Raman spectroscopic and fractal analysis of blood samples of dengue fever patients

Abstract

BACKGROUND:

The limitations of the existing techniques for the early detection of dengue fever necessitate the development of a powerful optical technique.

OBJECTIVE:

The present work is a study of Raman spectral modifications of blood on dengue infection and thereby to develop a spectroscopic method for its early detection. The images of the samples are subjected to fractal analysis to find the variation of fractal dimensions on dengue infection.

METHODS:

Correlation of platelet counts of dengue infected blood with Raman spectrum modification and fractal dimension. The effect of lowering of blood platelet count due to dengue infection is found to show some interesting changes in the spectrum.

RESULTS:

The full width at half maximum (FWHM) of the two bands in the region 950–1200 cm⁻¹ increase with the decrease of blood platelet count. The increase in fractal dimension gives an indication of the decrease of platelet count and hence the dengue infection.

CONCLUSIONS:

Raman spectrum and fractal analysis can effectively be used as potential techniques for the early detection of dengue infection.

Keywords

Raman spectroscopy dengue fever platelet blood fractals

1. Introduction

Raman spectroscopy is a proven method for structural determination of molecules that enables the detection of even a slight change in the molecular structure [1–3]. It is a powerful analytical technique that offers many advantages, in particular, a high specificity that allows to precisely determine the physicochemical makeup of Raman-active materials [4–6]. The spectrum recording is based on the inelastic scattering of photons as a result of the interaction between a monochromatic light and molecular vibrations. Raman spectra contain information about chemical composition, molecular bonding, symmetry, and structures, as well as other physical parameters [7–9]. These spectral signatures being unique, for a particular type of molecular structure, Raman spectroscopy is widely used in the field of medical diagnosis [10–12].

The present work is the first attempt of detecting dengue through Raman spectroscopic analysis of whole blood. A review of the literature reveals the potential application of Raman spectroscopy for blood serum analysis, platelet and hemoglobin characterizations [13,14]. The dengue fever has now become a common disease in several countries caused by viruses and spread by Aedes Aegypti mosquito. The usual symptoms of dengue fever are severe headache, high fever, vomiting, muscle and joint pain and skin rashes. When it becomes serious, some symptoms like bleeding, blood pressure fall, blood plasma leakage and a decrease in blood platelet counts can be seen [4,5]. The disease can be fatal if not identified and properly treated. Hence, proper diagnosis is essential. The existing detection methods are based on virus isolation, genome detection, and serological analysis [6,15,16].

Since virus isolation and genome detection require sophisticated equipment that is not commonly available, serology is the most widely used method for routine diagnosis. There are reports of using Raman spectrum for the diagnosis of dengue and malaria infection [14,17]. For the detection of dengue infection, M. Saleem et al. [14] have made Raman spectroscopic analysis of blood serum with the antibodies IgM and IgG. The detection of IgM is possible after five days and IgG after eight days of fever symptom [17]. Hence, development of a high sensitive and easy technique is essential. Since blood serum is devoid of platelets, the blood component that varies due to dengue infection, it will be good if the spectrum of isolated platelets, normal whole blood and dengue confirmed blood could be recorded and analyzed.

The human body is a highly complex structure with self-similar structures at the microscopic level. These complex shapes called fractals are self-similar structures, each part of which has the same statistical character as the whole. Mandelbrot [18] gave a method for analyzing these complex structures using non-integer dimensions [19]. These dimensions are measures of irregularities that can be greater than the Euclidean dimension of the object. Though fractal analysis has now emerged as a potential tool for the study of biological systems especially in the field of medicine [20], no report of employing the technique for blood characterization and diagnosis could be seen in literature. Thus, in the present work, Raman spectroscopic and fractal analysis are done for the first time on the blood samples of confirmed dengue patients for monitoring the variation of platelet count.

2. Materials and methods

2.1. Raman spectroscopic study

In the present work, the blood samples (in liquid form) of fifteen persons who were healthy otherwise were subjected to Raman spectroscopic study and attempted to relate the spectral variations to the changes in platelet count. Since platelets are found in blood (not in serum), samples of confirmed dengue-infected blood and normal blood were collected from the Doctors Diagnostic Research Centre, Medical College, Trivandrum, Kerala, India. It is well known that Raman spectra contain spectral signatures of chemical composition, it will be giving a peak at a particular wavenumber corresponding to a given constituent of blood that will not vary from person to person. In this study, a continuous monitoring of blood platelet count of fifteen persons admitted with fever is done. From the blood data analysis report it is found that the platelet count changes from 4 × 10⁵–1.2 × 10⁵ per microlitre and the serology results confirmed the fever to be dengue.

The fact that the blood platelet count changes on the dengue infection led to the present investigation of employing Raman spectroscopic analysis of normal blood and dengue-infected blood. In the present work, since the attempt is to find the spectral modifications of whole blood, Raman spectra are recorded with whole blood. Since platelet count varies on dengue infection, it is essential to record the Raman spectrum of isolated human blood platelet. For this, platelets are isolated by differential centrifugation procedure [21] from a normal healthy person. Platelets are mere circulating fragments of cells found in blood. Salim et al. [17] have done blood serum analysis using Raman spectroscopy for the diagnosis of dengue virus infection.

The Raman spectrometer used for the analysis is Lab RAM HR Horiba and the signals are analyzed using the software Lab spec. The Raman spectrum is recorded using an argon ion laser at 514.5 nm wavelength and a power of 5 mW as the excitation source. The Raman spectrometer is initially calibrated. Raman bands observed without interference from unwanted background fluorescence are subjected to effective baseline correction.

Fig. 1.

Raman spectrum of isolated blood platelets.

Fig. 2.

Raman spectrum of dengue infected blood with platelet count (a) 2.5 × 10⁵∕μl (b) 2 × 10⁵∕μl (c) 1.2 × 10⁵∕μl.

2.2. Fractal image study

The Raman spectrometer consists of a CCD (charge coupled device) camera that helps in imaging the sample. The samples are kept under a microscope of 100 $\times$ magnification and the images are recorded. The images are subjected to fractal analysis. The fractal dimension helps in describing objects with their degree of complexity [18,22]. Of several methods of determination of fractal dimension box counting method is more suitable for determining the dimension of self- similar mass fractals [23]. In this approach, the image is divided into a collection of squares and the number of squares (N (r), where r is the size of the box) with the image portions is counted. N (r) is directly proportional to r ^−D where D is the fractal dimension. Taking logarithm we get $\begin{eqnarray}\displaystyle \ln N(r)=D\ln \left(\frac{1}{r}\right)+\text{constant}. & & \displaystyle\end{eqnarray}$ (1) From the slope of the plot of ln N (r) Vs $\ln (\frac{1}{r})$ , we get the fractal dimension.

Table 1

Band assignment of Raman spectra of platelets

Raman shift (cm⁻¹)	Bands
958	Pure fibrin vibrations [24]
1005	Combination of 𝛿(C=CH) methyl in-plane rocking and 𝛿(C–H) out-of-plane bending modes [25,26]
	Aromatic ring of phenylalanine (Proteins) [27,28]
1122	C–C skeletal mode of acyl backbone in lipids [25–27]
1156	𝜈(C–C) stretching vibrations [27]
1268	𝜈(C–C) stretching vibrations [27]
1366	IgG [17]
1449	𝛿(CH₂) scissoring mode of methylene [27]
1521	𝜈(C=C) in-phase stretching vibrations [25–27]
1669	𝜈 (C=O) stretching of amide I (proteins) [28]

Fig. 3.

Raman spectra of dengue infected blood in the region 950–1200 cm⁻¹ with platelet count (a) 2.5 × 10⁵∕μl (b) 2 × 10⁵∕μl (c) 1.2 × 10⁵∕μl.

Fig. 4.

Variation of FWHM of band B and C with change in platelet count.

3. Results and discussions

3.1. Raman spectroscopic analysis

Since we know that the platelet count is the major variable in the blood samples of dengue patients, the Raman spectrum of blood platelets isolated from healthy normal blood is recorded in the range 950–1800 cm⁻¹. The Raman spectrum of blood platelets separated is recorded and is shown in Fig. 1. The spectrum of the platelet sample shows peaks at 958, 1005, 1122, 1156, 1268, 1366, 1449, 1521 and 1669 cm⁻¹. Among these, peaks at 1005, 1122, 1156 and 1521 cm⁻¹ are sharp. These peaks are used for dengue fever study.

The blood samples of dengue infected persons who were healthy otherwise are collected and analysed using Raman spectroscopy and are shown in Fig. 2(a), (b), and (c).

We collected the blood samples of fifteen patients admitted to a hospital with fever who was later confirmed to be dengue infected from serology tests. The blood platelet count was continuously decreasing from 4 × 10⁵ to 1.2 × 10⁵ per microlitre. The blood samples at different stages of fever were collected and analyzed using Raman Spectroscopy. The Raman spectrum of platelets with an initial count 4 × 10⁵∕μl is shown in Fig. 1. The Raman spectra of dengue-infected blood with platelet counts 2.5 × 10⁵∕μl, 2 × 10⁵∕μl, and 1.2 × 10⁵∕μl are shown in Fig. 2(a), (b), and (c) respectively. From Fig. 2, we can see that on dengue infection the sharp peaks present in the Raman spectrum of normal whole blood become broader with the various frequency components embedded. We can see only eight broad bands marked as A–H. The bands are centered around A – 958 cm⁻¹; B – 1015 cm⁻¹; C – 1156 cm⁻¹; D – 1276 cm⁻¹; E – 1414 cm⁻¹; F – 1503 cm⁻¹; G – 1622 cm⁻¹; H – 1673 cm⁻¹. The peaks appearing in the Raman spectrum of blood platelets are assigned to the vibrational modes as given in Table 1.

When the Raman spectrum of dengue-infected blood sample are closely examined in the region 950–1200 cm⁻¹, it is seen that the full width at half maximum of the two bands appearing in this region increases with a decrease of platelet count and is shown in Fig. 3.

From the comparison of Raman spectra of dengue-infected whole blood and platelets isolated from normal healthy blood, the following information can be deducted. The eight broad peaks (marked as A – H), appeared in the dengue-infected blood are shown in Fig. 2

The peak at 958 cm⁻¹ corresponding to the platelet remains unchanged on dengue infection.

The two sharp peaks at 1005 and 1122 cm⁻¹ disappears and gets merged into the broader band B and C.

The FWHM of bands B and C increase with decrease of platelet count.

The sharp peak of blood platelet at 1156 cm⁻¹ gets broadened on dengue infection.

The peak at 1449 cm⁻¹ of blood platelet disappears. We get a dip at 1449 cm⁻¹ between the bands E and F.

The sharp peak at 1521 cm⁻¹ in the spectrum of platelet merges with the band F.

The peak at 1669 cm⁻¹ remains unchanged on dengue infection.

Thus the Raman spectroscopic analysis of isolated platelets, normal whole blood and dengue confirmed blood revealed some interesting spectral changes that can help to understand the variation of platelet counts. In addition to the merging of sharp peaks found in whole blood, the peaks get broadened. The broadening of peaks may be due to the decrease in platelets and the variations in antibodies due to dengue infection. Whole blood in the liquid form selected for the analysis, eliminates the possible errors in the spectral recording due to smearing sample over the glass plate. The smearing of the sample over a glass plate may not give the correct result if the sample is not uniformly distributed. While using the sample in liquid form, the laser beam passes through the whole length of the sample in the cuvette. This gives greater interaction with the sample under study and gives a better result. Hence it is better to take the sample in the liquid form.

It is also interesting to note that the FWHM of the two bands in the region 950–1200 cm⁻¹ increase with the decrease of platelet count. The variation of FWHM of band B and C with the change in platelet count is shown in Fig. 4. The high R ² value obtained indicate a strong correlation between FWHM and change in platelet count. The R ² value tells how close the data are to the fitted regression line. From the graph, one can see high R ² value for the linear fit.

3.2. Fractal analysis

The images of the samples - (a) normal healthy blood (b) platelets and dengue-infected blood with platelet count (c) 2.5 × 10⁵∕μl (d) 2 × 10⁵∕μl (e) 1.2 × 10⁵∕μl - are also recorded using Lab RAM HR Horiba and are shown in Fig. 5. The platelets appear as spots of size about 20% of the diameter of red blood cells. From the very appearance, we can understand how the degree of complexity varies on dengue infection. Figure 5(b) is the least complex one as it contains only blood platelets.

The figures are analyzed by box counting method and the plots of ${\rm ln}N(r)$ Vs ${\rm ln}\left(\frac{1}{r}\right)$ are shown in Fig. 6(a) to 6(e). The graphs show strong correlation coefficient with high R ² values. The fractal dimensions and R ² values for the samples are shown in Table 2. The increase in fractal dimension is a measure of increase in complexity.

From the plot of fractal dimension Vs platelet count shown in Fig. 7, it can be seen that the two variables exhibit a strong correlation with R ² value 0.9909. This suggests that fractal dimension can be used for dengue analysis where the platelets count changes. The regression equation (equation 2) connecting fractal dimension (D) and platelet count can be used as a predictive equation for platelet count. Thus from fractal analysis, one can understand the platelet count variation.

Fig. 5.

Images of (a) normal healthy blood (b) platelets and dengue infected blood with platelet count (c) 2.5 × 10⁵∕μl (d) 2 × 10⁵∕μl (e) 1.2 × 10⁵∕μl.

Fig. 6.

The plot of ${\rm ln}N(r)$ Vs ${\rm ln}\left(\frac{1}{r}\right)$ for the samples of (a) normal healthy blood (b) platelets and dengue infected blood with platelet count (c) 2.5 × 10⁵∕μl (d) 2 × 10⁵∕μl (e) 1.2 × 10⁵∕μl.

Table 2

Fractal dimensions and R ² values

Samples	Fractal dimension (D)	R ² value
Normal blood with platelet count 4.0 × 10⁵∕μl	1.79	0.993
Dengue infected blood with platelet count 2.5 × 10⁵∕μl	1.71	0.983
Dengue infected blood with platelet count 2.0 × 10⁵∕μl	1.65	0.998
Dengue infected blood with platelet count 1.2 × 10⁵∕μl	1.55	0.994
Isolated platelets	0.94	0.980

Fig. 7.

Plot of fractal dimension Vs platelet count.

$\begin{eqnarray}\displaystyle \text{Platelet count}=12.821^{\ast }∼(D-1.48). & & \displaystyle\end{eqnarray}$ (2)

4. Conclusion

The variations of blood constituents due to dengue infection result in the broadening of Raman bands in the spectrum. Thus the broadening gives an indication of decrease of the platelet count and hence the dengue infection. Since Raman peaks are spectral signatures of molecular structure, study of a large number of samples of a particular platelet count (in the absence of other diseases) will not add any more information. The possibility of fractal analysis also is explored from the unique fractal dimension by box counting method. The platelet count and fractal dimension exhibit high correlation coefficient and increases linearly. The fractal analysis being a statistical procedure gives the possibility of predicting the platelet count from the blood images. Thus, the decrease in fractal dimension gives an indication of the decrease of platelet count and hence the dengue infection. The Raman spectroscopic and fractal analysis of blood samples with low platelet count due to the failure of bone marrow, other diseases, drugs, and pregnancy are in progress.

The data given in the manuscript are from the left over samples in the medical lab where the identity of the patient is not revealed. At the time of starting the work in 2016, there was no human ethical committee in University of Kerala, only an animal ethical committee clearance.

Footnotes

Conflict of interest

This research did not receive any specific grant from a funding agency in the public, commercial, or not-for-profit sectors.

References

Lacci

K.M.

and Dardik

, Platelet-rich plasma: Support for its use in wound healing, The Yale Journal of Biology and Medicine 83 (2010), 1.

Paudel

Raijada

and Rantanen

, Raman spectroscopy in pharmaceutical product design, Advanced Drug Delivery Reviews 89 (2015), 3.

Berger

J.A.

Koo

T.-W.

Itzkan

Horowitz

and Feld

S.M.

, Multicomponent blood analysis by near-infrared Raman spectroscopy, Applied Optics 38 (1999), 2916.

Guzman

M.G.

Halstead

S.B.

Artsob

Buchy

, Dengue: A continuing global threat, Nature Rev. Microbiology 8 (2010), S7–S16.

Guzman

M.G.

Vazquez

and Kouri

, Dengue: Where are we today?, Malaysian J. Med. Sci. 16 (2009), 5.

Guzman

M.G.

and Kouri

, Advances in dengue diagnosis, Clin. Diagn. Lab. Immunol. 3 (1996), 621.

Rosario

Alvarez

Vazquez

Amin

Rodriguez

Valdes

and M.G.

Guzman

, Application of molecular methods to the diagnosis and characterization of a dengue outbreak in Cuba, Biotecnol. Apl. 18 (2001), 203.

Lin

Huang

Wang

and Chen

, Raman spectroscopy of human hemoglobin for diabetes detection, Journal of Innovative Optical Health Sciences 7(1) (2014), 1350051.

Khan

Ullah

Khan

Wahab

Bilal

and Ahmed

, Analysis of dengue infection based on Raman spectroscopy and support vector machine (SVM), Biomedical Optics Express 7 (2016), 2249.

10.

Lorincz

Haddad

Naik

Fung

Cao

Manda

Pandya

Auner

Rabah

Langenburg

S.E.

and Klein

M.D.

, Raman spectroscopy for neoplastic tissue differentiation: A pilot study, J. Pediatr. Surg. 39 (2004), 953.

11.

Kong

Kendall

Stone

and Notingher

, Raman spectroscopy for medical diagnostics-From in-vitro biofluid assays to in-vivo cancer detection, Adv. Drug Deliv. Rev. 89 (2015), 121.

12.

Crow

Uff

J.S.

Farmer

J.A.

Wright

M.P.

and Stone

, The use of Raman spectroscopy to identify and characterize transitional cell carcinoma in vitro, BJU Int. 93 (2004), 1232.

13.

Aslanian

Vainer

Bolard

Guesdon

J.-P.

and Balkanski

, Resonance Raman spectrometric study of human blood platelets, Febs Letters 101(1) (1979), 39–42.

14.

Saleem

Bilal

Anwar

Rehman

and Ahmed

, Optical diagnosis of dengue virus infection in human blood serum using Raman spectroscopy, Laser Phys. Lett. 10 (2013), 035602.

15.

Rosario

Alvarez

Vazquez

Amin

Rodriguez

Valdes

and Guzman

M.G.

, Application of molecular methods to the diagnosis and characterization of a dengue outbreak in Cuba, Biotecnol. Apl. 18 (2001), 203.

16.

Kong

Y.Y.

Tin

C.H.

Tin

T.C.

and Devi

, Rapid detection, serotyping and quantitation of dengue viruses by TaqMan real-time one-step RT-PCR, J. Virol. Methods 138 (2006), 123.

17.

Bilal

Saleem

Amanat

S.T.

Shakoor

H.A.

Rashid

Mahmood

and Ahmeda

, Optical diagnosis of malaria infection in human plasma using Raman spectroscopy, J. Biomed. Opt. 20 (2015), 17002.

18.

Mandelbrot

B.B.

, The fractal geometry of nature, W. H. Freeman and Company, New York: NY, 1983.

19.

Gleick

, Chaos: Making A New Science, Penguin Books, New York: NY, 1987.

20.

Sonia

P.S.

and Fairgrieve

S.I.

, Exsanguinated blood volume estimation using fractal analysis of digital images, J. Forensic Sci. 57(3) (2012), 610.

21.

Vainer

, The platelet populations, Adv. Exp. Biol. Med. 34 (1972), 191.

22.

Barnsley

M.F.

, Fractals Everywhere, 2nd edn, Academic Press Professional, New York, 1993.

23.

Feder

, Fractals, Plenum Press, New York, 1988.

24.

Sikirzhytski

Sikirzhytskaya

and Lednev

I.K.

, Multidimensional Raman spectroscopic signatures as a tool for forensic identification of body fluid traces: A review, Appl. Spectrosc. 65 (2011), 1223.

25.

Rein

A.J.

Saperstein

D.D.

Pines

S.H.

and Radlick

P.C.

, Blood plasma investigations by resonance Raman spectroscopy: Detection of carotenoid pigments, Experientia 32 (1976), 1352.

26.

Movasaghi

Rehman

and Rehman

I.U.

, Raman spectroscopy of biological tissues, Appl. Spectrosc. Rev. 42 (2007), 493.

27.

Huang

Short

Zhao

Wang

Lui

Korbelik

and Zeng

, Full range characterization of the Raman spectra of organs in a murine model, Optics Express 23 (2011), 22892.

28.

Borges

R.C.F.

Navarro

R.S.

Giana

H.E.

Tavares

F.G.

Fernandes

A.B.

and Junior

L.S.

, Detecting alterations of glucose and lipid components in human serum by near-infrared Raman spectroscopy, Res. Biomed. Eng. 31 (2015), 160.