Abstract:Objectives
Platelet clumps present in anticoagulant specimens may generate a falsely decreased platelet count and lead to an incorrect diagnosis. A clear understanding of the ability of a haematology analyser (HA) to detect platelet clumps is important for routine work in the clinical laboratory.
Methods
Citrate-anticoagulated whole-blood samples were collected from various patients as a negative group. Adenosine diphosphate … Show more
“…Figure 4 shows that the CNN algorithm has no positive threshold determined by the platelet count, which is an algorithmic limitation of the XN-10 that was discussed in our previous study. 11 The above results indicate that the CNN can accurately detect platelet clumps using the optical information from the dedicated leukocyte channels and that the SSC and FSC of the WNR channel are the most suitable data for CNN training.Figure 2.The average ROC curves of the training groups. The average ROC curves of the CNNs trained by scattergrams from the WNR and WDF channels are plotted as solid and dashed lines, respectively.…”
Section: Resultsmentioning
confidence: 80%
“…The platelet aggregation rate (PAR) of each imitated PCD sample was calculated using the formula PAR = (Pb – Pa)/Pb, in which Pb represents the platelet count before ADP induction and Pa represents the platelet count after ADP induction. 11 The criteria for a PCD sample to be used for training were as follows: (a) The PAR was higher than 10%. (b) At least 1 platelet clump (containing a minimum of 3 platelets) was observed in the blood smear within 50 high magnification fields.…”
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 4,253 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.
“…Figure 4 shows that the CNN algorithm has no positive threshold determined by the platelet count, which is an algorithmic limitation of the XN-10 that was discussed in our previous study. 11 The above results indicate that the CNN can accurately detect platelet clumps using the optical information from the dedicated leukocyte channels and that the SSC and FSC of the WNR channel are the most suitable data for CNN training.Figure 2.The average ROC curves of the training groups. The average ROC curves of the CNNs trained by scattergrams from the WNR and WDF channels are plotted as solid and dashed lines, respectively.…”
Section: Resultsmentioning
confidence: 80%
“…The platelet aggregation rate (PAR) of each imitated PCD sample was calculated using the formula PAR = (Pb – Pa)/Pb, in which Pb represents the platelet count before ADP induction and Pa represents the platelet count after ADP induction. 11 The criteria for a PCD sample to be used for training were as follows: (a) The PAR was higher than 10%. (b) At least 1 platelet clump (containing a minimum of 3 platelets) was observed in the blood smear within 50 high magnification fields.…”
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 4,253 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.
“…In a study by Xu et al, 12 on the samples with low platelet counts showed that the total sensitivities of the platelet, CLP and platelet abnormal flags on platelet histogram/platelet distribution curve were 80.01% and 100%, respectively.…”
Background: Automated hematology analyzers have become mainstream of complete blood count (CBC) during the past two decades; they are potential, more accurate to produce results of CBCs. However, often automated analyzers can give false results too, specially, falsely low platelet count due to platelets aggregates, which have to be confirmed on peripheral smears.
Aims and Objectives: The aim and objective of the study was to diagnose thrombocytopenia (TCP) cases and differentiate them from pseudothrombocytopenia (PTCP) cases by analyzing platelet histogram and to determine the sensitivity and specificity of platelet histograms for diagnosing TCP and PTCP.
Materials and Methods: The present study was conducted in the Department of Pathology at Shyam Shah Medical College, Rewa, M.P., after obtaining ethical clearance from the Institutional Ethics Committee. It was prospective study conducted for a period of 15 months from January 2021 to May 2022 on 1000 samples.
Results: The sensitivity and specificity of platelet histogram flags and presence of multiple peaks to diagnose TCP and PTCP cases were calculated as 73.60% and 93.60%, respectively. The present study also interpret that mean platelet volume and platelet distribution width cannot use as indicator to differentiate between TCP and PTCP.
Conclusion: The present study concludes that analysis and interpretation of histogram more specifically platelet histogram flagging provide clue for early detection of PTCP cases prior peripheral smear examination and also helpful in differentiating them from true TCP cases. These platelet histogram flagging can be used as a screening parameter for the detection of PTCP and these are helpful in preventing unnecessary stress for clinician, patients, and their relatives. However, the peripheral smear examination will remain the gold standard to differentiate TCP from, PTCP.
“…Various studies have been conducted to explore PTCP detection, with varying sensitivity and specificity found with different devices. The sensitivity and specificity of the Sysmex instrument (Sysmex, Kobe, Japan) were 0.626 and 0.947, respectively, when platelet aggregation was induced using adenosine diphosphate in citrate-anticoagulated whole blood samples [17]. In a comparative analysis of the usefulness of platelet-clump flagging among the Sysmex, Beckman Coulter, and ADVIA (Siemens Healthcare Diagnostics, Eschborn, Germany) CBC analyzers, the Sysmex analyzer demonstrated the highest sensitivity of 67% [18].…”
Background: EDTA-induced pseudothrombocytopenia (PTCP) during whole blood collection requires significant laboratory resources to obtain accurate results. We evaluated platelet-deaggregation function in EDTA-induced PTCP and platelet-clump flagging by the BC-6800Plus hematology analyzer using integrated digital image analysis.Methods: We prospectively collected 132 whole blood samples suspected of platelet clumping (102 in EDTA and 30 in sodium citrate) from 88 individuals. We compared platelet counts determined using the platelet count by impedance (PLT-I) function of the DxH 900 hematology analyzer and the PLT-I or optical platelet count (PLT-O) function of the BC-6800Plus. Platelet clumping was verified through manual inspection and the MC-80 digital image analyzer.Results: Among the 132 whole blood samples, 43 EDTA samples showed platelet clumping. The DxH 900 PLT-I and BC-6800Plus PLT-I results demonstrated a strong correlation (r = 0.711) for the EDTA samples but only a moderate correlation with the BC-6800Plus PLT-O results (r = 0.506 and 0.545, respectively). The BC-6800Plus PLT-O results were consistent with the sodium citrate platelet counts, with a median dissociation rate of 102.5% (range, 74.9%-123.1%). The DxH 900 and BC-6800Plus analyzers had sensitivity values of 0.79 and 0.72, respectively, for platelet-clump flagging. When integrating the MC-80 digital image analysis results, the sensitivity of BC-6800Plus improved to 0.89 (standard mode) or 1.0 (PLT-Pro mode).Conclusions: BC-6800Plus PLT-O measurement results are close to the actual values obtained by platelet deaggregation with PTCP samples. Integrating the BC-6800Plus with a digital imaging analyzer effectively improved the diagnosis of PTCP and reduced the requirement for additional laboratory procedures.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
