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.

OBJECTIVE: Widely accepted standardized criteria for peripheral blood (PB) smear review do not exist. The aim of this study was to collect data regarding PB smear review practices across multiple institutions, with a focus on pathologist review.
METHODS: A 23-question survey was developed by members of the Society for Hematopathology (SH) Education Committee and distributed to SH members. The survey included questions on practice environment and PB smear review practices, including trainee involvement.
RESULTS: Of 725 members contacted, 137 (19%) completed the entire survey. Over half of practices examined 5 to 20 smears a day. All respondents reported using complete blood count/differential leukocyte count data and clinical history as part of smear review. The reported proportion of laboratory-initiated vs clinician-requested reviews varied across respondents. Clinician-requested smear reviews were more likely to be billed and issued as a separate pathology report. Glass slide review (as opposed to digital microscopy) was used by most respondents. All respondents affirmed that PB smear review is an essential component of pathology training programs. Numerous free-text comments were submitted by respondents regarding their own experiences with PB smear review and suggested improvements.
CONCLUSIONS: This survey elucidated the spectrum of practice patterns for pathologist review of blood smears and identified potential areas for process improvement.

暂无摘要。

BACKGROUND: Although widely used, the ADVIA 120 hematology analyzer has not been previously validated for determining the differential leukocyte count in goats.
OBJECTIVE: The aim of this study was to compare the differential leukocyte counts provided by the ADVIA 120 (A-diff) and the manual method (M-Diff) in goats.
METHODS: EDTA blood samples that were analyzed within 4 h of collection were used in the study. The following exclusion criteria were applied: inappropriately filled tubes or tubes containing clots, erroneous ADVIA peroxidase cytograms, and blood smears of poor quality. The A-Diff was compared with the M-Diff performed by two independent observers on 200 leukocytes.
RESULTS: Forty samples were included after previously excluding eight samples. The correlation between the A-Diff and M-Diff was very strong for eosinophils (r = .870, p < .001) and strong for lymphocytes (r = .796, p < .001) and neutrophils (r = .730, p < .001), while no significant correlation was observed for monocytes (r = .026, p = .872). The Passing-Bablok regression analyses revealed statistically significant constant errors for neutrophils (5.83%; 95% confidence interval [CI]: 0.41%, 12.18%) and eosinophils (1.89%; 95% CI: 1.17%, 2.71%). Bland-Altman analyses showed a statistically significant negative bias for lymphocytes (-5.0%) and a statistically significant positive bias for eosinophils (2.2%). The very low basophil percentages precluded a meaningful method comparison.
CONCLUSIONS: The ADVIA 120 overall demonstrated good performance for the differential WBC count in goats under the conditions of this study. Therefore, it can be considered suitable for routine hematologic screening in goats. Nonetheless, it should be emphasized that any abnormal result should be confirmed with a blood smear evaluation.

Delays in the diagnosis and management of sepsis are associated with higher mortality. Moreover, routine blood tests performed just before hospital discharge are still insufficient for sepsis survivors. In this report, for the first time, dramatic hematological changes found in the blood of a sepsis survivor are described. The pictorial information from microscope images associated with an appropriate set of multiparameter laboratory test results enabled for prediction of sepsis relapse four days before its clinical recognition. Thus, the role of this case report is to encourage medical practitioners to introduce (or re-introduce) blood smears as the helpful adjunctive extension of routine blood testing, especially when sepsis is suspected.

For many years, the treatment of malaria was based on clinical presumptive diagnosis, making its differential diagnosis with other causes of hyperthermia difficult. This drug pressure has led to the emergence of Plasmodium strains resistant to the most commonly used antimalarial drugs. This is why in 2004, the health authorities decided to revise the policy of malaria management by adopting a new strategy based on the rational use of artemisininbased combination therapies after the biological confirmation of suspected malaria cases. The biological diagnosis is an essential part of malaria management. The gold standard technique for diagnosis is the thick drop combined with the calculation of parasite density (PD), which is determined on the basis of the number of parasites counted in a microscopic field against a proposed standard number of leukocytes. The number of leukocytes used to calculate the parasite density should ideally be the actual number of leukocytes in the patient per cubic millimetre of blood. However, in the absence of the availability of a blood count at the time of the thick drop, an average number of 8 000 leukocytes/mm3 was used by the World Health Organisation (WHO) to estimate the parasite density. Nonetheless, in Benin the average number of leukocytes adopted by the National Malaria Control Programme (PNLP) is 6 000/mm3. The aim of our study was to determine the impact of the leukocyte count on the calculation of the parasite density in cases of uncomplicated malaria.
The study was a cross-sectional study with an analytical aim and took place in 2 hospitals in Benin, the Klouékanmey zone hospital in the south of Benin and the Djougou health centre in the north. It involved a population of 476 children aged between 6 and 59 months who were seen in consultation and in whom the clinical diagnosis of simple Plasmodium falciparum malaria was suspected. Children aged between 6 and 59 months, weighing at least 5 kg, with an axillary temperature ≥ 37.5°C at the time of consultation or a history of fever in the last 24 hours or other symptoms pointing to the diagnosis of malaria were included. Infestation was mono-specific for Plasmodium falciparum. Informed consent was required from the child\'s parents or guardian. The criteria for non-inclusion in our study were the presence of at least one sign of malaria severity, signs of severe malnutrition or a febrile state related to underlying infectious diseases other than malaria. Thick blood count and haemogram were systematically performed in all included children. Parasite density was calculated according to 3 methods, first using a weighted leukocyte count of 6 000/mm3 recommended by the Benin National Malaria Control Programme (PNLP), then a leukocyte count of 8 000/mm3 recommended by the World Health Organisation and finally the patient\'s actual leukocyte count obtained from the blood count. It should be noted that these different samples were respectively taken on the day of inclusion in compliance with the conditions of the pre-analytical phase in force in our medical biology laboratory.
At the end of our study, 313 children, i.e. 65.76% of our study population had a positive white blood cell count with a positivity rate of 62.14% in Djougou, i.e. 174 children, and 70.9% in Klouékanmey, i.e. 139 children. The average leukocyte count in these children was 11,580/mm3. Among them, 205 children had an abnormal white blood cell count, i.e. 17 cases of leukopenia (5.43%) and 188 cases of hyperleukocytosis (60.06%). Using successively the average number of 6 000 leukocytes/mm3 proposed by the Benin PNLP and that of 8 000 leukocytes/mm3 proposed by the WHO, the average parasite densities were respectively 47,943 and 63,936 trophozoïtes/µl against 92,290 trophozoïtes/µl when the real number of leukocytes of the patients was used for the calculation of the PD. By using an average of 6 000 leukocytes/mm3 for PD calculation, 60% of the calculated PDs were underestimated and 6% were overestimated. Using an average of 8 000 leukocytes/mm3 resulted in 49% of PD being underestimated and 15% being overestimated. The difference between the three calculation methods was considered statistically significant (p value <0.05).
The use of 6 000 or 8 000 coefficients for the estimation of parasitaemia could lead to a significant underestimation of the parasite load.

BACKGROUND: Cellular deterioration occurs with blood sample aging, impacting white blood cell (WBC) identification and differential accuracy. This may be exacerbated in samples from patients experiencing inflammation. Previously, bovine serum albumin (BSA) has been shown to improve cellular preservation of blood and other samples, but the effect on cell preservation in canine blood has not been assessed.
OBJECTIVE: We aimed to determine the effects of BSA on neutrophil nuclear area when added to potassium ethylene diamine tetra-acetic acid (K3 -EDTA)-anticoagulated canine blood prior to blood smear preparation. We evaluated the impact of inflammatory leukograms, sample storage temperatures (4° and 20°C), and time on outcomes.
METHODS: Canine K3 -EDTA-anticoagulated blood samples stored at 4° and 20°C were used from unique patients, 10 with and 10 without inflammatory leukograms. Blood smears were prepared from aliquots with or without the addition of 22% BSA at 0, 4, 8, 24, 48, and 72 h. The nuclear area was measured for 25 randomly selected neutrophils per slide using Fiji software. Mixed-effect linear regression modeling was performed (significance: P < 0.05).
RESULTS: Nuclear area increased over time with and without added BSA. Both sample storage temperatures and the presence or absence of an inflammatory leukogram significantly impacted neutrophil nuclear area. Samples with added BSA had slightly higher predicted nuclear areas than those without BSA, but this difference was not statistically significant.
CONCLUSIONS: BSA did not significantly impact neutrophil nuclear area and did not improve neutrophil preservation in canine blood samples.

We evaluated and compared the peripheral blood findings in patients with acute COVID-19 vs other viral respiratory infections.
We retrospectively reviewed peripheral blood counts and smear morphology in patients with a positive viral respiratory panel (VRP) or SARS-CoV-2 test.
A total of 97 peripheral blood samples (COVID-19 infection, 53; VRP positive, 44) from 50 patients (mean [SD] age, 45.8 [20.8] years; females 52%) were reviewed. There were no statistically significant differences in the demographic characteristics between the 2 groups. The most common peripheral blood abnormalities were anemia, thrombocytopenia, absolute lymphopenia, and reactive lymphocytes. The following peripheral blood findings were significantly associated with other viral respiratory infections compared with COVID-19 infection: low red blood cell count, low hematocrit, high mean corpuscular volume, thrombocytopenia, low mean platelet volume, high red cell distribution width, band neutrophilia, and toxic granulation in neutrophils.
Our study showed that there are several peripheral blood count and morphologic abnormalities seen in patients with COVID-19, but most of these findings lack specificity as they are also seen in the other viral respiratory infections.

This study compares the effectiveness of an interactive e-learning module with a traditional text-based method for teaching peripheral blood smear analysis.
Pathology trainees at Accreditation Council for Graduate Medical Education residency programs were asked to participate. Participants completed a multiple-choice test on peripheral blood smear findings. Trainees were randomized into completing an e-learning module or a PDF reading exercise with the same educational content. Respondents rated their experience and completed a postintervention test composed of the same questions.
In total, 28 participants completed the study; 21 improved their score in the posttest (mean, 21.6 correct answers) compared with the pretest (19.8; P < .001). This improvement was seen in both the PDF (n = 19) and interactive (n = 9) groups, with no difference in performance between the 2 groups. Trainees with less clinical hematopathology experience showed a trend of having the largest performance improvement. Most participants completed the exercise within 1 hour, rated the exercise as easy to navigate, were engaged, and reported learning new information about peripheral blood smear analysis. All participants indicated that they would likely complete a similar exercise in the future.
This study suggests that e-learning is an effective tool for hematopathology education and equivalent to traditional narrative-based methods. This module could easily be incorporated into a curriculum.

BACKGROUND: A peripheral blood smear is a basic test for hematological disease diagnosis. This test is performed manually in many places worldwide, which requires both time and qualified staff. Large laboratories are equipped with digital morphology analyzers, some of which are based on deep learning methods. However, it is difficult to explain to scientists how they work. In this paper, we proposed to add an explanatory factor to enhance the interpretability of deep learning models in leukocyte classification.
METHODS: 10 297 single images of leukocytes obtained from peripheral blood smears were included in this study. Pre-trained and fully trained VGG16 and VGG19 models were used to classify the leukocytes, and Shapley Additive Explanations (SHAP) DeepExplainer was applied to visualize the area of cells that were significant for classification. The output images from the DeepExplainer were compared with cellular elements that are essential to laboratory practice.
RESULTS: The accuracy of our fully trained models was 99.81% for VGG16 and 99.79% for VGG19. It achieved slightly better results than the partially trained model, which scored 98.67% for VGG16 and 98.33% for VGG19. Their SHAP explanations indicated the significance of cellular structures in microscopic examination. Explanations in the pre-trained models have proved the cell and nucleus contours to be relevant to classification, while explanations in the fully trained models pointed to the cytoplasm area.
CONCLUSIONS: Despite different SHAP DeepExplainer explanations for fully and partially trained models, this method appears to be helpful for the verification of leukocyte classification in automated peripheral blood smear examination.