Multiwavelength transmission (MWT) spectroscopy was applied to the investigation of the morphological parameters and composition of red blood cells (RBCs). The MWT spectra were quantitatively analyzed with a Mie theory based interpretation model modified to incorporate the effects of the nonsphericity and orientation of RBCs. The MWT spectra of the healthy and anemic samples were investigated for the RBC indices in open and blinded studies. When MWT performance was evaluated against a standard reference system, very good agreement between two methods, with
Numerous diseases result in pathomorphological changes in the red blood cell (RBC) shape, structure, composition, and count [
Multiwavelength transmission (MWT) spectroscopy provides substantial information on the physical, chemical, and physiological character of the cells and therefore it is capable of detection and identification of changes in the cells due to diseases [
The absorption component provides information on the chemical composition and the scattering component on the size and structure of the cells. Since scattering properties depend on the refractive index of the measured objects, which, in turn, is governed by chemical composition, simultaneous measurement of both absorption and scattering gives clear advantage over the methods of traditional spectroscopy such as angular light scattering [
In this paper, we review the method based on MWT spectroscopy coupled with MMT interpretation model for the estimation of the RBC parameters (size, dimensions, and hemoglobin composition) and RBC indices in whole blood (RBC number density, hematocrit, and total hemoglobin). The whole blood samples from 43 healthy donors and from 51 oncology patients undergoing chemotherapy and having cancer related and/or chemotherapy-induced anemia were used for the demonstration. The samples from anemic patients were analyzed blindly. The samples were analyzed with a reference laboratory hematology analyzer (SYSMEX SF-3000) in parallel to the MWT method. In addition, the MWT method was applied for the analysis of laboratory culture samples of RBCs infected with malaria parasite
Whole blood samples from 43 healthy donors were provided by One Blood (St. Petersburg, FL). The samples were collected using lavender top tubes containing K2EDTA and analyzed for the complete blood count (CBC) parameters within 24 hours of collection with a laboratory reference system (SYSMEX SF-3000 hematology analyzer). The MWT spectroscopy analysis of the blood samples from healthy donors was conducted in parallel.
Fifty-one EDTA tubes with whole blood from anemic patients were provided by One Blood (St. Petersburg, FL). The samples were collected from the oncology patients at Moffit Cancer Center (Tampa, FL) 1–5 days prior to the analysis. The analyses with MWT and the reference system (SYSMEX SF-3000 hematology analyzer) were conducted in parallel and the results from the reference system were sealed until the end of the study.
In vitro cultures of the W2 strain of
All MWT spectra were recorded using a diode array spectrometer (HP 8453 Hewlett-Packard, Palo Alto, CA) having an acceptance angle smaller than 2°. The measurements were conducted with spectra acquisition time of 0.5 sec, signal to noise ratio greater than 104, and 1 nm wavelength resolution. All measurements were conducted at room temperature using a 1 cm pathlength cuvette. Prior to recording a sample spectrum, the spectrometer was zeroed to account for any stray light. To avoid the effect of inhomogeneities in the suspending medium, the background spectrum was taken using a sample from the same batch of phosphate buffered saline (PBS) utilized for the sample measurements.
3
1 mL of a whole blood sample was centrifuged at 13,000 rpm for 1 min, plasma and buffy coat fractions were aspired, and RBCs were resuspended in 1 mL of PBS. The centrifugation, supernatant aspiration, and resuspension in PBS were repeated 3 times to completely clear the RBC fraction from other blood components. To measure a MWT spectrum, 3–5
1 mL of 4% hematocrit culture containing
The measured spectra of RBCs from healthy donors, anemic patients, and noninfected control RBCs were interpreted using the homogeneous case of modified Mie theory (MMT) model described in detail in [
The numerical interpretation procedure consisted of the prediction of spectra as functions of the model parameters, comparison of the predicted and measured spectra, subsequent adjustment of the parameters, new prediction, and comparison. An iterative least squares minimization procedure based on a Nelder-Meade downhill simplex optimization algorithm and variable transformation techniques [
The measured spectra of whole blood from healthy donors and anemic patients were analyzed for the RBC indices in accordance with (
The interpretation of the MWT spectra of IRBCs followed (
The composition of the structural groups of the IRBC interpretation model.
Structural group (shell/core) | Composition of the shell | Composition of the shell | Composition of the core |
---|---|---|---|
IRBC body/parasite body/digestive vacuole | Hemoglobin |
Proteins |
Hemozoin |
IRBC body/parasite body/nucleus | Proteins | ||
IRBC body/parasite body/organelles | Proteins |
Each structural group was modelled as a three-layer structure such that the outer layer was the IRBC cytosol, the intermediate layer was the parasite’s cytoplasm, and the core was a parasite’s structural element that provided distinct spectral contribution (digestive vacuole (DV), nucleus (NU), and “average” organelle (ORG)). The validated approximation of additivity of the spectral contributions from the structural groups was used [
The statistical analysis of the whole blood samples from healthy donors and anemic patients included correlation coefficients, outliers, and bias calculations. Correlation coefficients (
The variance in the size, dimensions, orientation, and hemoglobin composition of the RBCs resulted in the breadth in the RBC spectral features as can be appreciated in Figure
The MWT spectra of RBCs in PBS from healthy donors and anaemic patients.
The interpretation of the MWT spectra of RBC fractions with MMT model is illustrated with examples in Figure
Illustration of the interpretation of the measured MWT spectra
The values for selected RBC indices (MCV, MCHC, HCT, and HGB) obtained from the interpretation of the MWT spectra and those from the reference SYSMEX SF-3000 analyzer are compared in Figure
The slopes and correlation coefficients of the regressions between the MWT analysis and reference analyzer SYSMEX SF-3000 values, bias, and standard deviation (S.D.) for the RBC indices (MCV, MCHC, MCH, RBC#, HCT, and HGB) for the samples from healthy donors.
Stat/parameter | MCV | MCHC | MCH | RBC# | HCT | HGB |
---|---|---|---|---|---|---|
Slope ( |
1.00 | 1.00 | 1.00 | 0.99 | 0.99 | 0.99 |
|
0.93 | 0.90 | 0.91 | 0.89 | 0.86 | 0.89 |
Bias ( |
1.2% | 0.9% | 1.6% | 2.1% | 2.3% | 2.5% |
S.D. ( |
0.95 | 0.40 | 0.23 | 0.09 | 0.79 | 0.27 |
The slopes and correlation coefficients of the regressions between the MWT analysis and SYSMEX SF-3000 values, bias, and standard deviation (S.D.) for the RBC indices (MCV, MCHC, MCH, RBC#, HCT, and HGB) calculated for the full data set of samples from anemic patients (
Stat/parameter | MCV | MCHC | MCH | RBC# | HCT | HGB |
---|---|---|---|---|---|---|
Slope ( |
1.00 | 1.00 | 1.00 | 1.08 | 1.07 | 1.07 |
|
0.72 | 0.70 | 0.70 | 0.68 | 0.61 | 0.65 |
Bias ( |
2.1% | 1.8% | 2.5% | 10% | 11% | 11% |
S.D. ( |
5.0 | 0.8 | 1.4 | 0.3 | 2.8 | 0.8 |
|
||||||
Slope ( |
1.00 | 1.00 | 1.00 | 1.05 | 1.04 | 1.05 |
|
0.99 | 0.93 | 0.98 | 0.86 | 0.86 | 0.85 |
Bias ( |
0.8% | 1.0% | 1.3% | 4.9% | 5.0% | 4.4% |
S.D. ( |
0.6 | 0.3 | 0.25 | 0.2 | 1.4 | 0.5 |
Regressions between selected RBC indices ((a) MCV (fL), (b) MCHC (g/dL), (c) HCT (%), and (d) HGB (g/dL)) obtained through the MMT interpretation of the MWT spectra (
Very good agreement between two methods was achieved for the first set of RBC indices (MCV, MCHC, and MCH) for the samples from healthy donors as the
Table
The mean (±2 S.D.) and the range of values of the RBC indices (MCV, MCHC, and MCH) and whole blood RBC indices (RBC#, HCT, and HGB) for the samples from healthy donors (
Samples | Stat/parameter | MCV (fL) | MCHC (g/dL) | MCH (pg) | RBC# (×106) | HCT (%) | HGB (g/dL) |
---|---|---|---|---|---|---|---|
Healthy donors | Mean ± 2 S.D. | 88 ± 11 | 33.4 ± 2.2 | 29.4 ± 4.3 | 4.7 ± 0.9 | 41 ± 7 | 13.7 ± 2.8 |
Range | 76–100 | 31.3–36.5 | 24.9–33.8 | 3.6–6.0 | 34–53 | 11.0–18.2 | |
|
|||||||
Anemic patients | Mean ± 2 S.D. | 94 ± 16 | 30.2 ± 3.6 | 28.3 ± 4.6 | 2.9 ± 1.2 | 27 ± 11 | 8.2 ± 3.2 |
Range | 73–108 | 24.9–34.7 | 21.7–34.6 | 0.09 | 12–38 | 3.9–11.4 |
The MCV distribution for samples from healthy donors approached normal as can be seen in Figure
The values (mean ± 2 S.D.) of the RBC parameters obtained with the MMT interpretation analysis from the MWT spectra of RBC fractions of samples from healthy donors and anemic patients. The samples from anemic patients were divided into three categories: microcytic with MCV < 82 fL, normocytic with MCV > 82 fL and <97 fL, and macrocytic with MCV > 97 fL.
RBC samples/parameter | MCV (fL) | Surface area ( |
Length (nm) | Width (nm) | Sphericity | Aspect ratio | MCHC (g/dL) | MCH (pg) |
---|---|---|---|---|---|---|---|---|
Healthy donors |
88 ± 11 | 119 ± 12 | 8.0 ± 0.4 | 2.2 ± 0.2 | 0.71 ± 0.03 | 0.28 ± 0.02 | 33.4 ± 2.2 | 29.4 ± 4.3 |
Microcytic anemic patients |
77 ± 5 | 126 ± 9 | 8.4 ± 0.4 | 1.7 ± 0.4 | 0.61 ± 0.08 | 0.20 ± 0.06 | 30.4 ± 3.8 | 23.6 ± 3.1 |
Normocytic anemic patients |
92 ± 6 | 121 ± 11 | 8.0 ± 0.4 | 2.2 ± 0.2 | 0.71 ± 0.04 | 0.27 ± 0.04 | 30.5 ± 3.1 | 28.1 ± 3.0 |
Macrocytic anemic patients |
102 ± 5 | 130 ± 10 | 8.4 ± 0.4 | 2.1 ± 0.2 | 0.69 ± 0.06 | 0.26 ± 0.04 | 29.5 ± 3.3 | 30.0 ± 2.5 |
Histogram distribution plots of MCV (fL) index for samples from healthy donors (a) and anemic patients (b).
The MWT spectroscopy is highly sensitive to the changes in the size, shape, and composition of RBC that occurred with infection of
Illustration of the interpretation of the measured MWT spectra
Note that the reconstructed spectra of the noninfected and infected RBCs in Figure
Table
The values (mean ± 2 S.D.) of the RBC parameters obtained with the MMT interpretation analysis from the MWT spectra of noninfected control and
RBC samples/parameter | MCV (fL) | Surface area ( |
Length (nm) | Width (nm) | Sphericity | Aspect ratio | MCHC (g/dL) | MCH (pg) |
---|---|---|---|---|---|---|---|---|
Noninfected |
83 ± 2 | 124 ± 6 | 8.2 ± 0.2 | 2.1 ± 0.1 | 0.71 ± 0.02 | 0.27 ± 0.02 | 31.0 ± 1.2 | 25.7 ± 0.9 |
Ring-infected |
82 ± 3 | 124 ± 3 | 8.2 ± 0.1 | 2.2 ± 0.1 | 0.70 ± 0.02 | 0.27 ± 0.02 | 31.4 ± 2.4 | 22.4 ± 1.7 |
Early troph. infected |
91 ± 10 | 128 ± 6 | 8.2 ± 0.3 | 2.6 ± 0.5 | 0.76 ± 0.05 | 0.32 ± 0.05 | 29.1 ± 7.0 | 19.6 ± 5.2 |
Late troph. infected |
110 ± 12 | 130 ± 8 | 7.9 ± 0.2 | 3.3 ± 0.3 | 0.85 ± 0.03 | 0.42 ± 0.03 | 17.6 ± 1.8 | 9.1 ± 4.0 |
Schizont infected |
76 ± 3 | 110 ± 6 | 7.5 ± 0.3 | 2.6 ± 0.1 | 0.79 ± 0.03 | 0.34 ± 0.03 | 20.4 ± 1.4 | 2.6 ± 0.6 |
A long-standing question in the study of malaria is how parasites prevent premature lysis of infected RBCs during the 48 h parasite asexual cycle and the nature of the final release of merozoites from the infected RBCs [
Graphic illustration of how diseased RBCs can be readily distinguished from normal healthy RBCs using morphological parameters is presented in Figure
Scatter diagram of RBC aspect ratio versus MCV (fL) estimated from the MMT interpretation of the MWT spectra of RBCs samples from healthy donors, anemic patients, and laboratory RBC cultures infected with
The aim of this study was to demonstrate that MWT spectroscopy combined with appropriate spectral interpretation techniques provides reliable quantitative estimates of the morphological parameters and composition of normal and diseased RBCs. We showed adequate performance of the MWT method as its estimates of the RBC indices were in very good agreement with those obtained with a reference laboratory system in both open and blinded studies. Morphological parameters extracted from the MWT spectra were extensively applied in RBC recognition to distinguish between normal cells, diseased cells, and parasite infected cells. The MWT analysis of the samples from anemic patients undergoing chemotherapy allowed comprehensive characterization of the morphological features of three types of RBCs in anemia (microcytic, normocytic, and macrocytic). Microcytic and macrocytic RBCs had greater length and surface area than normocytic RBCs. Further, microcytic RBCs appeared as flat discs with significantly lower sphericity and aspect ratio than other RBC types. Analysis of RBC morphology can give insights into the causes of anemia and even its severity. Our results showed the predominance of abnormal RBC shapes (microcytic or macrocytic) in samples with severe anemia (71%). Further, morphologic characteristics of the infected RBCs derived from the interpretation of the MWT spectra allowed tracing the parasite’s development and provide insights into parasite-host interactions. Changes in the infected RBC volume, surface area, aspect ratio, and hemoglobin composition during the parasites’ development allow for quantitative assessment of the parasites’ mediation of the host cells. The MWT method demonstrated in this study determines the RBC morphological parameters and proves to be valuable for identification of RBC pathologic changes and disease states. Given the portability of the MWT instrumentation and the reagentless nature of the method, it could be used in places where standard complex technologies for RBC analysis are not possible.
The authors declare that there is no conflict of interests regarding the publication of this paper.
The authors are grateful to Dr. German Leparc at One Blood (St. Petersburg, FL) for the assistance with whole blood samples and reference hematology analysis and to Janus Patel and Dr. Wilbur Milhous at College of Public Health, University of South Florida, for the participation in