Developing and validating a highly sensitive platelet clump detection model for the Sysmex haematology analyser

BACKGROUND

Mainstream haematology analysers (HAs) are reported to have low detection sensitivity for platelet clumps. In this study, a deep learning (DL) algorithm, convolutional neural network (CNN), was implemented to detect platelet clumps.

METHODS

Adenosine diphosphate (ADP) was used to induce platelet aggregation to mimic platelet clumps detected (PCD) samples. Six types of leukocyte scattergrams were collected from the Sysmex XN-10. Then, multiple CNNs were trained and validated by scattergrams in a fivefold cross-validation (CV) method. Finally, the CNN model with the best CV accuracy was tested with practical routine work samples.

RESULTS

A total of 386 samples (190 PCD and 196 negative samples) and 4253 samples (150 PCD and 4103 negative samples) were eligible for CNN training and practical test, respectively. The CNN with the highest CV accuracy was trained by using scattergrams of side scatter (SSC) vs. forward scatter (FSC) from the white count and nucleated red blood cells (WNR) channel, whose mean area under the curve (AUC), accuracy, specificity and sensitivity were 0.968, 0.940, 0.937 and 0.942, respectively, in the CV. In the practical test, the AUC, accuracy, specificity and sensitivity of the CNN were 0.916, 0.961, 0.860 and 0.965, respectively. The dispersed spots presenting around the leucocytes in the WNR channel may be a sign of platelet clumping.

CONCLUSIONS

This study demonstrates that the CNN algorithms can identify platelet clumps based on optical information from dedicated leukocyte channels and has a higher ability to detect platelet clumps than the XN-10 device's internal algorithm under practical circumstances.

NEUT-SFL in Patients with COVID-ARDS: A Novel Biomarker for Thrombotic Events?

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an enveloped RNA virus first identified in December 2019 in Wuhan, China, and responsible for coronavirus disease 2019 (COVID-19). The ongoing COVID-19 pandemic is impacting healthcare worldwide. Patients who develop coagulopathy have worse outcomes. The pathophysiology of COVID-19 suggests a strong interplay between hemostasis and immune cells, especially neutrophils. Our purpose was to assess neutrophil fluorescence as a potential biomarker of deep vein thrombosis (DVT) in patients with COVID-acute respiratory distress syndrome (COVID-ARDS). Sixty-one patients with COVID-ARDS admitted to the four intensive care units (ICUs) of a French general hospital were included in this prospective study. Neutrophil activation was assessed by measuring neutrophil fluorescence (NEUT-Side Fluorescence Light, NEUT-SFL) with a specific fluorescent dye staining analyzed by a routine automated flow cytometer Sysmex XN-3000™ (Sysmex, Kobe, Japan). DVT was diagnosed by complete duplex ultrasound (CDU). We found that NEUT-SFL was elevated on admission in patients with COVID-ARDS (49.76 AU, reference value 46.40 AU, p < 0.001), but did not differ between patients with DVT (49.99 AU) and those without (49.52 AU, p = 0.555). NEUT-SFL is elevated in patients with COVID-ARDS, reflecting neutrophil activation, but cannot be used as a marker of thrombosis. Because neutrophils are at interface between immune response and hemostasis through release of neutrophil extracellular traps, monitoring their activation could be an interesting approach to improve our management of coagulopathy during COVID-ARDS. Further research is needed to better understand the pathophysiology of COVID-19 and identify high-performance biomarkers.

The integration of Sysmex XN-9100' WPC channel reflex testing in the detection of reactive versus malignant blood samples

INTRODUCTION

Sysmex XN-9100^™ (Sysmex, Kobe, Japan) system has an optional White Progenitor Cell (WPC) channel. While the White Differentiation (WDF) channel reports a combined flag for blasts/abnormal lymphocytes, WPC channel specifies flagging into a separate flag for each cell type or removes the flag entirely. Aim of this study was to evaluate the added value of this WPC channel in the detection of malignant samples.

METHODS

Routine blood samples analyzed on Sysmex XE-5000 with flagging for either blasts, abnormal lymphocytes, or atypical lymphocytes (n = 630) were selected for testing on Sysmex XN-9100, resulting in a reflex WPC analysis in 420 samples. All samples were microscopically evaluated with DI-60 digital cell imaging analyzer.

RESULTS

WPC reflex testing resulted in a suspect flag ("blasts" and/or "abnormal lymphocytes") in 334 samples, which was confirmed microscopically in 38% (128/334). In all true positive samples, WPC correctly classified the initial WDF flag in either "blasts" flag or "abnormal lymphocytes?" flag in 75%. Only 12% (50/420) of WDF "blasts/abnormal lymphocytes" positive samples became negative after WPC reflex testing. Subgroup analysis showed differences between the "pediatric" versus "adult" groups and the "hematological/chemotherapy" versus "nonhematological/nonchemotherapy" groups in specificity and smear reduction.

CONCLUSION

Overall, WPC reflex testing showed good sensitivity (99%); however, the specificity remains poor (29%). Using reflex WPC to the WDF channel resulted in a 12% reduction of the smear review rate. Although the WPC channel offers different interesting advantages, additional topics and a specific workflow should be applied to optimize the use of this channel.

The Negative Predictive Ability of Immature Neutrophils for Bacteremia in Children With Community-Acquired Infections

Background: Bacteremia is a serious condition. We aimed to assess the role of immature neutrophils in peripheral blood smears for prediction of bacteremia in children. Methods: In this cross-sectional study conducted in Salvador, Brazil, blood cultures collected from patients aged ≤18 years were identified. White Blood Cell count (WBC) was performed upon admission. Medical charts were reviewed and cases from the community were included. Results: Out of 833 potentially eligible patients, 263 (31.6%) were excluded. Therefore, the study group comprised 570 patients being blood collected for culture upon admission from all of them and WBC performed upon admission from 566. The median age was 2 years (IQR: 9.4 mo-5 y) and 300 (52.6%) were male. Acute respiratory infection was the most frequent diagnosis (n = 388; 68.1%), being 250 (43.9%) lower (LRTI) and 138 (24.2%) upper respiratory tract infections. Blood culture was positive in 9 (1.6%; 95% CI: 0.8-2.9%) cases, out of which 7 (2.8%) had LRTI. Streptococcus pneumoniae (n = 3), Haemophilus (n = 2), Neisseria meningitidis, viridans streptococci, Streptococcus agalactiae, and Acinetobacter baumanii (n = 1 each) were isolated. The total WBC/mm3 did not differ when children with positive or negative blood culture were compared (12,100 [IQR: 6,950-15,250] vs. 11,000 [IQR: 7,900-14,900]; P = 0.9). However, presence of any immature neutrophil was significantly more frequent among patients with bacteremia in comparison with patients without bacteremia (100% [9/9] vs. 40% [223/557]; P < 0.001). The absolute number of immature neutrophils was significantly lower among children without bacteremia (0 [IQR: 0-259] vs. 325 [IQR: 275-1,106]; P < 0.001). Overall, the area under the ROC curve of the number of immature neutrophils in regard to bacteremia was 0.82 (95% CI: 0.76-0.88; P = 0.001). Among 413 patients with absolute number of immature neutrophils <242/mm³, none had bacteremia; among 153 patients with absolute number of immature neutrophils ≥242/mm³, 9 (5.9%) had bacteremia. Absolute number of immature neutrophils ≥242/mm3 showed: sensitivity 100% (95% CI: 71.7-100%), specificity 74.1% (95% CI: 70.4-77.7%), negative predictive value 100% (95% CI: 99.3-100.0%), and positive predictive value 5.9% (95% CI: 2.9-10.5%). When only children with LRTI were analyzed, the results were similar. Conclusion: The absolute number of immature neutrophils in peripheral blood smear is a potential tool to rule out bacteremia among children with community-acquired infections.

Performance of automated digital cell imaging analyzer Sysmex DI-60

BACKGROUND

The Sysmex DI-60 system (DI-60, Sysmex, Kobe, Japan) is a new automated digital cell imaging analyzer. We explored the performance of DI-60 in comparison with Sysmex XN analyzer (XN, Sysmex) and manual count.

METHODS

In a total of 276 samples (176 abnormal and 100 normal samples), white blood cell (WBC) differentials, red blood cell (RBC) classification and platelet (PLT) estimation by DI-60 were compared with the results by XN and/or manual count. RBC morphology between pre-classification and verification was compared according to the ICSH grading criteria. The manual count was performed according to the Clinical and Laboratory Standards Institute guidelines (H20-A2).

RESULTS

The overall concordance between DI-60 and manual count for WBCs was 86.0%. The agreement between DI-60 pre-classification and verification was excellent (weighted κ=0.963) for WBC five-part differentials. The correlation with manual count was very strong for neutrophils (r=0.955), lymphocytes (r=0.871), immature granulocytes (r=0.820), and blasts (r=0.879). RBC grading showed notable differences between DI-60 and manual counting on the basis of the ICSH grading criteria. Platelet count by DI-60 highly correlated with that by XN (r=0.945). However, DI-60 underestimated platelet counts in samples with marked thrombocytosis.

CONCLUSIONS

The performance of DI-60 for WBC differential, RBC classification, and platelet estimation seems to be acceptable even in abnormal samples with improvement after verification. DI-60 would help optimize the workflow in hematology laboratory with reduced manual workload.

Performance evaluation of the automated nucleated red blood cell count of five commercial hematological analyzers

INTRODUCTION

Recent automated hematology analyzers (HAs) can identify and report nucleated red blood cells (NRBC) count as a separate population out of white blood cells (WBC). The aim of this study was to investigate the analytical performances of NRBC enumeration on five top of the range HAs.

METHODS

We evaluated the within-run and between-day precision, limit of blank (LoB), limit of detection (LoD), and limit of quantitation (LoQ) of XE-2100 and XN-module (Sysmex), ADVIA 2120i (Siemens), BC-6800 (Mindray), and UniCel DxH 800 (Beckman Coulter). Automated NRBC counts were also compared with optical microscopy (OM).

RESULTS

The limits of detection for NRBC of the BC-6800, XN-module, XE-2100, UniCel DxH 800, and ADVIA 2120i are 0.035×10⁹ /L, 0.019×10⁹ /L, 0.067×10⁹ /L, 0.038×10⁹ /L, and 0.167×10⁹ /L, respectively. Our data indicated excellent performance in terms of precision. The agreement with OM was excellent for BC-6800, XN-module, and XE-2100 (Bias 0.023, 0.019, and 0.033×10⁹ /L, respectively). ADVIA 2120i displayed a significant constant error and UniCel DxH 800 both proportional and small constant error.

CONCLUSION

Regards to NRBC counting, the performances shown by BC-6800, XN-module, and XE-2100 are excellent also a low count, ADVIA 2120i and UniCel DxH 800 need to be improved.