Automated imaging, identification, and counting of similar cells from digital hologram reconstructions
Spotlight summary: Imaging based assays (especially in X-ray and optical regimes) have significant potential of providing automated and accurate assessment of biological samples for clinical diagnosis. In both regimes, new approaches are emerging that do not require intrusive labeling of the specimen but rather utilize the phase delay induced by transparent specimens. In the X-ray regime (where useful labels are scarce), emergence of coherent and partially coherent phase imaging methods is considered ‘the most important development after computed tomography.’ In the optical regime (where labeling is relatively more frequent and easier), researchers are developing new phase imaging based assays that can improve speed and fidelity of clinical assessment. The current paper is a case in point. It develops and verifies a holographic approach of distinguishing and quantifying mature and immature red blood cells (RBCs) from a patient sample.
Mature RBCs (which have biconcave shape and do not contain DNA or RNA) are carriers of oxygen, whereas immature RBCs lack this capacity. In the human circulatory system, a very small amount (about 1% of red cells) are immature blood cells. Immature RBCs are not as symmetrically biconcave as mature RBCs and contain ribosomal RNA. It is clinically known that higher abundance of immature red blood cells may be caused by acute blood loss, hypoxia, RBC destruction, sickle cell disease, and autoimmune hemolytic anemia. Measurement of abundance of immature RBCs is diagnostically important and has been achieved mostly by labeling the RNA present in immature RBCs. Rather than relying on presence of RNA, the authors use the other distinguishing property of immature RBCs: their morphology.
Mihailescu et al. show that the diffraction patterns produced by mature RBCs and immature RBCs carry a signature because of the difference in their concavity. This observation is supported by experimental data as well as simulation. They assume the RBCs to be ellipsoids (immature RBCs are modeled as simple oblate spheroids and mature RBCs as biconcave oblate spheroids) and simulate the diffraction pattern produced by them in the Fresnel region. The simulations results match with diffraction patterns observed from RBCs in experiments—where RNA labeling was used to identify the maturity of the RBCs. The authors proceed to develop algorithms for automatic identification and counting of mature RBCs and immature RBCs.
But simple image analysis of holographically recorded diffraction patterns does not suffice owing to experimental noise and the fact that cells may overlap each other. Overlap of cells is a significant problem for holography as it is not a true 3D imaging method; holography provides poor discrimination of two RBCs that may overlap in the direction of propagation. Mihailescu et al. develop and verify detailed algorithms for identifying overlapping RBCs and distinguishing them.
Put together, the imaging and image analysis approach developed by these authors may provide a promising method of measuring the abundance of immature RBCs.
The most challenging aspect of the presented approach (viz., the need for detailed image analysis to distinguish overlapping cells) may be overcome by using phase microscopy methods that use partially coherent rather than coherent light. Partially coherent phase microscopy approaches offer intrinsic sectioning and hence provide another exciting set of methods that could be useful for measuring the abundance of immature RBCs.
Technical Division: Optical Design and Instrumentation
ToC Category: Holography
|OCIS Codes:||(090.0090) Holography : Holography|
|(090.1970) Holography : Diffractive optics|
|(170.0180) Medical optics and biotechnology : Microscopy|
|(090.1995) Holography : Digital holography|
You must log in to add comments.