Malaria is a common disease in tropical countries that poses serious public health concerns. It is one of the leading causes of mortality globally. Nearly 0.5 million deaths were attributed to the disease in the year 2015 according to the world malaria report [1]. The majority of these deaths comprise of children less than five years old from sub-Saharan Africa. Besides, it is estimated that approximately 200 million people are infected with the disease and 3.2 billion people are at risk of contracting the disease annually according to the report. This has created a heavy social-economic burden to the affected countries due to huge treatment costs and absenteeism from work and schools [2].

The disease is caused by protozoan parasites of the genus Plasmodium which lives in peripheral blood, spleen or liver of humans. The parasites are introduced to the human host through bites by infected mosquitoes—malaria vectors. Once in the bloodstream, the parasites move to the liver for incubation and multiplication. They later move back to the bloodstream and attack the Red Blood Cells (RBCs). The parasites feed on the cells’ hemoglobin reducing or completely inhibiting the cells’ ability to bind oxygen. Hemoglobin is a complex bio-molecule made of two simpler molecules; the iron compound called heme and a protein molecule globin. The parasite digests globin to meet its nutritional needs but converts heme (which is toxic to the parasite [3] ) into a by-product called hemozoin. Hemozoin, an inert insoluble crystal composed of iron nano-rods [4], is then deposited in parasites’ food vacuoles. Presence of hemozoin in RBCs in peripheral blood is, therefore, an indicator of malaria infection. Notably, hemozoin production is also associated with other blood-feeding parasites such as Schistosoma mansoni and Rhodnius prolixus [5] [6]. However, neither Schistosomamansoni nor Rhodniusprolixus resides in peripheral blood.

There are five species ofPlasmodium parasites that infect humans [7], namely; Plasmodium falciparum,Plasmodium ovale,Plasmodiumvivax,Plasmodium malariae, and Plasmodium knowlesi.Plasmodium falciparum is the most fatal while other species cause mild to moderate illnesses. Each of the five species has distinct morphological features which are used for Plasmodium parasites species identification during optical microscopy diagnosis. The parasites undergo three main inter-erythrocyte life stages, namely; the early stage is also known as ring stage, intermediate stage also called trophozoite stage and Mature stage also called schizont stage (the reproductive stage). Upon maturity of the parasites, the infected cells’ rapture and the merozoites (offspring of the parasites) are released into the bloodstream where they attack other health erythrocytes and the cycle starts afresh. However, some Plasmodium merozoites formed during asexual reproduction of the schizonts differentiate to become the sexual form of the parasites (gametocytes) that are taken up by mosquitoes during blood meals and aid in the transmission of the disease [8]. Figure 1 [9] shows the life cycle of the Plasmodium parasite in the human host.

Malaria diagnosis entails detection of the presence of the Plasmodium parasites in the patient blood and determination of the following parameters; species, life stages, and parasitemia estimation—an indicator of disease severity given by either the number of Plasmodium parasites or the number of infected erythrocytes

per unit volume of blood. A number of malaria diagnostic techniques exist and more are under development. The gold standard method is optical microscopy [10] [11]. The technique involves examination of stained blood smear samples using a light microscope to aid visualization of Plasmodium parasites in erythrocytes. Labeling (staining) of blood smear samples with Giemsa is undertaken prior to examination of the slides. Staining helps to highlight the parasites’ pigment (hemozoin) and hence makes the parasites conspicuous. This technique is time-consuming and labor-intensive due to the need for sample preparation and the enumeration of detected parasites in the microscope field of view. A single malaria test takes between 30 to 60 minutes. In addition, reliability of the results from this method depends on the experience of the microscope operator.

Other malaria diagnostic methods include; Polymerase Chain Reaction (PCR) based technique [12], Plasmodium parasite antigen detection based technique also referred as Rapid Diagnostic Test (RDT) [13] and Fluorescence microscopy-based techniques [11]. Antigen-based technique is widely used and boasts of a relatively fast detection speed of about fifteen minutes. However, the technique has low parasite detection sensitivity and specificity especially when the parasite load is low in the blood. The test is also relatively expensive with a single test costing about 1US$. In contrast, PCR technique has high detection sensitivity and specificity. It probes and amplifies genetic material of the Plasmodium parasites present in the blood. However, the technique demands the use of sophisticated equipments and expensive reagents. Besides, a highly skilled operator is also required to perform the test. The technique is therefore used in research laboratories and confirmation of malaria tests. Fluorescent based technique relies on staining of blood smears with a fluorescent dye that would highlight the parasite nucleus. A fluorescent microscope is then used to observe the stained blood sample. The technique has also not yet found wide clinical adoption in malaria screening due to high cost of equipment, reagents and demand for highly skilled expertise.

Some of the malaria diagnostic techniques reported by literature as being under development include; Third Harmonic Generation (THG) technique [14], Magneto-optic based technique [15], and computer vision assisted optical microscopy technique [16] [17]. Label-free optical microscopy using multispectral imaging and chemometrics techniques has also been explored [18] [19]. Recently, a few publications have reported on the development of Photo-Thermal (PT) and Photo-Acoustic (PA) based techniques for in vivo malaria diagnosis [4] [20] [21] [22] [23]. Preliminary results regarding the techniques’ performance are impressive but more work is needed before the method can be clinically accepted for malaria screening.

Generally, the mainstream malaria diagnostic techniques suffer from a number of shortcomings [10] [11] which include low detection sensitivity and specificity, labor and time intensiveness and demand for reagents and sophisticated equipments which are often expensive and inaccessible to people in resource scarce settings where often malaria is endemic. There is therefore an urgent need for development of new and more efficient techniques for malaria diagnosis. The remaining part of this paper explains in detail the principle behind photoacoustic-based malaria diagnostic technique and the reported progress.

The major chromophores in malaria infected blood are total hemoglobin (Hbt) and hemozoin [20]. Hbt is the sum of oxy-hemoglobin (HbO₂) and deoxy-hemoglobin (Hb) molecules in the blood. Figure 2 is an excerpt of the spectral plots of hemozoin (Hz) and Hemoglobin (Hb) extracted from the work of saha et al. [16], and Cai et al. [18]. Table 1 gives values of the optical absorption coefficients and molar concentrations of Hemozoin (Hz) in an infected erythrocyte

as extracted from the work of Saha et al. [20]. Based on the values of molar concentration of Hz present in the cell, it is possible to differentiate various stages of infection. Previous studies have demonstrated the possibility of in vivo imaging of oxygen saturation and malignant tumor hypoxia [50] [68] - [73], oxygen metabolic rate [74], methemoglobin [75] and oxygen release rate in blood vessels and individual erythrocytes using sub-wavelength Optical Resolution Photoacoustic Microscope (OR-PAM) [76]. This suggests that a similar approach can be used to detect the presence of other dominant optical absorbers in individual erythrocytes such as hemozoin in case of malaria.

The following is a brief analysis of the problem. Optical absorption coefficient due to a malaria infected erythrocyte can be described by the following equation.

μ a λ i = ε H z λ i [ c H z ] + ε t H b λ i [ c t H b ] (1)

ε i λ i is the molar extinction coefficient of the i^th chromophore present in the cell. Both μ a and ε are wavelength dependent hence the superscript λ i . [ c i ] is the molar concentration of the i^th bio-molecule in the cell. The initial photoacoustic signal generated due to absorption of an optical pulse of wavelength λ i is the product of the deposited optical energy A e = μ a λ i Φ λ i and a factor called gruneisen parameter (Γ). Φ is the local optical fluence distribution in the blood which depends on the intensity of the incident light as well as blood optical properties (mainly, optical absorption and scattering coefficients). In case the chromophore relaxation after photon absorption is not 100% non-radiative, a down scaling factor is included in the product to account for the fraction of photon energy converted to heat energy. Therefore, the initial PA pressure is given by the following equation [63].

P 0 λ i = Γ n t h μ a λ i Φ λ i (2)

With the transformed thermal energy density due to non-radiative relaxation of chromophores given by

H = n t h μ a λ i Φ λ i (3)

n t h is the fraction of optical energy converted to heat energy ( n t h = 1 for non-fluorescent and non-photoluminescent chromophores). The grueneisen parameter is a thermodynamic quantity which gives the photoacoustic efficiency of the tissue, i.e., the fraction of the thermal energy converted to acoustic waves.

The generated acoustic signal P_o (also referred as initial PA pressure) propagates through the tissue and is detected by an ultrasonic transducer placed at the tissue surface, a distance r from the target absorber (in this case a skin subsurface blood vessel). The detected acoustic signal, P_d differs from the initial PA signal due to; acoustic attenuation through the tissue, partial detection geometry of the traducer and the transducer’s detection efficiency η ( f ) . The transducer output signal, P_d is therefore a product of the initial signal, P_o and a scaling factor, T_d which accounts for the described signal attenuation i.e.,

P d ( f ) = T d ( f ) P o (4)

with

T d ( f ) = η ( f ) N A 2 e − α o f 2 l (5)

where; NA is the traducers’ numerical aperture, l is its focal length and α o is the tissue attenuation coefficient. Numerical values for the parameters in Equation (4) for some commercially available ultrasonic transducers can be found in reference [77].

A limited number of studies have so far been carried out regarding in vivo photoacoustic diagnosis of malaria. Generally, the principle of in vivo Photoacoustic Flow Cytometry (PAFC) [78] has been adopted. The technique entails use of a focused ultrasonic transducer to detect PA signals emitted due to optical excitation of the flowing erythrocytes in blood vessels located below the skin surface. The signal peak intensity is monitored in time to keep track of the flowing erythrocytes (infected and non-infected). Healthy erythrocytes produce a nearly constant amplitude signal. Hemozoin – an endogenous malaria biomarker has either higher or lower optical absorption than hemoglobin depending on the optical wavelength used. This implies that the amplitude of Plasmodium infected blood would differ from that of healthy blood. The following is a brief review of the few reported literature.

Cai et al. [22] proposed an in vivo (PAFC) technique for early diagnosis of malaria. They used a focused US transducer for detection of PA signal from flowing erythrocytes irradiated by three pulsed laser lights of wavelengths 532 nm, 671 nm, and 820 nm. The signal was then amplified, digitized and loaded to a PC for processing. The system’s spatial resolution was reported to be between 1 - 6 µm. Both in vitro and in vivo experiments using mice as the animal model were conducted to determine the sensitivity of the device. The authors reported that their technique was 1000 times more sensitive than optical microscopy method and could detect an infected erythrocyte 3 hours after inoculation of Plasmodium parasites into a healthy mouse compared to 3 days taken before detection using optical microscopy. They observed that when the optical energy used was increased beyond a certain threshold, hemozoin vapor nano-bubble formation occurred in the infected cells and this produced PA signal of increased intensity—a process termed as nonlinear photoacoustics. They also used the process to perform real time destruction of infected cells (photo-thermal malaria therapy) while monitoring the efficiency of the therapy. They observed that the PA signal ratio of infected cells to non-infected cells was highest at 671 nm optical excitation wavelength.

Lukianova-Hleb et al. (2014) [21] described a picosecond pulsed laser technique for generation and detection of Hemozoin induced Vapor Nano-Bubbles (H-VNB). The authors argued that when hemozoin crystals in infected RBCs absorbs the ultra-short pulsed radiation (having picoseconds range pulse width) of sufficient energy, nano-volume liquid around the hemozoin crystals is overheated and vaporized resulting in generation of transient vapor bubbles which undergo explosive expansion and collapsing—a process that is accompanied by emission of high intensity acoustic waves. Both in vitro and in vivo sets of hemozoin positive and hemozoin negative samples consisting of blood smears, whole blood and live mice were studied. Laser pulses were shone on samples and the emitted acoustic waves detected by an ultrasonic transducer. The transducer output was coupled to a digital oscilloscope. The detected signal was then analyzed in the time domain. It was found that samples containing hemozoin produced acoustic signals of higher amplitudes than hemozoin negative samples. It was reported that the technique yielded high correlation in differentiating Plasmodium parasite life stages and parasitemia estimation in comparison with optical microscopy. The optical wavelength used for excitation of PA signals was 672 nm—an absorption peak for hemozoin. The optical energies per pulse used for in vitro and in vivo experiments were 10 μJ and 15 μJ respectively. Lukianova-Hleb et al. (2015) [23] used the technique for in vivo detection of parasites in humans and infected mosquitoes. Similar results were obtained.

In general, photoacoustic based malaria diagnosis boasts of several advantages over other diagnostic methods which includes; use of endogenous contrast agent (hemozoin) for parasite detection as opposed to exogenous contrast agents as is the case in optical microscopy, potential for in vivo diagnosis. Use of non-ionizing radiation, which makes it safe, real time monitoring and capability for early detection of the disease hence making treatment more effective. In addition, it is possible to use the technique for diagnosis of other blood related disease. However, the reported photoacoustic diagnostic systems suffer from a number of challenges which must be addressed before the technology can be clinically adopted. For example, the technique cannot perform quantitative measurement of chromophore molar concentration in absolute units. In most of the reported work [21] [22] [23], the distinction between healthy and diseased samples is made based on the relative amplitudes values of the induced PA signal between the two sets of samples. The limitation of this approach is wide signal amplitudes variability within subjects of the same class, which makes system calibration difficult. Secondly, the reported PA systems are bulky and expensive due to the demand for high-energy nanosecond pulsed multispectral laser sources, high frequency sensitive ultrasonic transducers and lock-in amplifiers. Besides, the use of fluence estimation models for quantitative PA spectroscopy makes the measurements time intensive and quasi real time. Their accuracy also depends on how well the tissue geometry is known a priori.

There is a promising trend in addressing these challenges going by the number of publications and research groups working on the problem. Simplification of PA hardware has been attempted by substitution of the bulky and expensive Optical Parametric Oscillator (OPO) tuned laser sources with much smaller and cheap diode laser sources [79] [80] [81] and Light Emitting Diodes (LEDs) [82] that have a wide range of optical wavelengths and are commercially available. Notably, Quantellaser^TM has developed a compact high energy multispectral pulsed diode laser module operating in four different wavelengths (808, 915, 940 or 980 nm) in the NIR region customized for PA applications. Its high optical pulse energy (of 1 mJ per emitted wavelength per 80 nS pulse width) is achieved by stacking multiple diode lasers of the same wavelength. Attempts to develop a compact hand held photoacoustic probe for in vivo imaging have also been reported [83] [84] [85] [86].

Use of chirped FD techniques in photoacoustic spectroscopy applications [50] [68] has the potential of simplifying the instrumentation complexity of the PA systems by enabling use of cheap and portable CW diode lasers as PA optical sources. Moreover, eliminate the need for computationally demanding numerical algorithms for fluence estimation. Development of a sensitive FD photoacoustic technique for hypoxia monitoring and early cancer detection has been reported [34] [35]. The technique is termed as Wavelength Modulated Differential Photoacoustic Technique (WM-DPAS). The technique uses two optical wavelengths (680 nm and 808 nm) to monitor variations in oxyhemoglobin concentration in tissue. One wavelength (680 nm in this case) should have maximum variation in extinction coefficients for oxyhemoglobin and deoxyhemoglobin while the other wavelength (808 nm which is the isosbestic wavelength) should ideally have the extinction coefficients coinciding. Two CW laser sources are intensity modulated out of phase (differentially) and their output coupled to the sample for photoacoustic excitation. The emitted PA signal is detected by an ultrasonic transducer. Upon signal preprocessing, the Fourier transform of the signal is performed to obtain both amplitude and phase signals of the differential PA time domain signal. The phase signal is said to be independent of the optical fluence and its variation is only due to variations in the molar concentration of the probed chromophore. Judging from the reported features and performance of WM-DPAS, this technique is likely to be very suitable for in-vivo detection of malaria.