range. The observed deviations of HGB and HGB-dependent parameters are attributable to the methodology of ADVIA 2120 as a reference method. As described for other hematology analyzers, the scil POCA has weaknesses in meeting the quality requirements for rare cell populations of the differential count, so that manual verification of the automated differential count by blood smear, not only with respect to LYM and MON, but also in case of morphologic instrument flags or abnormalities of the leukocyte cytogram is recommended. Specific characteristics of dotplots were found to be useful in detecting hematologic abnormalities, such as the presence of immature or atypical cells, although diagnostic accuracy remains to be evaluated in follow-up studies.
