International journal of laboratory hematology最新文献

Background: The peripheral blood film (PBF) analysis traditionally relies on manual microscopy (MM), a labour-intensive method with inter-observer variability. This study evaluates Blade (a semi-supervised AI model) and CellaVision DM9600 (commercial benchmark) against MM in automated leukocyte classification.

Methods: PBFs from 168 patients were prepared using automated staining and scanned digitally. Blade, trained on 185 412 cells (75 435 labelled, 109 977 unlabelled) via ResNet34 and RetinaNet architectures, underwent pseudo-labelling and AdamW optimisation. Performance was evaluated on 1675 cells against MM using the concordance correlation coefficient (CCC), Bland-Altman analysis, Deming/Passing-Bablok regression and diagnostic accuracy measures across nine leukocyte subtypes.

Results: When evaluated individually against MM, both systems showed high agreement. Blade achieved excellent correlation for common cells (neutrophils: ccc = 0.988; lymphocytes: ccc = 0.985; eosinophil: ccc = 0.953) and comparable results to CellaVision for monocytes (ccc = 0.852 vs. 0.847) and basophils (ccc = 0.762 vs. 0.794). Blade performed better for metamyelocytes (ccc = 0.905 vs. 0.756) and showed higher sensitivity for monocytes (75% vs. 63%) and myelocytes (87% vs. 74%). Regression analysis showed slopes close to 1.0 for most cell types, with Blade displaying narrower Limits of Agreement in Bland-Altman analysis. Both systems achieved 100% sensitivity for blasts and reactive lymphocytes. Overall macro-averaged performance was comparable between Blade (sensitivity 89.2%, specificity 96.3%) and CellaVision (86.3% and 96.7%).

Conclusion: Blade and CellaVision demonstrated strong concordance with MM, validating their clinical utility. Blade's semi-supervised learning confers marginal advantages in rare cell detection and stability, highlighting AI's potential to enhance diagnostic accuracy. While both systems reduce labour and variability, Blade's performance has potential for integration into haematology workflows. Future validation in diverse cohorts is recommended.

Introduction: Point-of-care testing (POCT) for hematology enables clinicians to access actionable results during patient visits, supporting timely decisions such as transfusions or chemotherapy adjustments. However, POCT analyzers must demonstrate clinical performance comparable to centralized laboratory systems. We conducted a proof-of-concept evaluation of the HemoScreen (PixCell Medical, Israel) hematology POCT analyzer across diverse clinical settings in an integrated healthcare system.

Methods: EDTA whole blood specimens from inpatient wards, emergency departments, outpatient clinics, and infusion centers underwent routine CBC testing on a Sysmex XN analyzer (Sysmex Corporation, Japan) at the central laboratory. Residual, de-identified samples were tested within 10 min on a HemoScreen analyzer. All parameters and instrument flags were recorded.

Results: From 9/15/2020 to 11/20/2020, 199 samples were analyzed, including 49 from the emergency department and 50 each from inpatient, outpatient, and infusion center settings. HGB, RBC, PLT, WBC, and absolute neutrophil count showed strong linear correlations (R² = 0.91-0.98) and no significant differences between instruments (p > 0.29). HCT and MCV values were significantly lower on HemoScreen (bias -5% and -3.6%; p = 0.010 and p < 0.001, respectively), with modestly reduced correlations (R² = 0.80-0.89). Neutrophil percentage was higher on HemoScreen (bias +9.0%, p < 0.001), whereas monocyte counts were significantly lower (bias -50%, p < 0.001).

Conclusion: HemoScreen demonstrated generally acceptable agreement with the Sysmex XN for most parameters. However, systematic differences in HCT, MCV, neutrophil, and monocyte results warrant further investigation to assess clinical impact before broader implementation.

Background: Interpretation of laboratory data is a comparative process that requires reliable reference intervals (RIs) to support clinical decision-making. Although population-based reference intervals (popRIs) are commonly used for this purpose, more accurate interpretation can be achieved using RIs derived from an individual's own data-that is, personalized reference intervals (prRIs). In this study, we estimated prRIs for common hematology parameters and compared them with popRIs indirectly derived from laboratory data.

Method: popRIs were estimated for a total of 17 complete blood count (CBC) subparameters using patient laboratory data. prRIs for the same parameters were calculated using longitudinal data from 200 healthy individuals, based on a prediction interval model. Additionally, the reference interval index (RII) was calculated as the ratio of the range of prRIs to the range of popRIs for each measurand.

Results: prRIs calculated for CBC parameter differed from the corresponding popRIs. The median of RII varied from 0.41 for eosinophils to 2.97 for platelet distribution width (PDW).

Conclusion: We concluded that the prRIs of the CBC parameters differ from their corresponding popRIs, and that prRIs should be used for a more accurate interpretation of patients' CBC results.

Background: A fixed D-dimer cut-off threshold of 0.500 μg/mL is used in Blackrock Health Hermitage Clinic (BHHC) and most Haematology laboratories across Ireland, however, international guidelines have evolved over the past decade, and a growing body of evidence supports the use of age-adjusted D-dimer thresholds. This study evaluates the appropriateness of age-adjusted cut-offs in the patient population served by BHHC. Notably, inconsistent adherence to a formal hospital policy for the screening and diagnosis of venous thromboembolism (VTE) was observed throughout the clinical areas; the potential benefits of implementing such a policy are discussed herein. The integration of age-adjusted D-dimer thresholds with a validated clinical probability assessment tool may reduce unnecessary diagnostic imaging in patients investigated for VTE and prevent D-dimer testing in those deemed to have a 'likely' pre-test probability of VTE.

Methods: A cross-sectional study was conducted to determine age-related D-dimer thresholds in a representative sample of patients attending BHHC. D-dimer measurements were performed using the ACL TOP 350 coagulation analyser with HemosIL D-dimer HS 500 reagents. Statistical analysis was carried out using IBM SPSS (v29) and Microsoft Excel. Ethical approval was obtained from the BHHC Clinical Governance Committee.

Results: An age-related increase in D-dimer levels was observed. Applying an age-adjusted threshold (defined as patient age × 0.01 μg/mL for individuals aged 50 years and older) reduced false-positive results by 12.6% in this age group.

Conclusion: Age-adjusted D-dimer thresholds are recommended, provided they are combined with a validated clinical probability assessment tool, such as the two-level Wells score. This approach improves diagnostic specificity and may reduce unnecessary imaging, particularly in older adults.

Introduction: Thrombin generation assay (TGA) is an assay that activates the coagulation cascade using a trigger reagent containing tissue factor (TF) and phospholipids (PL), and the generated thrombin is detected using a fluorogenic substrate. The standardisation of assay conditions and validation of in-house methods are recommended for research investigations. This study aimed to investigate the influence of TF and PL concentrations on TGA parameters in normal and abnormal plasma samples and the usefulness of commercial quality control materials for standardisation.

Methods: Five samples were employed for the TGA: Pooled Normal Plasma and Ci-Trol 1 as normal samples as well as Ci-Trol 2, Ci-Trol 3 and 2.5-fold diluted Pooled Normal Plasma. Trigger reagents were prepared with various TF and PL concentrations. The final TF concentrations were 0-50 pM with 4 μM PL, while the final PL concentrations were 0-40 μM with 5 pM TF. The lag time, peak height, time-to-peak and endogenous thrombin potential were calculated.

Results: The lag time and time-to-peak prolonged, and the peak height decreased as TF concentration decreased. In contrast, PL concentration primarily affected peak height. Clear thrombin peaks were confirmed in Ci-Trol 2 and Ci-Trol 3.

Conclusions: Changes in TF concentration affected the lag time time-to-peak, and peak height, while alterations in PL concentration primarily affected the peak height in both normal and abnormal plasma samples. These findings are crucial for assay standardisation; Ci-Trol 2 and Ci-Trol 3 may serve as quality control materials.

Introduction: Thalassemia is a major global health concern, as declared by the World Health Organization (WHO). Accurate screening of heterozygotes is of paramount importance for diagnostic guidance and disease mitigation.

Methods: Multiclass machine-learning (ML) models were developed to classify alpha, beta, and delta/beta-thalassemia heterozygotes, individuals with iron deficiency microcytic anemia and healthy individuals. Complete blood count (CBC) data were derived from 1518 individuals, simultaneously measured in four different hematological analyzers [Sysmex K-1000 (Sysmex), Cell-Dyn Sapphire (Abbott), ADVIA 2120 (Siemens), BC-6800 (Mindray)]. The Recursive Feature Elimination (RFE) method was applied to investigate the optimal number and combination of hematological parameters required for adequate model training, and their importance was determined using the SHAP method. Random Forest classifier and Active Learning (AL) method were used to evaluate diagnostic accuracy and enhance the model's performance, respectively. Dimensionality reduction techniques were applied for visualization purposes. Additionally, the existence of subclusters within each class was investigated using unsupervised techniques.

Results: Four hematological parameters (MCH, MCV, HGB, and RDW) are sufficient to classify basic types of microcytosis, achieving an accuracy of over 80%. Adding further parameters to the training process can improve the stability and reliability of the model when applied to unseen data, but does not increase test accuracy beyond 90%. Each microcytosis cluster is discriminated by assigning different weights to selected hematological parameters, forming five distinct clusters in the 2D plane (heterozygous alpha-, beta-, and delta/beta-thalassemia, iron deficiency anemia, and normal individuals). The greatest overlap between clusters occurs between alpha- and beta-thalassemia. The alpha-thalassemia cluster appears to have a more diffuse scatter on the 2D plane, but two distinct subclusters are identified within this class, characterized by the differential expression of specific parameters (PCT, MPV, P-LCR, PDW, MCV, MCH, and RBC). The developed model can estimate the probabilities of classifying a new case into the core established clusters and, in the case of alpha-thalassemia, the possible subcluster.

Conclusion: The application of ML models for the automated differential diagnosis of microcytosis, using CBC data, revealed that the diagnostic accuracy is not proportionally dependent on the use of an increasing number of hematological parameters. A developed model is proposed which can correctly classify a new case in the core clusters or subclusters of patients with microcytosis.