Can automated blood film analysis replace the manual differential? An evaluation of the CellaVision DM96 automated image analysis system

A critical analysis of CellaVision systems in the modern hematology laboratory

OBJECTIVE

This review aims to provide a comprehensive analysis of the current literature regarding the use of CellaVision digital morphology systems and to assess the emerging applications, their potential to supplement current diagnostic pathways, and-importantly-appreciate their practical and perceived limitations.

METHODS

A manual literature review was conducted to identify relevant journal articles and published abstracts as they relate to the application of CellaVision systems to hematology laboratory practice in human patients.

RESULTS

CellaVision systems can characterize cellular morphology with overall a high degree of accuracy-in particular, for neutrophils; lymphocytes; and common red blood cell changes, including target cells. Challenges remain, however, with the detection of particular important findings, including immature granulocytes and red blood cell agglutination. The application of CellaVision systems to emerging areas such as telepathology and parasitology are evolving, with an increasing volume of literature highlighting the technology's utility.

CONCLUSION

CellaVision systems and associated digital technologies are poised to play an increasingly important role in hematology laboratories, with promises of enhanced accuracy and improved workflows. Several critical deficiencies highlight the necessity for continued research to support their transition into routine clinical practice.

Real-World Application of Digital Morphology Analyzers: Practical Issues and Challenges in Clinical Laboratories

Digital morphology (DM) analyzers have advanced clinical hematology laboratories by enhancing the efficiency and precision of peripheral blood (PB) smear analysis. This review explores the real-world application of DM analyzers with their benefits and challenges by focusing on PB smear analysis and less common analyses, such as bone marrow (BM) aspirates and body fluids (BFs). DM analyzers may automate blood cell classification and assessment, reduce manual effort, and provide consistent results. However, recognizing rare and dysplastic cells remains challenging due to variable algorithmic performances, which affect diagnostic reliability. The quality of blood film as well as staining techniques significantly influence the accuracy of DM analyzers, and poor-quality samples may lead to errors. In spite of reduced inter-observer variability compared with manual counting, an expert's review is still needed for complex cases with atypical cells. DM analyzers are less effective in BM aspirates and BF examinations because of their higher complexity and inconsistent sample preparation compared with PB smears. This technology relies heavily on artificial intelligence (AI)-based pre-classifications, which require extensive, well-annotated datasets for improved accuracy. The performance variation across platforms in BM aspirates and rare-cell analysis highlights the need for AI algorithm advancements and DM analysis standardization. Future clinical practice integration will likely combine advanced digital platforms with skilled oversight to enhance diagnostic workflow in hematology laboratories. Ongoing research aims to develop robust and validated AI models for broader clinical applications and to overcome the current limitations of DM analyzers. As technology evolves, DM analyzers are set to transform laboratory efficiency and diagnostic precision in hematology.

Diagnosis and typing of leukemia using a single peripheral blood cell through deep learning

Leukemia is highly heterogeneous, meaning that different types of leukemia require different treatments and have different prognoses. Current clinical diagnostic and typing tests are complex and time-consuming. In particular, all of these tests rely on bone marrow aspiration, which is invasive and leads to poor patient compliance, exacerbating treatment delays. Morphological analysis of peripheral blood cells (PBC) is still primarily used to distinguish between benign and malignant hematologic disorders, but it remains a challenge to diagnose and type these diseases solely by direct observation of peripheral blood(PB) smears by human experts. In this study, we apply a segmentation-based enhanced residual network that uses progressive multigranularity training with jigsaw patches. It is trained on a self-built annotated dataset of 21,208 images from 237 patients, including five types of benign white blood cells(WBCs) and eight types of leukemic cells. The network is not only able to discriminate between benign and malignant cells, but also to typify leukemia using a single peripheral blood cell. The network effectively differentiated acute promyelocytic leukemia (APL) from other types of acute myeloid leukemia (non-APL), achieving a precision rate of 89.34%, a recall rate of 97.37%, and an F1 score of 93.18% for APL. In contrast, for non-APL cases, the model achieved a precision rate of 92.86%, but a recall rate of 74.63% and an F1 score of 82.75%. In addition, the model discriminates acute lymphoblastic leukemia(ALL) with the Ph chromosome from those without. This approach could improve patient compliance and enable faster and more accurate typing of leukemias for early diagnosis and treatment to improve survival.

Robustness assessment of an automated AI-based white blood cell morphometric analysis system using different smear preparation methods

INTRODUCTION

Numerous AI-based systems are being developed to evaluate peripheral blood (PB) smears, but the feasibility of these systems on different smear preparation methods has not been fully understood. In this study, we assessed the impact of different smear preparation methods on the robustness of the deep learning system (DLS).

METHODS

We collected 193 PB samples from patients, preparing a pair of smears for each sample using two systems: (1) SP50 smears, prepared by the DLS recommended fully automated slide preparation with double fan drying and staining (May-Grunwald Giemsa, M-G) system using SP50 (Sysmex) and (2) SP1000i smears, prepared by automated smear preparation with single fan drying by SP1000i (Sysmex) and manually stained with M-G. Digital images of PB cells were captured using DI-60 (Sysmex), and the DLS performed cell classification. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were used to evaluate the performance of the DLS.

RESULTS

The specificity and NPV for all cell types were 97.4%-100% in both smear sets. The average sensitivity and PPV were 88.9% and 90.1% on SP50 smears, and 87.0% and 83.2% on SP1000i smears, respectively. The lower performance on SP1000i smears was attributed to the intra-lineage misclassification of neutrophil precursors and inter-lineage misclassification of lymphocytes.

CONCLUSION

The DLS demonstrated consistent performance in specificity and NPV for smears prepared by a system different from the recommended method. Our results suggest that applying an automated smear preparation system optimized for the DLS system may be important.

Advances in Hematology Analyzers Technology

The evolution of complete blood count (CBC) methodology from manual calculations to sophisticated high throughput hematology analyzers is the focus of this article. In recent years, hematology testing has greatly benefitted from the combination of various technologies with automated neural networks. In addition to an increasing complexity of the laboratory instrumentation, there is a demand on point of care CBC testing with its benefits and drawbacks. This article highlights exciting advancements of hematology testing from the past to the present and into the future.

Digital Imaging and AI Pre-classification in Hematology

A leukocyte differential of peripheral blood can be performed using digital imaging coupled with cellular pre-classification by artificial neural networks. Platelet and erythrocyte morphology can be assessed and counts estimated. Systems from a single vendor have been used in clinical practice for several years, with other vendors' systems, in a development. These systems perform comparably to traditional manual optical microscopy, however, it is important to note that they are designed and intended to be operated by a trained morphologist. These systems have several benefits including increased standardization, efficiency, and remote-review capability.

Application of image recognition technology in pathological diagnosis of blood smears

Traditional manual blood smear diagnosis methods are time-consuming and prone to errors, often relying heavily on the experience of clinical laboratory analysts for accuracy. As breakthroughs in key technologies such as neural networks and deep learning continue to drive digital transformation in the medical field, image recognition technology is increasingly being leveraged to enhance existing medical processes. In recent years, advancements in computer technology have led to improved efficiency in the identification of blood cells in blood smears through the use of image recognition technology. This paper provides a comprehensive summary of the methods and steps involved in utilizing image recognition algorithms for diagnosing diseases in blood smears, with a focus on malaria and leukemia. Furthermore, it offers a forward-looking research direction for the development of a comprehensive blood cell pathological detection system.

Evaluating the Sysmex DI-60 Integrated Slide Processing System's impact on hematology differential turnaround times and patient care: Real-world implementation experience in a large Veterans Affairs hospital

OBJECTIVES

This quality improvement study conducted at the Kansas City VA Medical Center in Kansas City, Missouri, investigated the Sysmex DI-60 Integrated Slide Processing System's ability to improve hematology turnaround times when integrated into daily practices. It further addressed potential patient care factors associated with changes in turnaround times.

METHODS

Three months of manual and Sysmex DI-60 patient data were examined between May 2022 and February 2023. White blood cell (WBC) ranges, turnaround times, working hours, and study months were analyzed using 2-tailed unpaired t testing and percentage change. The number of specimens in these categories was further analyzed using 2-tailed, 2-sample proportion testing.

RESULTS

This quality improvement study indicated that the Sysmex DI-60 system produced a statisitcally significant reduction in turnaround times overall and for various ranges of WBCs plus work shifts. The most statistically significant improvement in turnaround times occurred for WBC concentrations less than 2.0 × 103/µL and concentrations within the reference range. In addition, the off shifts experienced a notable improvement in turnaround times.

CONCLUSIONS

The Sysmex DI-60 system substantially decreases turnaround times for differentials, thus potentially benefiting patient care by providing prompt results. It is possible that reducing turnaround times could expedite emergency department admissions and discharges as well as enhance care for the oncology department's patients. It could also lead to more timely results for patients with false-positive flags by the Sysmex analyzer, which may also help with clinician decision-making.

Diagnostic performance of the CellaVision preclassification neutrophil count - time to bypass the reclassification?

OBJECTIVES

The objective of this study was to perform a method comparison between the CellaVision preclassification neutrophil count and the reclassification neutrophil count performed by trained laboratory technicians, and to evaluate the diagnostic performance of the preclassification neutrophil count at clinical decision levels.

METHODS

We retrospectively identified patient samples through 2019-2022 in which the differential count was performed on Cellavision (n = 4,354). Data on sample characteristics and leukocyte- and differential counts was extracted from the electronic medical journal. For each sample, data containing the pre- and reclassification leukocyte classification, respectively, was extracted from the Cellavision software. Method comparison between the pre-and reclassification neutrophil count was performed using Bland Altman analysis. Diagnostic performance of the preclassification neutrophil count was evaluated according to four pre-specified categories of results with the reclassification as reference method.

RESULTS

The median difference between the pre- and reclassification neutrophil count was 0.044 x 109/L. The preclassification neutrophil count categorised 95.6% of all samples correctly according to the four categories. The sensitivity, specificity, positive predictive value and negative predictive value for detecting neutrophilia > 7.00 x 109/L was 98.8%, 97.2%, 95.8%, and 99.2%, respectively. In samples with leukopenia (n = 543), the sensitivity, specificity, positive predictive value and negative predictive value for detecting severe neutropenia (< 0.50 x 109/L) was 97.7%, 99.1%, 98.6%, and 98.5%, respectively.

CONCLUSION

The diagnostic performance of the CellaVision preclassification neutrophil count was satisfactory. The preclassification neutrophil count may be released to the electronic medical journal to improve turnaround time and benefit laboratory management.

Performance comparison of two automated digital morphology analyzers for leukocyte differential in patients with malignant hematological diseases: Mindray MC-80 and Sysmex DI-60

BACKGROUND

The MC-80 (Mindray, Shenzhen, China), a newly available artificial intelligence (AI)-based digital morphology analyzer, is the focus of this study. We aim to compare the leukocyte differential performance of the Mindray MC-80 with that of the Sysmex DI-60 and the gold standard, manual microscopy.

METHODS

A total of 100 abnormal peripheral blood (PB) smears were compared across the MC-80, DI-60, and manual microscopy. Sensitivity, specificity, predictive value, and efficiency were calculated according to the Clinical and Laboratory Standards Institute (CLSI) EP12-A2 guidelines. Comparisons were made using Bland-Altman analysis and Passing-Bablok regression analysis. Additionally, within-run imprecision was evaluated using five samples, each with varying percentages of mature leukocytes and blasts, in accordance with CLSI EP05-A3 guidelines.

RESULTS

The within-run coefficient of variation (%CV) of the MC-80 for most cell classes in the five samples was lower than that of the DI-60. Sensitivities for the MC-80 ranged from 98.2% for nucleated red blood cells (NRBC) to 28.6% for reactive lymphocytes. The DI-60's sensitivities varied between 100% for basophils and reactive lymphocytes, and 11.1% for metamyelocytes. Both analyzers demonstrated high specificity, negative predictive value, and efficiency, with over 90% for most cell classes. However, the DI-60 showed relatively lower specificity for lymphocytes (73.2%) and lower efficiency for blasts and lymphocytes (80.1% and 78.6%, respectively) compared with the MC-80. Bland-Altman analysis indicated that the absolute mean differences (%) ranged from 0.01 to 4.57 in MC-80 versus manual differential and 0.01 to 3.39 in DI-60 versus manual differential. After verification by technicians, both analyzers exhibited a very high correlation (r = 0.90-1.00) with the manual differential results in neutrophils, lymphocytes, and blasts.

CONCLUSIONS

The Mindray MC-80 demonstrated good performance for leukocyte differential in PB smears, notably exhibiting higher sensitivity for blasts identification than the DI-60.

Performance Assessment of Sysmex DI-60: Is Digital Morphology Analyzer Reliable for White Blood Cell Differentials in Body Fluids?

BACKGROUND

Few studies have evaluated digital morphology (DM) analyzers on body fluids (BF). We evaluated the performance of a DM analyzer, Sysmex DI-60 (Sysmex, Kobe, Japan) for white blood cell (WBC) differentials in BF samples.

METHODS

In five BF samples (two pleural fluids and three ascites) containing a single, dominant cell type (>80%, neutrophils, lymphocytes, macrophages, abnormal lymphocytes, and malignant cells in each sample), we evaluated the precision of the DI-60 and compared the WBC differentials and turnaround times (TAT) between DI-60 and manual counting.

RESULTS

The precision of the DI-60 pre-classification and verification was excellent (%CV, 0.01-3.16%). After verification, the DI-60 showed high sensitivity, specificity, and efficiency (ranges: 90.8-98.1%, 96.8-97.9%, and 92.5-98.0%, respectively) for the dominant cell types in neutrophil- and lymphocyte-dominant samples. For all samples, the DI-60 and manual counting showed high correlations for major cell types (neutrophils, lymphocytes, macrophages, and others, r = 0.72 to 0.94) after verification. The agreement between the pre-classification and verification of the DI-60 was strong in the neutrophil-dominant sample (κ = 0.81). The DI-60 showed a significantly longer TAT (min: s) than manual counting for all samples (median TAT/slide: 6:28 vs. 1:53, p < 0.0001), with remarkable differences in abnormal lymphocyte- and malignant cell-dominant samples (21:05 vs. 2:06; 12:34 vs. 2:25).

CONCLUSIONS

The DI-60 may provide reliable data in neutrophil- and lymphocyte-dominant BF samples. However, it may require longer times and higher workloads for WBC differentials especially in BF samples containing atypical cells. Further improvement would be needed before applying DM analyzers for routine clinical practice in BF analysis.

Performance of the digital cell morphology analyzer MC-100i in a multicenter study in tertiary hospitals in China

BACKGROUND

This study investigated the performance of the MC-100i, a pre-commercial digital morphology analyzer utilizing a convolutional neural network algorithm, in a multicentric setting involving up to 11 tertiary hospitals in China.

METHODS

Blood smears were analyzed by MC-100i, verified by morphologists, and manually differentiated. The classification performance on WBCs and RBCs was evaluated by comparing the classification results using different methods. The PLT and PLT clump counting performance was also assessed. The total assay time including hands-on time was evaluated.

RESULTS

The agreements between pre- and post-classification were high for normal WBCs (κ > 0.96) and lower for overall abnormal WBCs (κ = 0.90). The post-classification results correlated well with manual differentials for both normal and abnormal WBCs (r > 0.93), except for basophils (r = 0.8480) and atypical lymphocytes (r = 0.8211). The clinical sensitivity and specificity of each RBC abnormality after verification were above 90 % using microscopy reviews as the reference. The PLTs counted by the MC-100i before and after verification correlated well with those measured by the PLT-O mode (r = 0.98). Moreover, PLT clumps were successfully classified by the analyzer in EDTA-dependent pseudothrombocytopenia blood samples.

CONCLUSIONS

The MC-100i is an accurate and reliable digital cell morphology analyzer, offering another intelligent option for hematology laboratories.

Evaluation of canine and feline leukocyte differential counts obtained with the scil vCell 5 compared to the Advia 2120 hematology analyzer and a manual method

The vCell 5 (scil Animal Care), a point-of-care hematology analyzer (POCA), was recently introduced to veterinary laboratories. This laser- and impedance-based analyzer is capable of providing a CBC with 5-part WBC differential count (Diff) along with WBC cytograms and flags serving as interpretation aids for numerical results. We compared the scil POCA-Diff to reference methods (i.e., manual differential count, Advia 2120 hematology analyzer [Siemens]) for canine and feline blood samples and considered WBC cytograms and flags. Total observed error (TEo), calculated from CV and bias%, was compared to total allowable error (TEa). Data were analyzed before and after a review process (exclusion of flagged and samples with invalid cytograms). For both species, correlation was good-to-excellent (rs = 0.81-0.97) between both analyzers for all variables, except for feline monocytes (rs = 0.21-0.63) and canine monocyte% (rs = 0.50). Smallest biases were seen for neutrophils (dog: -5.7 to 0.8%; cat: 1.5-9.4%) with both reference methods. Quality requirements (TEo < TEa) were fulfilled for canine and feline neutrophils (TEo = 5.3-10.6%, TEa = 15%) and eosinophils (TEo = 67.1-83%, TEa = (90)-50%) considering at least one reference method. Our review process led to mildly higher rs-values for most variables. Although not completely satisfactory, the scil POCA provides reliable results in compliance with ASVCP quality goals for canine and feline neutrophils and eosinophils. Analyzer flag and cytogram analysis served as useful tools for QA, indicating the necessity for manual review of blood smears, and contributed to improvement of scil POCA performance.

Development and Pilot Testing of a Comprehensive Mobile Application to Assist Cell Count Determination During Peripheral Smear and Bone Marrow Examination

BACKGROUND

In the modern era of complete blood count analysis, manual differential count is performed whenever 'flags' are generated by an automated hematology analyzer. Traditionally, tally counters with five or eight keys are used for manual differential count. A few mobile applications are available to perform this task; however, the application features and cell representation are limited.

OBJECTIVES

The primary objective of our study was to develop an indigenous, comprehensive mobile application to assist with manual blood cell differential count. The secondary objective was to measure the usability of a newly developed application among undergraduate medical students.

MATERIALS AND METHODS

A new mobile application was developed using a Java development kit, Version 11.0.13 (Oracle Corporation, Austin, USA) in Android Studio Dolphin (2021.3.1) (Google, California, USA). The application content was validated by three pathologists with more than five years of experience. The app's usability was tested among 60 participants using a validated mHealth App Usability Questionnaire (MAUQ). The questionnaire had 18 items covering three domains: ease of use, interface & satisfaction, and usefulness.

RESULTS

The newly developed application supports peripheral smear WBC differential count, platelet count, reticulocyte count, malaria parasite quantification, and bone marrow differential count. During usability testing, the app was easy to use in 95% (57/60) of participants, time-efficient in 91.7% (55/60), and helpful for healthcare practice learning in 96.7% (58/60). The total mean score was 6.11, indicating high usability.

CONCLUSION

A comprehensive mobile application to assist manual differential count with adequate cell representation was developed. The mobile application was easy to use, time-efficient, and valuable among the study participants.

White blood cell differentials performance of a new automated digital cell morphology analyzer: Mindray MC-80

INTRODUCTION

The manual differential count has been recognized for its disadvantages, including large interobserver variability and labor intensiveness. In this light, automated digital cell morphology analyzers have been increasingly adopted in hematology laboratories for their robustness and convenience. This study aims to evaluate the white blood cell differential performance of the Mindray MC-80, the new automated digital cell morphology analyzer.

METHODS

The cell identification performance of Mindray MC-80 was evaluated for sensitivity and specificity using pre-classification and post-classification of each cell class. The method comparison study used manual differentials as the gold standard for calculating Pearson correlation, Passing-Bablok regression, and Bland-Altman analysis. In addition, the precision study was performed and evaluated.

RESULTS

The precision was within the acceptable limit for all cell classes. Overall, the specificity of cell identification was higher than 95% for all cell classes. The sensitivity was greater for 95% for most cell classes, except for myelocytes (94.9%), metamyelocytes (90.9%), reactive lymphocytes (89.7%), and plasma cells (60%). Pre-classification and post-classification results correlated well with the manual differential results for all the cell types investigated. The regression coefficients were greater than 0.9 for most cell classes except for promyelocytes, metamyelocytes, basophils, and reactive lymphocytes.

CONCLUSION

The performance of Mindray MC-80 for white blood cell differentials is reliable and seems to be acceptable even in abnormal samples. However, the sensitivity is less than 95% for certain abnormal cell types, so the user should be aware of this limitation where such cells are suspected.

Analytical validation of white blood cell differential and platelet assessment on the Sysmex DI-60 digital morphology analyzer

INTRODUCTION

Digital morphology analyzers are increasingly replacing light microscopy in laboratory hematology practice. This study aimed to perform the analytical validation of the white blood cell (WBC) differential and of reliability of platelet assessment on Sysmex DI-60 (Kobe, Japan).

METHODS

Validation included determination of within-run and between-run precision for WBC differential according to the CLSI EP15-A3 protocol, accuracy and method comparison with light microscopy and with the automated WBC differential from the Sysmex XN-10 hematology analyzer, reliability of platelet clump detection and platelet count estimation.

RESULTS

Standard deviations of both pre- and post-classification mostly satisfied manufacturer's criteria for imprecision. Accuracy assessment revealed that only eosinophil count (1.4%) in one peripheral blood smear (PBS) remained outside the declared range (2-10%) after reclassification. Method comparison between DI-60 and light microscopy yielded Spearman's correlation coefficients from 0.37 (basophils) to 0.94 (neutrophils and lymphocytes), minor proportional difference for bands, constant difference for monocytes, both constant and proportional difference for lymphocytes and statistically significant biases for bands, lymphocytes, monocytes and basophils. Diagnostic sensitivity (Se) and specificity (Sp) of DI-60 in detecting immature/pathological cells were 88.7% (95%CI:81.1-94.0) and 83.0% (95%CI:78.7-86.7), respectively, with the area under the curve (AUC) of 0.86 (95%CI:0.82-0.89). Agreement in detection of platelet clumps was 94.8% (kappa coefficient = 0.67, 95%CI:0.53-0.80). Se and Sp of DI-60 to detect platelet clumps were 65.7% (95%CI: 47.8-80.9) and 96.9% (95%CI: 93.9-98.6), respectively, while AUC was 0.81 (95%CI: 0.76-0.86).

CONCLUSION

DI-60 provides reliable WBC differential and platelet assessment. In doubtful cases, the use of light microscopy is still mandatory.

Performance of digital morphology analyzer Medica EasyCell assistant

OBJECTIVES

The EasyCell assistant (Medica, Bedford, MA, USA) is one of the state-of-the-art digital morphology analyzers. We explored the performance of EasyCell assistant in comparison with manual microscopic review and Pentra DX Nexus (Horiba ABX Diagnostics, Montpellier, France).

METHODS

In a total of 225 samples (100 normal and 125 abnormal samples), white blood cell (WBC) differentials and platelet (PLT) count estimation by EasyCell assistant were compared with the results by manual microscopic review and Pentra DX Nexus. The manual microscopic review was performed according to the Clinical and Laboratory Standards Institute guidelines (H20-A2).

RESULTS

WBC differentials between pre-classification by EasyCell assistant and manual counting showed moderate correlations for neutrophils (r=0.58), lymphocytes (r=0.69), and eosinophils (r=0.51) in all samples. After user verification, they showed mostly high to very high correlations for neutrophils (r=0.74), lymphocytes (r=0.78), eosinophils (r=0.88), and other cells (r=0.91). PLT count by EasyCell assistant highly correlated with that by Pentra DX Nexus (r=0.82).

CONCLUSIONS

The performance of EasyCell assistant for WBC differentials and PLT count seems to be acceptable even in abnormal samples with improvement after user verification. The EasyCell assistant, with its reliable performance on WBC differentials and PLT count, would help optimize the workflow of hematology laboratories with reduced workload of manual microscopic review.

Digital assessment of peripheral blood and bone marrow aspirate smears

The diagnosis of benign and neoplastic hematologic disorders relies on analysis of peripheral blood and bone marrow aspirate smears. As demonstrated by the widespread laboratory adoption of hematology analyzers for automated assessment of peripheral blood, digital analysis of these samples provides many significant benefits compared to relying solely on manual review. Nonetheless, analogous instruments for digital bone marrow aspirate smear assessment have yet to be clinically implemented. In this review, we first provide a historical overview detailing the implementation of hematology analyzers for digital peripheral blood assessment in the clinical laboratory, including the improvements in accuracy, scope, and throughput of current instruments over prior generations. We also describe recent research in digital peripheral blood assessment, particularly in the development of advanced machine learning models that may soon be incorporated into commercial instruments. Next, we provide an overview of recent research in digital assessment of bone marrow aspirate smears and how these approaches could soon lead to development and clinical adoption of instrumentation for automated bone marrow aspirate smear analysis. Finally, we describe the relative advantages and provide our vision for the future of digital assessment of peripheral blood and bone marrow aspirate smears, including what improvements we can soon expect in the hematology laboratory.

Efficient diagnosis of hematologic malignancies using bone marrow microscopic images: A method based on MultiPathGAN and MobileViTv2

BACKGROUND AND OBJECTIVES

Hematologic malignancies, including the associated multiple subtypes, are critically threatening to human health. The timely detection of malignancies is crucial for their effective treatment. In this regard, the examination of bone marrow smears constitutes a crucial step. Nonetheless, the conventional approach to cell identification and enumeration is laborious and time-intensive. Therefore, the present study aimed to develop a method for the efficient diagnosis of these malignancies directly from bone marrow microscopic images.

METHODS

A deep learning-based framework was developed to facilitate the diagnosis of common hematologic malignancies. First, a total of 2033 microscopic images of bone marrow analysis, including the images for 6 disease types and 1 healthy control, were collected from two Chinese medical websites. Next, the collected images were classified into the training, validation, and test datasets in the ratio of 7:1:2. Subsequently, a method of stain normalization to multi-domains (stain domain augmentation) based on the MultiPathGAN model was developed to equalize the stain styles and expand the image datasets. Afterward, a lightweight hybrid model named MobileViTv2, which integrates the strengths of both CNNs and ViTs, was developed for disease classification. The resulting model was trained and utilized to diagnose patients based on multiple microscopic images of their bone marrow smears, obtained from a cohort of 61 individuals.

RESULTS

MobileViTv2 exhibited an average accuracy of 94.28% when applied to the test set, with multiple myeloma, acute lymphocytic leukemia, and lymphoma revealed as the three diseases diagnosed with the highest accuracy values of 98%, 96%, and 96%, respectively. Regarding patient-level prediction, the average accuracy of MobileViTv2 was 96.72%. This model outperformed both CNN and ViT models in terms of accuracy, despite utilizing only 9.8 million parameters. When applied to two public datasets, MobileViTv2 exhibited accuracy values of 99.75% and 99.72%, respectively, and outperformed previous methods.

CONCLUSIONS

The proposed framework could be applied directly to bone marrow microscopic images with different stain styles to efficiently establish the diagnosis of common hematologic malignancies.

Retracted: White blood cell differentials performance of a new automated digital cell morphology analyzer: Mindray MC-80

White blood cell differentials performance of a new automated digital cell morphology analyzer: Mindray MC-80, K. Paisooksantivatana; N. Khongjaroensakun; P. Chinudomwong; N. Chaothai; L. Chamchomdao; K. Suriyachand, International Journal of Laboratory Hematology, 10.1111/ijlh.13995 The above article, published online on 18 November 2022, in Wiley Online Library (wileyonlinelibrary.com), had been retracted by agreement between the authors, the journal's Editors-in-Chief, Giuseppe D'Onofrio and Ian Mackie, and John Wiley & Sons. The authors contacted the journal after publication to propose extensive changes to the data presented in the accepted article such that it would no longer reflect the version that was peer reviewed. As a result, this retraction has been undertaken.

Performance of the CellaVision DC-1 digital cell imaging analyser for differential counting and morphological classification of blood cells

AIMS

Recently, a new automated digital cell imaging analyser (Sysmex CellaVision DC-1), intended for use in low-volume and small satellite laboratories, has become available. The purpose of this study was to compare the performance of the DC-1 with the Sysmex DI-60 system and the gold standard, manual microscopy.

METHODS

White blood cell (WBC) differential counts in 100 normal and 100 abnormal peripheral blood smears were compared between the DC-1, the DI-60 and manual microscopy to establish accuracy, within-run imprecision, clinical sensitivity and specificity. Moreover, the agreement between precharacterisation and postcharacterisation of red blood cell (RBC) morphological abnormalities was determined for the DC-1.

RESULTS

WBC preclassification and postclassification results of the DC-1 showed good correlation compared with DI-60 results and manual microscopy. In addition, the within-run SD of the DC-1 was below 1 for all five major WBC classes, indicating good reproducibility. Clinical sensitivity and specificity were, respectively, 96.7%/95.9% compared with the DI-60% and 96.6%/95.3% compared with manual microscopy. The overall agreement on RBC morphology between the precharacterisation and postcharacterisation results ranged from 49% (poikilocytosis) to 100% (hypochromasia, microcytosis and macrocytosis).

CONCLUSIONS

The DC-1 has proven to be an accurate digital cell imaging system for differential counting and morphological classification of WBCs and RBCs in peripheral blood smears. It is a compact and easily operated instrument that can offer low-volume and small satellite laboratories the possibilities of readily available blood cell analysis that can be stored and retrieved for consultation with remote locations.

Performance of digital morphology analyzer CellaVision DC-1

OBJECTIVES

CellaVision DC-1 (DC-1, Sysmex, Kobe, Japan) is a newly launched digital morphology analyzer that was developed mainly for small to medium-volume laboratories. We evaluated the precision, qualitative performance, comparison of cell counts between DC-1 and manual counting, and turnaround time (TAT) of DC-1.

METHODS

Using five peripheral blood smear (PBS) slides spanning normal white blood cell (WBC) range, precision and qualitative performance of DC-1 were evaluated according to the Clinical and Laboratory Standards Institute (CLSI) EP15-A3, EP15-Ed3-IG1, and EP12-A2 guidelines. Cell counts of DC-1 and manual counting were compared according to the CLSI EP 09C-ED3 guidelines, and TAT of DC-1 was also compared with TAT of manual counting.

RESULTS

DC-1 showed excellent precision (%CV, 0.0-3.5%), high specificity (98.9-100.0%), and high negative predictive value (98.4-100.0%) in 18 cell classes (12 WBC classes and six non-WBC classes). However, DC-1 showed 0% of positive predictive value in seven cell classes (metamyelocytes, myelocytes, promyelocytes, blasts, plasma cells, nucleated red blood cells, and unidentified). The largest absolute mean differences (%) of DC-1 vs. manual counting was 2.74. Total TAT (min:s) was comparable between DC-1 (8:55) and manual counting (8:55).

CONCLUSIONS

This is the first study that comprehensively evaluated the performance of DC-1 including its TAT. DC-1 has a reliable performance that can be used in small to medium-volume laboratories for assisting PBS review. However, DC-1 may make unnecessary workload for cell verification in some cell classes.

Multicentric evaluation of the variability of digital morphology performances also respect to the reference methods by optical microscopy

INTRODUCTION

Despite the important diagnostic role of peripheral blood morphology, cell classification is subjective. Automated image-processing systems (AIS) provide more accurate and objective morphological evaluation. The aims of this multicenter study were the evaluation of the intra and inter-laboratory variation between different AIS in cell pre-classification and after reclassification, compared with manual optical microscopy, the reference method.

METHODS

Six peripheral blood samples were included in this study, for each sample, 70 May-Grunwald and Giemsa stained PB smears were prepared from each specimen and 10 slides were delivered to the seven laboratories involved. Smears were processed by both optical microscopy (OM) and AIS. In addition, the assessment times of both methods were recorded.

RESULTS

Within-laboratory Reproducibility ranged between 4.76% and 153.78%; between-laboratory Precision ranged between 2.10% and 82.2%, while Total Imprecision ranged between 5.21% and 20.60%. The relative Bland Altman bias ranged between -0.01% and 20.60%. The mean of assessment times were 326 ± 110 s and 191 ± 68 s for AIS post reclassification and OM, respectively.

CONCLUSIONS

AIS can be helpful when the number of cell counted are low and can give advantages in terms of efficiency, objectivity and time saving in the morphological analysis of blood cells. They can also help in the interpretation of some morphological features and can serve as learning and investigation tools.

Red and white blood cell morphology characterization and hands-on time analysis by the digital cell imaging analyzer DI-60

Background

The Sysmex DI-60 digital morphology analyzer is a fully automated, cell-locating image analysis system. This study aimed to evaluate the analytical performance of DI-60.

Methods

A total of 822 peripheral blood smears were used. The diagnostic performance of DI-60 in terms of red blood cell (RBC) morphology characterization, white blood cell (WBC) differentials, and the total assay time including hands-on time was evaluated.

Results

In comparison with manual slide review, DI-60 demonstrated acceptable accuracy in recognizing polychromasia, target cells, and ovalocytes. However, for schistocytes, DI-60 demonstrated low specificity (10.4%) despite the high sensitivity (97.2%). In the precision analysis of RBC morphology characterization, borderline samples harboring specific RBCs showed inconsistencies in the positive results among 20 replicates. Particularly, 6 of 10 samples showed inconsistencies in the precision for schistocytes. For WBC differentials, the overall agreement between pre-classification results and user-verified results was 89.4%. Except for basophils, normal WBCs showed a good correlation between DI-60 (after user verification) and manual counts. The sensitivities in detecting immature granulocytes, blasts, atypical lymphocytes, and normoblasts were 85.9%, 92.0%, 37.5%, and 77.6%, respectively. Although the total assay time of DI-60 was longer than that of manual review, the hands-on time was considerably shorter with a difference of 144.1 s/slide for abnormal samples.

Conclusion

DI-60 demonstrated acceptable performance for normal samples. However, for abnormal WBC differentials and RBC morphology characterization, it should be utilized carefully. DI-60 may contribute to an improvement in laboratory efficiency with increased feasibility.

Differential diagnosis of hereditary anemias from a fraction of blood drop by digital holography and hierarchical machine learning

Anemia affects about the 25% of the global population and can provoke severe diseases, ranging from weakness and dizziness to pregnancy problems, arrhythmias and hearth failures. About 10% of the patients are affected by rare anemias of which 80% are hereditary. Early differential diagnosis of anemia enables prescribing patients a proper treatment and diet, which is effective to mitigate the associated symptoms. Nevertheless, the differential diagnosis of these conditions is often difficult due to shared and overlapping phenotypes. Indeed, the complete blood count and unaided peripheral blood smear observation cannot always provide a reliable differential diagnosis, so that biomedical assays and genetic tests are needed. These procedures are not error-free, require skilled personnel, and severely impact the financial resources of national health systems. Here we show a differential screening system for hereditary anemias that relies on holographic imaging and artificial intelligence. Label-free holographic imaging is aided by a hierarchical machine learning decider that works even in the presence of a very limited dataset but is enough accurate for discerning between different anemia classes with minimal morphological dissimilarities. It is worth to notice that only a few tens of cells from each patient are sufficient to obtain a correct diagnosis, with the advantage of significantly limiting the volume of blood drawn. This work paves the way to a wider use of home screening systems for point of care blood testing and telemedicine with lab-on-chip platforms.

How Reproducible Is the Data from Sysmex DI-60 in Leukopenic Samples?

Digital morphology (DM) analyzers are widely applied in clinical practice. It is necessary to evaluate performances of DM analyzers by focusing on leukopenic samples. We evaluated the analytical performance, including precision, of a Sysmex DI-60 system (Sysmex, Kobe, Japan) on white blood cell (WBC) differentials in leukopenic samples. In a total of 40 peripheral blood smears divided into four groups according to WBC count (normal, mild, moderate, and severe leukopenia; each group n = 10), we evaluated precision of WBC preclassificaiton by DI-60. %coefficients of variation (%CVs) of precision varied for each sample and for each cell class; the fewer cells per slide, the higher %CV. The overall specificity and efficiency were high for all cell classes except plasma cells (95.9-99.9% and 90.0-99.4%, respectively). The largest absolute value of mean difference between DI-60 and manual count in each group was: 10.77, normal; 10.22, mild leukopenia; 19.09, moderate leukopenia; 47.74, severe leukopenia. This is the first study that evaluated the analytical performance of DI-60 on WBC differentials in leukopenic samples as the main subject. DI-60 showed significantly different performance depending on WBC count. DM analyzers should be evaluated separately in leukopenic samples, even if the overall performance was acceptable.

Accurate white blood cell differential by Alinity hq: A comparison with flow cytometry and manual differential

INTRODUCTION

White blood cell (WBC) differential by flow cytometry can report a six-part WBC differential and enumerate blasts. Some modern hematology analyzers are also able to provide a six-part WBC differential (including immature granulocytes). Our goal was to compare the WBC differential obtained by the Abbott Alinity hq hematology analyzer to an 8-color single-tube flow cytometry method and to manual WBC differential.

METHODS

Samples from 144 patients were tested with Alinity hq, flow cytometry, and microscopic WBC differential. The WBC count ranged from 1.22 to 359 × 10⁹ /L, and 34 samples were flagged by the analyzer for abnormal WBC morphology.

RESULTS

Strong concordance was demonstrated between Alinity hq and flow cytometry for all six components of the differential, with correlation coefficients ranging from 0.86 (basophils) to 1.00 (lymphocytes). Small, clinically insignificant positive difference was observed between Alinity hq and flow cytometry for mature and total neutrophils (2.51% and 1.85%) and eosinophils (0.14%), and small negative difference for immature granulocytes (-0.65%), lymphocytes (-0.61%), and basophils (-0.21%). No bias was detected between the Alinity hq and flow cytometry monocyte counts. Alinity hq and flow cytometry results agreed with the manual differential, apart from small, clinically insignificant differences. Alinity hq nucleated red blood cell concentrations were equivalent with the manual results (r = 0.95, slope = 1.16). The percentage of blasts by flow cytometry demonstrated good correlation and agreement with the manual count (r = 0.99, slope = 1.35).

CONCLUSION

Alinity hq has produced accurate six-part WBC differential in this three-way comparison, equivalent to flow cytometry and morphological classification.

Highly accurate differentiation of bone marrow cell morphologies using deep neural networks on a large image data set

A data set of >170 000 microscopic images allows training neural networks for identification of BM cells with high accuracy.

Neural networks outperform a feature-based approach to BM cell classification and can be analyzed with explainability and feature embedding methods.

Biomedical applications of deep learning algorithms rely on large expert annotated data sets. The classification of bone marrow (BM) cell cytomorphology, an important cornerstone of hematological diagnosis, is still done manually thousands of times every day because of a lack of data sets and trained models. We applied convolutional neural networks (CNNs) to a large data set of 171 374 microscopic cytological images taken from BM smears from 945 patients diagnosed with a variety of hematological diseases. The data set is the largest expert-annotated pool of BM cytology images available in the literature. It allows us to train high-quality classifiers of leukocyte cytomorphology that identify a wide range of diagnostically relevant cell species with high precision and recall. Our CNNs outcompete previous feature-based approaches and provide a proof-of-concept for the classification problem of single BM cells. This study is a step toward automated evaluation of BM cell morphology using state-of-the-art image-classification algorithms. The underlying data set represents an educational resource, as well as a reference for future artificial intelligence–based approaches to BM cytomorphology.

Interchangeability of multiple Sysmex XN10 and XN20 modules for six types of leukocytes

INTRODUCTION

Differential counts of leukocytes are frequent, and often several automated blood cell counters are needed in contemporary laboratories. However, these modules are often individually quality assured. Our aim was therefore to validate the interchangeability of five hematology modules in a large modern laboratory and to compare them with our gold standard (GS) manual white blood cell differential count.

METHODS

At Copenhagen University Hospital, we compared five Sysmex XN-modules for neutrophils, lymphocytes, monocytes, eosinophils, basophils, and immature granulocytes (IG). We analyzed control samples in three levels to evaluate intra- and intermodular precision. Bias between modules was evaluated by analyzing 93 random patient samples within reference intervals. XN-modules' mean counts were compared with GS.

RESULTS

We found acceptable intramodular CV% (0.92%-8.76%), only neutrophils and eosinophils exceeded state-of-the-art imprecision or desirable specifications for medium control levels. Intermodular CV% showed significance difference for only monocytes (ANOVA, P < .0001). For patient samples, there were significant differences between XN-modules regarding four WBC types (ANOVA); however, proportional bias ranged from 1.7% to 3.8%, being within desirable specifications except basophils and IG (bias = 13.3% and 24.9%, respectively). Comparisons with GS, XN-modules exceeded desirable bias for basophils (lower than GS); monocytes and IG (higher than GS).

CONCLUSION

This multimodule comparison shows acceptable intermodular imprecision and bias for clinical purposes, which is important for patient safety. Similar multimodule study should be performed with samples out of reference range in large-scale laboratories to confirm the interchangeability.

Artificial Intelligence and Mapping a New Direction in Laboratory Medicine: A Review

BACKGROUND

Modern artificial intelligence (AI) and machine learning (ML) methods are now capable of completing tasks with performance characteristics that are comparable to those of expert human operators. As a result, many areas throughout healthcare are incorporating these technologies, including in vitro diagnostics and, more broadly, laboratory medicine. However, there are limited literature reviews of the landscape, likely future, and challenges of the application of AI/ML in laboratory medicine.

CONTENT

In this review, we begin with a brief introduction to AI and its subfield of ML. The ensuing sections describe ML systems that are currently in clinical laboratory practice or are being proposed for such use in recent literature, ML systems that use laboratory data outside the clinical laboratory, challenges to the adoption of ML, and future opportunities for ML in laboratory medicine.

SUMMARY

AI and ML have and will continue to influence the practice and scope of laboratory medicine dramatically. This has been made possible by advancements in modern computing and the widespread digitization of health information. These technologies are being rapidly developed and described, but in comparison, their implementation thus far has been modest. To spur the implementation of reliable and sophisticated ML-based technologies, we need to establish best practices further and improve our information system and communication infrastructure. The participation of the clinical laboratory community is essential to ensure that laboratory data are sufficiently available and incorporated conscientiously into robust, safe, and clinically effective ML-supported clinical diagnostics.

The Use and Effectiveness of an Online Diagnostic Support System for Blood Film Interpretation: Comparative Observational Study

Background

The recognition and interpretation of abnormal blood cell morphology is often the first step in diagnosing underlying serious systemic illness or leukemia. Supporting the staff who interpret blood film morphology is therefore essential for a safe laboratory service. This paper describes an open-access, web-based decision support tool, developed by the authors to support morphological diagnosis, arising from earlier studies identifying mechanisms of error in blood film reporting. The effectiveness of this intervention was assessed using the unique resource offered by the online digital morphology Continuing Professional Development scheme (DM scheme) offered by the UK National External Quality Assessment Service for Haematology, with more than 3000 registered users. This allowed the effectiveness of decision support to be tested within a defined user group, each of whom viewed and interpreted the morphology of identical digital blood films.

Objective

The primary objective of the study was to test the effectiveness of the decision support system in supporting users to identify and interpret abnormal morphological features. The secondary objective was to determine the pattern and frequency of use of the system for different case types, and to determine how users perceived the support in terms of their confidence in decision-making.

Methods

This was a comparative study of identical blood films evaluated either with or without decision support. Selected earlier cases from the DM scheme were rereleased as new cases but with decision support made available; this allowed a comparison of data sets for identical cases with or without decision support. To address the primary objectives, the study used quantitative evaluation and statistical comparisons of the identification and interpretation of morphological features between the two different case releases. To address the secondary objective, the use of decision support was assessed using web analytical tools, while a questionnaire was used to assess user perceptions of the system.

Results

Cases evaluated with the aid of decision support had significantly improved accuracy of identification for relevant morphological features (mean improvement 9.8%) and the interpretation of those features (mean improvement 11%). The improvement was particularly significant for cases with higher complexity or for rarer diagnoses. Analysis of website usage demonstrated a high frequency of access for web pages relevant to each case (mean 9298 for each case, range 2661-24,276). Users reported that the decision support website increased their confidence for feature identification (4.8/5) and interpretation (4.3/5), both within the context of training (4.6/5) and also in their wider laboratory practice (4.4/5).

Conclusions

The findings of this study demonstrate that directed online decision support for blood morphology evaluation improves accuracy and confidence in the context of educational evaluation of digital films, with effectiveness potentially extending to wider laboratory use.

A deep learning model (ALNet) for the diagnosis of acute leukaemia lineage using peripheral blood cell images

BACKGROUND AND OBJECTIVES

Morphological differentiation among blasts circulating in blood in acute leukaemia is challenging. Artificial intelligence decision support systems hold substantial promise as part of clinical practise in detecting haematological malignancy. This study aims to develop a deep learning-based system to predict the diagnosis of acute leukaemia using blood cell images.

METHODS

A set of 731 blood smears containing 16,450 single-cell images was analysed from 100 healthy controls, 191 patients with viral infections and 148 with acute leukaemia. Training and testing sets were arranged with 85% and 15% of these smears, respectively. To find the best architecture for acute leukaemia classification VGG16, ResNet101, DenseNet121 and SENet154 were evaluated. Fine-tuning was implemented to these pre-trained CNNs to adapt their layers to our data. Once the best architecture was chosen, a system with two modules working sequentially was configured (ALNet). The first module recognised abnormal promyelocytes among other mononuclear blood cell images, such as lymphocytes, monocytes, reactive lymphocytes and blasts. The second distinguished if blasts were myeloid or lymphoid lineage. The final strategy was to predict patients' initial diagnosis of acute leukaemia lineage using the blood smear review. ALNet was assessed with smears of the testing set.

RESULTS

ALNet provided the correct diagnostic prediction of all patients with promyelocytic and myeloid leukaemia. Sensitivity, specificity and precision values of 100%, 92.3% and 93.7%, respectively, were obtained for myeloid leukaemia. Regarding lymphoid leukaemia, a sensitivity of 89% and specificity and precision values of 100% were obtained.

CONCLUSIONS

ALNet is a predictive model designed with two serially connected convolutional networks. It is proposed to assist clinical pathologists in the diagnosis of acute leukaemia during the blood smear review. It has been proved to distinguish neoplastic (leukaemia) and non-neoplastic (infections) diseases, as well as recognise the leukaemia lineage.

Performance of digital morphology analyzer Vision Pro on white blood cell differentials

OBJECTIVES

Vision Pro (West Medica, Perchtoldsdorf, Austria) is a recently developed digital morphology analyzer. We evaluated the performance of Vision Pro on white blood cell (WBC) differentials.

METHODS

In a total of 200 peripheral blood smear samples (100 normal and 100 abnormal samples), WBC preclassification and reclassification by Vision Pro were evaluated and compared with manual WBC count, according to the Clinical and Laboratory Standards Institute guidelines (H20-A2).

RESULTS

The overall sensitivity was high for normal WBCs and nRBCs (80.1-98.0%). The overall specificity and overall efficiency were high for all cell classes (98.1-100.0% and 97.7-99.9%, respectively). The absolute values of mean differences between Vision Pro and manual count ranged from 0.01 to 1.31. In leukopenic samples, those values ranged from 0.09 to 2.01. For normal WBCs, Vision Pro preclassification and manual count showed moderate or high correlations (r=0.52-0.88) except for basophils (r=0.34); after reclassification, the correlation between Vision Pro and manual count was improved (r=0.36-0.90).

CONCLUSIONS

This is the first study that evaluated the performance of Vision Pro on WBC differentials. Vision Pro showed reliable analytical performance on WBC differentials with improvement after reclassification. Vision Pro could help improve laboratory workflow.

Automated cell differential count in sputum is feasible and comparable to manual cell count in identifying eosinophilia

INTRODUCTION

Cell differential count (CDC) of induced sputum is considered the gold standard for inflammatory phenotyping of asthma but is not implemented in routine care due to its heavy time- and staff demands. Digital Cell Morphology is a technique where digital images of cells are captured and presented preclassified as white blood cells (neutrophils, eosinophils, lymphocytes, macrophages, and unidentified) and nonwhite blood cells for review. With this study, we wanted to assess the accuracy of an automated CDC in identifying the key inflammatory cells in induced sputum.

METHODS

Sputum from 50 patients with asthma was collected and processed using the standard processing protocol with one drop 20% albumin added to hinder cell smudging. Each slide was counted automatically using the CellaVision DM96 and manually by an experienced lab technician. Sputum was classified as eosinophilic or neutrophilic using 3% and 61% cutoffs, respectively.

RESULTS

We found a good agreement using intraclass correlation for all target cells, despite significant differences in the cell count rate. The automated CDC had a sensitivity of 65%, a specificity of 93%, and a kappa-coefficient of 0.61 for identification of sputum eosinophilia. In contrast, the automated CDC had a sensitivity of 29%, a specificity of 100%, and a kappa-coefficient of 0.23 for identification of sputum neutrophilia.

CONCLUSION

Automated- and manual cell counts of sputum agree with regards to the key inflammatory cells. The automated cell count had a modest sensitivity but a high specificity for the identification of both neutrophil and eosinophil asthma.

Basophil counting in hematology analyzers: time to discontinue?

Basophils (basophilic granulocytes) are the least abundant cells in blood. Nowadays, basophils are included in the complete blood count performed by hematology analyzers and therefore reported in practically all patients in whom hematologic investigations are requested. However, hematology analyzers are not reliable enough to report clinically useful results. This is due to a combination of very high analytical imprecision and poor specificity, because the chemical and physical methods used for basophil counting in hematology analyzers are ill-defined and thus basophils are not well recognized by the analyzers. As a result, false basophil counts are quite common. In view of increasing analytical performance demands, hematology laboratories should stop reporting basophil counts produced by hematology analyzers. Suggestions for alternative pathways are presented for those situations where basophils are of clinical relevance.

Haematology and blood chemistry in free-ranging quokkas (Setonix brachyurus): Reference intervals and assessing the effects of site, sampling time, and infectious agents

Quokkas (Setonix brachyurus) are small macropodid marsupials from Western Australia, which are identified as of conservation concern. Studies on their blood analytes exist but involve small sample sizes and are associated with very little information concerning the health of the animals. Blood was collected from free-ranging quokkas from Rottnest Island (n = 113) and mainland (n = 37) Western Australia, between September 2010 and December 2011, to establish haematology and blood chemistry reference intervals. Differences in haematology and blood chemistry between sites (Rottnest Island v mainland) were significant for haematology (HMT, p = 0.003), blood chemistry (BLC, p = 0.001) and peripheral blood cell morphology (PBCM, p = 0.001). Except for alkaline phosphatase, all blood chemistry analytes were higher in mainland animals. There were also differences with time of year in HMT (p = 0.001), BLC (p = 0.001) and PBCM (p = 0.001) for Rottnest Island quokkas. A small sample of captive animals (n = 8) were opportunistically sampled for plasma concentrations of vitamin E and were found to be deficient compared with wild-caught animals. Fifty-eight of the 150 quokkas were also tested for the presence of Salmonella, microfilariae, Macropodid herpesvirus-6, Theileria spp., Babesia spp., trypanosomes, Cryptococcus spp. and other saprophytic fungi. All eight infectious agents were detected in this study. Infectious agents were detected in 24 of these 58 quokkas (41%), with more than one infectious agent detected for all 24 individuals. Salmonella were detected concurrently with microfilariae in 8 of these 24 quokkas, and this mixed infection was associated with lower values across all haematological analytes, with Salmonella having the greater involvement in the decreased haematological values (p < 0.05). There was no evidence for an effect of sex on HMT, BLC and PBCM. Our data provide important haematological and blood chemistry reference intervals for free-ranging quokkas. We applied novel methods of analyses to HMT and BLC that can be used more broadly, aiding identification of potential disease in wildlife.

Multicentric evaluation of analytical performances digital morphology with respect to the reference methods by manual optical microscopy

AIMS

Optical microscopic (OM) evaluation of peripheral blood (PB) cells is still a crucial step of the laboratory haematological workflow. The morphological cell analysis is time-consuming and expensive and it requires skilled operator. To address these challenges, automated image-processing systems, as digital morphology (DM), were developed in the last few years. The aim of this multicentre study, performed according to international guidelines, is to verify the analytical performance of DM compared with manual OM, the reference method.

METHODS

Four hundred and ninety PB samples were evaluated. For each sample, two May Grunwald-stained and Giemsa-stained smears were performed and the morphological evaluation of cells was analysed with both DM and OM. In addition, the assessment times of both methods were recorded.

RESULTS

Comparison of DM versus OM methods was assessed with Passing-Bablok and Deming fit regression analysis: slopes ranged between 0.17 for atypical, reactive lymphocytes and plasma cells (LY(AT)) and 1.24 for basophils, and the intercepts ranged between -0.09 for blasts and 0.40 for LY(AT). The Bland-Altman bias ranged between -6.5% for eosinophils and 21.8% for meta-myemielocytes. The diagnostic agreement between the two methods was 0.98. The mean of assessment times were 150 s and 250 s for DM and OM, respectively.

CONCLUSION

DM shows excellent performance. Approximately only 1.6% of PB smears need the OM revision, giving advantages in terms of efficiency, standardisation and assessment time of morphological analysis of the cells. The findings of this study may provide useful information regarding the use of DM to improve the haematological workflow.

Evaluation of the CellaVision Advanced RBC Application for Detecting Red Blood Cell Morphological Abnormalities

BACKGROUND

The Advanced RBC Application of the CellaVision DM9600 system (CellaVision AB, Lund, Sweden) automatically characterizes and classifies red blood cells (RBCs) into 21 morphological categories based on their size, color, shape, and inclusions. We evaluated the diagnostic performance of the CellaVision Advanced RBC Application with respect to the classification and grading of RBC morphological abnormalities in accordance with the 2015 International Council for Standardization in Haematology (ICSH) guidelines.

METHODS

A total of 223 samples, including 123 with RBC morphological abnormalities and 100 from healthy controls, were included. Seven RBC morphological abnormalities and their grading obtained with CellaVision DM9600 pre- and post-classification were compared with the results obtained using manual microscopic examination. The grading cut-off percentages were determined in accordance with the 2015 ICSH guidelines. The sensitivity and specificity of the CellaVision DM9600 system were evaluated using the manual microscopic examination results as a true positive.

RESULTS

In pre-classification, >90% sensitivity was observed for target cells, tear drop cells, and schistocytes, while >90% specificity was observed for acanthocytes, spherocytes, target cells, and tear drop cells. In post-classification, the detection sensitivity and specificity of most RBC morphological abnormalities increased, except for schistocytes (sensitivity) and acanthocytes (specificity). The grade agreement rates ranged from 35.9% (echinocytes) to 89.7% (spherocytes) in pre-classification and from 46.2% (echinocytes) to 90.1% (spherocytes) in post-classification. The agreement rate of samples with within-one grade difference exceeded 90% in most categories, except for schistocytes and echinocytes.

CONCLUSIONS

The Advanced RBC Application of CellaVision DM9600 is a valuable screening tool for detecting RBC morphological abnormalities.

Performance of the Sysmex White Precursor Channel to discover circulating leukemic blast cells

INTRODUCTION

Circulating immature precursor cells indicate malignant diseases like acute myeloid or lymphoid leukemia, and blast cells are key finds for disease management. Automatized cell counters are an essential contemporary appliance for blast detection, but false-positive samples remain challenging in terms of time and resources. To reduce this issue, the White Precursor Channel (WPC) was introduced to Sysmex XN series; however, sensitivity may reduce when accommodating low specificity. Therefore, our aim was to evaluate WPC blast alarm flag performance with regard to detecting blast cells.

METHODS

At two major Danish hospitals, random blood samples were collected from the routine setting   in a four-week period and analyzed on WPC XN20 (Sysmex, Japan). Results were compared with manual differential white blood cell count (Manual WBCC) assisted by CellaVisionDM96.

RESULTS

In 117 samples, we found 0.2 to 34.4% blasts, WPC blast flag specificity = 82% and a low sensitivity = 40%. However, other XN alarm flags forwarded samples to Manual WBCC, so blast cells were detected despite missing a specific blast flag: With all alarm flags, combined sensitivity increased to 88%. Overall, the WPC application stopped 18% of the 117 samples going to Manual WBCC (three false negatives). Q values are arbitrary probability measurements for the blast flag, and in five samples (0.5 to 47.3% blasts) imprecision ranged from 5.3 to 122 CV%.

CONCLUSIONS

WPC blast alarm flags are imprecise and inaccurate, especially when blast counts are low. However, the XN20 will alarm samples with other flags so that most samples containing blast cells will be manually reviewed after all. Hence, the presented flag types should not bias the decisions of manual reviewers.

Machine-based detection and classification for bone marrow aspirate differential counts: initial development focusing on nonneoplastic cells

Bone marrow aspirate (BMA) differential cell counts (DCCs) are critical for the classification of hematologic disorders. While manual counts are considered the gold standard, they are labor intensive, time consuming, and subject to bias. A reliable automated counter has yet to be developed, largely due to the inherent complexity of bone marrow specimens. Digital pathology imaging coupled with machine learning algorithms represents a highly promising emerging technology for this purpose. Yet, training datasets for BMA cellular constituents, critical for building and validating machine learning algorithms, are lacking. Herein, we report our experience creating and employing such datasets to develop a machine learning algorithm to detect and classify BMA cells. Utilizing a web-based system that we developed for annotating and managing digital pathology images, over 10,000 cells from scanned whole slide images of BMA smears were manually annotated, including all classes that comprise the standard clinical DCC. We implemented a two-stage, detection and classification approach that allows design flexibility and improved classification accuracy. In a sixfold cross-validation, our algorithms achieved high overall accuracy in detection (0.959 ± 0.008 precision-recall AUC) and classification (0.982 ± 0.03 ROC AUC) using nonneoplastic samples. Testing on a small set of acute myeloid leukemia and multiple myeloma samples demonstrated similar detection and classification performance. In summary, our algorithms showed promising early results and represent an important initial step in the effort to devise a reliable, objective method to automate DCCs. With further development to include formal clinical validation, such a system has the potential to assist in disease diagnosis and prognosis, and significantly impact clinical practice.

Blast flagging of the Sysmex XN‐10 hematology analyzer with supervised cell image analysis: Impact on quality parameters

Introduction

The Sysmex XN‐10 automated hematology analyzer (Sysmex Corporation) is routinely used in hematology laboratories to perform complete blood cell count with differential (CBC w/ diff). The sensitivity of this system for blast detection is unclear, since many prior studies evaluating the blast flagging capabilities of Sysmex XN series used the white precursor cell (WPC) channel, which is not cleared for use in the United States.

Methods

We assessed the blast flagging capabilities of the Sysmex XN‐10 compared with CellaVision (a cell image analyzer)‐assisted visual hematology results. We evaluated the following flags: “blasts?/abnormal lymph?” and “immature granulocytes present” and compared differences in turnaround time between methods.

Results

We collected data on 2239 CBC w/ diff Sysmex automated analyzer differential and CellaVision‐assisted visual differential from the inpatient hematology‐oncology population of a tertiary care medical center. Solely analyzing the first CBC/diff from each unique patient, both flags had a combined sensitivity of 100%, specificity of 50.2%, PPV of 21.7%, and NPV of 100%. The mean turnaround time for the automated differential was 19.5 minutes (SD 35.9 minutes) compared with 66.4 minutes for the CellaVision‐assisted visual differential (SD 68.5 minutes; P < 0.001; Figure 1).

Conclusion

The Sysmex XN‐10 abnormal lymphocyte/blast and immature granulocytes flags had excellent sensitivity and acceptable specificity in detecting circulating blasts with shorter turnaround time than the CellaVision‐assisted visual differential. Our study suggests that automated differentials performed on Sysmex XN‐10 can replace visual differentials as a first‐line screening method for blast detection with improved turnaround time in hematology‐oncology populations.

Automatic recognition of different types of acute leukaemia in peripheral blood by image analysis

AIMS

Morphological differentiation among different blast cell lineages is a difficult task and there is a lack of automated analysers able to recognise these abnormal cells. This study aims to develop a machine learning approach to predict the diagnosis of acute leukaemia using peripheral blood (PB) images.

METHODS

A set of 442 smears was analysed from 206 patients. It was split into a training set with 75% of these smears and a testing set with the remaining 25%. Colour clustering and mathematical morphology were used to segment cell images, which allowed the extraction of 2,867 geometric, colour and texture features. Several classification techniques were studied to obtain the most accurate classification method. Afterwards, the classifier was assessed with the images of the testing set. The final strategy was to predict the patient's diagnosis using the PB smear, and the final assessment was done with the cell images of the smears of the testing set.

RESULTS

The highest classification accuracy was achieved with the selection of 700 features with linear discriminant analysis. The overall classification accuracy for the six groups of cell types was 85.8%, while the overall classification accuracy for individual smears was 94% as compared with the true confirmed diagnosis.

CONCLUSIONS

The proposed method achieves a high diagnostic precision in the recognition of different types of blast cells among other mononuclear cells circulating in blood. It is the first encouraging step towards the idea of being a diagnostic support tool in the future.

Rise of the Machines: Advances in Deep Learning for Cancer Diagnosis

Deep learning refers to a set of computer models that have recently been used to make unprecedented progress in the way computers extract information from images. These algorithms have been applied to tasks in numerous medical specialties, most extensively radiology and pathology, and in some cases have attained performance comparable to human experts. Furthermore, it is possible that deep learning could be used to extract data from medical images that would not be apparent by human analysis and could be used to inform on molecular status, prognosis, or treatment sensitivity. In this review, we outline the current developments and state-of-the-art in applying deep learning for cancer diagnosis, and discuss the challenges in adapting the technology for widespread clinical deployment.