Combining imaging flow cytometry and machine learning for high-throughput schistocyte quantification: A SVM classifier development and external validation cohort

Demagny, Julien; Roussel, Camille; Guyader, Maïlys Le; Guiheneuf, Eric; Harrivel, Véronique; Boyer, Thomas G.; Dussiot, Michaël; Demont, Yohann; Garçon, Loïc

doi:10.1016/j.ebiom.2022.104209

Cited by 10 publications

(10 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This method has very high specificity (91.3%) and sensitivity (100%) for schizocytosis (schistocytes >1%). 26 The investigators designed and implemented several clinical cohort studies. The criteria for RBC-DIFF counts was used to distinguish TTP and hemolytic uremic syndrome from other thrombotic microangiopathies and a higher specificity than the clinical morphology grading (72%) was provided while maintaining a high sensitivity (94%-100%).…”

Section: Other Research Progress Of Ai Related To Schistocyte Detectionmentioning

confidence: 99%

“…Keras features recognized whole‐blood elements with higher accuracy (94.03%, CI: 93.75%–94.31%) than IDEAS features (91.54%, CI: 91.2191 0.87%), and the combination of the two had the highest accuracy (95.64%, CI: 95.3995.88%). This method has very high specificity (91.3%) and sensitivity (100%) for schizocytosis (schistocytes >1%) 26 . Foy et al devised a new machine‐learning algorithm (RBC‐diff algorithm) for quantifying RBC morphologies in PB smear images.…”

Section: Ai and Morphological Detection Of Schistocytesmentioning

confidence: 99%

See 1 more Smart Citation

Schistocyte detection in artificial intelligence age

Zhang,

Yang,

Wang

2024

Int J Lab Hematology

View full text Add to dashboard Cite

Schistocytes are fragmented red blood cells produced as a result of mechanical damage to erythrocytes, usually due to microangiopathic thrombotic diseases or mechanical factors. The early laboratory detection of schistocytes has a critical impact on the timely diagnosis, effective treatment, and positive prognosis of diseases such as thrombocytopenic purpura and hemolytic uremic syndrome. Due to the rapid development of science and technology, laboratory hematology has also advanced. The accuracy and efficiency of tests performed by fully automated hematology analyzers and fully automated morphology analyzers have been considerably improved. In recent years, substantial improvements in computing power and machine learning (ML) algorithm development have dramatically extended the limits of the potential of autonomous machines. The rapid development of machine learning and artificial intelligence (AI) has led to the iteration and upgrade of automated detection of schistocytes. However, along with significantly facilitated operation processes, AI has brought challenges. This review summarizes the progress in laboratory schistocyte detection, the relationship between schistocytes and clinical diseases, and the progress of AI in the detection of schistocytes. In addition, current challenges and possible solutions are discussed, as well as the great potential of AI techniques for schistocyte testing in peripheral blood.

show abstract

Section: Other Research Progress Of Ai Related To Schistocyte Detectionmentioning

confidence: 99%

Section: Ai and Morphological Detection Of Schistocytesmentioning

confidence: 99%

Schistocyte detection in artificial intelligence age

Zhang,

Yang,

Wang

2024

Int J Lab Hematology

View full text Add to dashboard Cite

show abstract

“…However, most of the past works rely on heavy computation models for both classification and localization 26,32,35 Additionally, there has been no mention of poorly focused datasets and attempts to solve the issue of defocus for IFC data. Additionally, past works typically do not optimize different pre-training modes and model configurations 8,22,35,40 .…”

Section: Related Workmentioning

confidence: 99%

Advancing precision single-cell analysis of red blood cells through semi-supervised deep learning using database of patients with post-COVID-19 syndrome

Kurenkov,

Kussanova,

Barteneva

2024

Imaging, Manipulation, and Analysis of Biomolecules, Cells, and Tissues XXII

View full text Add to dashboard Cite

We developed a semi-supervised deep learning-based system classifying different types of red blood cells (RBCs) images based on their shape, texture, and size. Specifically, pre-training a convolutional neural network was done on over 35,000 brightfield images of RBCs acquired with an imaging flow cytometer from a post-COVID-19 patient cohort. The system utilizes object localization powered by a YOLO-inspired block for cell identification and a de-blurring CNN block based on FocalNet. A series of convolutional and fully connected layers classifies images into side-view, biconcave, elongated, and additional categories for reticulocytes and erythrocytes. Fine-tuning was done using 7,000 manually labeled brightfield images. Consequent evaluation on a test dataset of 3,000 samples yielded an accuracy of 98.2%. This system can be used for other cell analysis tasks, not requiring large fine-tuning datasets while maintaining high efficiency.

show abstract

“…Previous work has applied ML to prediction problems in forensic science, including recent applications predicting tissue age and parameter selection for fluorescent molecular topography (10,11). ML has also been applied to IFC measurements to classify white blood cell types (12) as well as red blood cell types (13), to differentiate cancer cells from blood cells (14), and to predict gene expression from blood cells (15). To our knowledge, ML has never been applied to estimate TSD in epithelial cells from touch samples using IFC measurements.…”

Section: Introductionmentioning

confidence: 99%

Comparison of three quantitative approaches for estimating time-since-deposition from autofluorescence and morphological profiles of cell populations from forensic biological samples

Gentry

Ingram

Philpott

et al. 2023

Preprint

View full text Add to dashboard Cite

Determining when DNA recovered from a crime scene transferred from its biological source, i.e., a sample’s ‘time-since-deposition’ (TSD), can provide critical context for biological evidence. Yet, there remains no analytical techniques for TSD that are validated for forensic casework. In this study, we investigate whether morphological and autofluorescence measurements of forensically-relevant cell populations generated with Imaging Flow Cytometry (IFC) can be used to predict the TSD of ‘touch’ or trace biological samples. To this end, three different prediction frameworks for estimating the number of day(s) for TSD were evaluated: the elastic net, gradient boosting machines (GBM), and generalized linear mixed model (GLMM) LASSO. Additionally, we transformed these continuous predictions into a series of binary classifiers to evaluate the potential utility for forensic casework. Results showed that GBM and GLMM-LASSO showed the highest accuracy, with mean absolute error estimates in a hold-out test set of 29 and 21 days, respectively. Binary classifiers for these models correctly binned 94-96% and 98-99% of the age estimates as over/under 7 or 180 days, respectively. This suggests that predicted TSD using IFC measurements coupled to one or, possibly, a combination binary classification decision rules, may provide probative information for trace biological samples encountered during forensic casework.

show abstract

Combining imaging flow cytometry and machine learning for high-throughput schistocyte quantification: A SVM classifier development and external validation cohort

Cited by 10 publications

References 53 publications

Schistocyte detection in artificial intelligence age

Schistocyte detection in artificial intelligence age

Advancing precision single-cell analysis of red blood cells through semi-supervised deep learning using database of patients with post-COVID-19 syndrome

Comparison of three quantitative approaches for estimating time-since-deposition from autofluorescence and morphological profiles of cell populations from forensic biological samples

Contact Info

Product

Resources

About