Abstract:Abstract Background Performing basic hematology analysis in a near-patient setting enables more rapid clinical decision making in the process of diagnosis. Traditional blood analyzers in central labs use a complex combination of sensing principles and chemical modification of the sample to derive the necessary parameters for a complete blood count (CBC). To reduce cost and maintenance, varying approaches have been employed to determine a CBC with a single-step measurement technique. One common option is the use of optical microscopy to deliver highly reliable results, examples are fluorescence microscopy, digital holographic microscopy, or spatial light interference microscopy. However, most of these technologies lack a proper determination of cellular-level RBC parameters. Therefore, a non-interferometric, quantitative-phase measurement technique was developed to derive relevant CBC parameters from a single measurement, comparable to a standard lab analyzer. Methods The optical setup employed by this method enables the parallel acquisition of brightfield and differential interference contrast (DIC) images by polarization-sensitive cameras. By using three cameras with individual optical filters, quantitative phase and absorption information can be calculated, which enables the determination of mean cellular volume (MCV), mean cellular hemoglobin (MCH), and mean cellular hemoglobin concentration (MCHC) for red blood cells (RBCs). The whole-blood sample is diluted in a buffer containing a non-cell-penetrating dye to enhance optical refraction contrast and therefore precision in the dependent calculated quantities. The sample is then introduced to a glass flow cell, and the cells are allowed to sediment to the bottom for imaging. Machine learning is applied to the derived images to classify unstained native white blood cells (WBCs) to yield a 5-part differential, which distinguishes between neutrophil, monocyte, lymphocyte, eosinophil and basophil. The algorithm is trained by imaging FACS-validated, purified WBC subgroups. All blood samples are measured on an ADVIA® 2120 Hematology System for control. Results A one-step, optical CBC hematology system based on quantitative phase and absorption image analysis was developed. The correlations for MCV and MCHC compared to a standard lab CBC analyzer were 0.77 and 0.84. The accuracy for the 5-part differential WBC classification was approximately 0.8. To enable a visual inspection of the classification results, a pseudo-colored image of each cell was generated and uploaded in a review tool. Conclusion With a more-refined, less-expensive optical setup, a high-quality CBC analyzer suitable for use at the point of care can be derived with the presented method. A possible simplification could be the use of low-cost but high-resolution Fourier ptychography microscopy (FPM) technology. Imaging of blood samples on a single-cell level offers the potential for the analysis of even more morphological parameters, for example, disease-related changes in the cell membrane. Also, the adhesion behavior of the WBCs in the glass flow cell could indicate cell activation status and therefore give information about infection processes. *Trademark disclaimer: “ADVIA, and all associated marks are trademarks of Siemens Healthcare Diagnostics, Inc. or its affiliates. All other trademarks and brands are the property of their respective owners.”

3-D image analysis system and megakaryocyte quantitation

A Segmentation Method Based on HMRF for the Aided Diagnosis of Acute Myeloid Leukemia

Computer Aided Disector Analysis and 3D Reconstruct System

Quantitative analysis of microvascular network

Label Free Monitoring of Megakaryocytic Development and Proplatelet Formation in Vitro

An Automatic Method for Robust and Fast Cell Detection in Bright Field Images from High-Throughput Microscopy.

A-075 Development and Validation of Image-based Cell Counter Microscanner C3 for CD34-positive Hematopoietic Stem Cell Enumeration

A Neurocomputational Method for Fully Automated 3D Dendritic Spine Detection and Segmentation of Medium-Sized Spiny Neurons

Automatic retinal image diagnosis system for mass health screenings

Visualization and Correction of Automated Segmentation, Tracking and Lineaging from 5-D Stem Cell Image Sequences

Alignment, Segmentation and 3-D Reconstruction of Serial Sections Based on Automated Algorithm

Bone Marrow Cells Detection: A Technique for the Microscopic Image Analysis

Automatic Correspondence Finding In Deformed Serial Sections

ROI-BMC-DNNet: An efficient automatic analysis model of whole-slide scanned bone marrow aspirate images for the diagnosis of hematological disorders

A-144 Three-wavelength Polarization Imaging Method for the Analysis of Individual Human Blood Cells With Dye-Enhanced Dispersion of Dilution Buffer

Quantifying Cellular Interaction Dynamics in 3d Fluorescence Microscopy Data

A new method for automatic counting of marrow cells.

Development of an image processing software for quantification of histological calcification staining images

PACESS: Practical AI-based Cell Extraction and Spatial Statistics for large 3D bone marrow tissue images

Automated 3D scoring of fluorescence in situ hybridization (FISH) using a confocal whole slide imaging scanner

Performance Evaluation of a Novel Artificial Intelligence-Assisted Digital Microscopy System for the Routine Analysis of Bone Marrow Aspirates