Malaria and the ‘last’ parasite: how can technology help?

The first record of malaria fevers dates back to the 5th century BC [1]. Today, malaria remains one of the four most life-threatening infectious diseases worldwide, together with tuberculosis, HIV/AIDS and hepatitis [2]. Latest data published by the World Health Organization (WHO) are staggering: more than 216 million cases in 91 countries and more than 400,000 deaths occurred globally in 2016 [3]. These figures are the same as in 2015, indicating that despite the unprecedented efforts in recent years, progress has stalled. This calls for more effective tools to reduce malaria and finally to eliminate this scourge. If this historical milestone can be accomplished, it could save the global economies $2 trillion by 2040 [4].

This review only focuses on relevant innovative diagnostic technologies for malaria elimination settings where the malaria transmission is low; therefore, there is a critical need to detect asymptomatic individuals. Together with other effective interventions, ultra-sensitive rapid diagnostic tests are much needed to identify the invisible reservoirs. The role of innovative tools becomes crucial in the fight against malaria and the WHO identifies three strategic pillars (universal access to prevention, drugs and diagnosis, elimination and surveillance), of which accurate and effective diagnostics at the point-of-care (POC) is the first step towards appropriate diagnosis and treatment for malaria infection [5, 6].

Table 1 compares the performance of currently available malaria diagnostic tests for case management and surveillance. The landscape for malaria diagnosis can be divided into two main groups, POC methods in case management and laboratory-based methods for surveillance [7]. In case management, microscopy and RDTs are the two diagnostic methods that are recommended in primary settings whilst highly sensitive RDTs and molecular diagnostics [polymerase chain reaction (PCR) and loop mediated isothermal amplification (LAMP)] are often used in laboratory settings [8]. While presenting ultra-sensitivity (less than 2 parasites/μL for both Pan and Pf-LAMP) in the field [9, 10], implementing malaria diagnostic tools in the field still requires addressing of several critical challenges such as simplified sample preparation steps, ready to use kits that require no cold chain [11]. Further, there is no reported literature referring to the use of malaria LAMP as a diagnostic tool in populations, or of being endorsed and procured by any programs or governments. In the meantime, also being less sensitive, conventional RDTs are at much lower cost of approximately 1 $USD per test [12]. Field studies have shown that POC methods such as microscopy and rapid diagnostic tests (RDTs) are effective in low-resource settings (LRS) [10, 13‐25].

Microscopy is the reference standard for visualization of parasites in blood smears with an analytical sensitivity under normal circumstances approximately tenfold inferior than that of molecular testing [26]. Microscope has been commonly used as a diagnostic tool in peripheral health centres for various reasons, including availability [27]. However, the quality of such diagnosis depends on the availability and skills of trained microscopists, which might not always be available in the LRS, where malaria is endemic.

Field studies have confirmed the benefits of introducing RDTs into routine testing such as better case management, improved adherence to test results, and having more rational treatments [28, 29]. Characteristics of current malaria RDTs are summarized in Table 2. Key advantages of RDTs are the ease to use and quick result delivery time (15–20 min). Unlike PCR or microscopy, RDTs detect circulating antigen; therefore they can also be used to detect placental malaria [30]. Diagnosis of malaria in pregnancy is challenging because of placental sequestration, which is specific to Plasmodium falciparum infections, can make microscopy detection of parasites difficult.

Although using the same technology of lateral flow immunoassays, the performance of malaria RDTs varies greatly from brand-to-brand, and lot-to-lot, especially with specimens having low parasite density (< 200 parasites/μL). In a collaboration between the Foundation for Innovative New Diagnostics (FIND), the WHO and the Centers for Disease Control and Prevention, 293 malaria RDTs were evaluated from 2008 to 2016 [31]. Most of the evaluated malaria RDTs detect P. falciparum histidine-rich protein 2 (PfHRP2) or P. falciparum lactase dehydrogenase (PfpLDH). In the last round of evaluation, anomalies that interfered with result interpretation were also recorded [31]. The most common anomalies were incomplete clearing and red background, which were observed in 48 and 24% of products. The second most common anomalies were failed migration of liquid, incomplete migration and patchy broken test lines, which occurred in 15, 11 and 11% of the products, respectively.

The performance of lateral flow-based RDTs depends on two main factors: the sensitivity and specificity of antibody-antigen combinations, and the ability to facilitate reliable liquid migration on the nitrocellulose membrane. Much research has focused on new biomarker discovery [32‐34], and only limited attention has been paid to reduce limitations imposed by the inhomogeneous migration of liquid across porous nitrocellulose membranes [35].

Figure 1 illustrates how unstructured the flow paths could be in a nitrocellulose membrane [36]. As the migration of liquid occurs in a porous network and is not actively controlled, a number of limitations arise: large volumes of sample needed, accumulation of reagents at the leading edge of the liquid flow, and increased cross-reactivity [37]. It is, therefore, time to consider alternative options to facilitate a more precise liquid migration, hence more accurate test results.

Table 3 summarizes six major classes of technologies used for detecting malaria and indicates their maturity levels. These technologies are individually reviewed in depth elsewhere [38] and most of them rely on standard concepts using immunoassays [39, 40], molecular diagnostics [41‐49] and the visualization of parasites [50‐53]. Table 4 provides specifications of some recently entered market malaria diagnostic [38]. Of those market-ready products, four of them are molecular diagnostics, three are immunoassays and one is based on automated microscopy. Several promising proof-of-concepts for the next generation of malaria RDTs are emerging. For example, prototypes have been built to detect the presence of haemozoin in blood sample [54‐57]. Haemozoin crystals are produced by Plasmodium parasites as a final non-toxic compound of haemoglobin metabolism. In a specific example, a portable light meter was built to image crystalized haemozoin pigment [58]. These pigments are birefringent, so the detection of haemozoin is based on rotating a plane of polarized light through them and observing anisotropic output of the light. The minimum concentration of haemozoin that could be detected with this polarized light system was 15 pg/mL, equivalent to 30 parasites/μL of blood. Applications in the field are to be tested.

Another example utilizes a portable breath analyzer: breaths of malaria-infected patients were found to contain terpenes, a family of aromatic chemicals that are produced by parasites that can further attract mosquitoes [59, 60]. A pilot study in Malawi confirmed that these aromatic compounds could be transported into the lungs and hence could be detected in the exhalation of infected patients [61].

Despite being unquestionably novel, these abovementioned methods of detection still need to prove their practicality for POC in LRS and demonstrate a clinically relevant limit of detection (LOD). For instance, in the breath analyzer, it would be useful to be able to convert the level of terpenes detected in breath into parasite density.

Different settings require different target product profiles (TPP) [8]. Unlike previous malaria control campaigns, the key characteristics of malaria elimination efforts are to interrupt endemic transmission and to prevent its re-establishment [62]. The Program for Appropriate Technology in Health (known as PATH) and FIND are pioneering the development and validation of sensitive rapid tests for mass screening in LRS. They also proposed a TPP for malaria RDTs in elimination settings, stating specific requirements for the ideal rapid tests according to concept of Affordable, Sensitive, Specific, User-friendly, Equipment-free and Deliverable (ASSURED) [63]. The desired LOD is 5 parasites/µL or less, or concentration range of 6–12 ng/mL PfHRP2 [63]. For RDT developers it is important to note the caveat of the prozone phenomenon that might prevent detection of high parasite density [64]. Poor specificity could lead to over-treatment, thus depreciation of the intended value of RDTs (from public health perspectives); therefore, the required specificity for effective malaria diagnosis is at least 97% or ideally 99% [63].

Additional requirements for ideal RDTs are suitability and appropriateness for LRS where most malaria cases occur. To make an impact simplicity and affordability are of utmost importance. Simplicity means, the system should be equipment-free and should require very little resources [65]. A simple and automated test could obviate false results caused by user-errors [66]. Affordability is difficult to measure and depends on the cost–benefit equation of a specific situation. Also, tests should be designed to minimize impact of inappropriate storage conditions (2–40 °C) on reagent stability and usability of the devices [67].

Microfluidics enable the miniaturization and simplification of complicated analytical processes while consuming less reagents, minimizing waste, and requiring less supporting instrumentation [68]. This stems out from the predictable behaviour of liquids at the microscale where flow is typically laminar. At microscale, minute amounts of liquids can be manipulated using microstructures, such as microvalves, micromixers or micropumps [69]. Low volumes of reagents, fast reaction times, compact and portable platforms are just a few advantages that make microfluidics technology attractive for POC applications [70, 71]. Figure 2 shows several examples demonstrating the archetype of microfluidic-based diagnostics for POC applications, which is an integrated system composed of a disposable unit (where analysis takes place) and a signal acquisition and processing module to process the results. (a) [72], (b) [73], (c) [74].

Currently, microfluidic-based diagnostic devices can be divided into two categories: non-paper-based “traditional” microfluidics and paper-based microfluidics [75, 76]. Research on traditional microfluidics often focuses on miniaturizing conventional techniques. For example, a collection of passive and active mixing elements were designed to facilitate mixing processes on chips [77]. Recent work in developing microfluidic-based diagnostic devices has focused on integrating all necessary elements into stand-alone platforms [78, 79] because such integrated systems can operate without bulky accessories and do not require water, buffer, or a constant supply of electricity [80]. There are many ways to control liquid flows on microfluidic platforms, for instance, acoustic forces, mechanical forces, magnetic forces, as well as capillary and centrifugal forces [81‐85]. To satisfy the stringent requirements for LRS, devices based on capillary and centrifugal forces have shown promising results. Table 5 presents some examples of microfluidic-based systems that have been designed to detect PfHRP2 and PfpLDH antigens or genetic materials from the parasites using on-chip molecular testing, cell deformation mechanism, electrical, optical, and magnetic detections amongst others [54, 58, 79, 81, 86‐94].

Standard protocols to perform immunodiagnostics on microfabricated platforms require sample pre-concentration, flow control and detection of biomarkers (analytes and/or parasites). These multi-step protocols can benefit greatly from miniaturization, and in fact, microfluidic-based immunoassays have demonstrated their potential for reliable and accurate performance [95, 96]. Figure 3 presents some examples to illustrate how microfluidics technology can be used to detect malaria by different methods of detection, such as molecular testing, size-based cell sorting, electrical differentiation of healthy and infected red blood cells, optical detection of antigen and magnetic detection of haemozoin. (a) [97], (b) [88], (c) [91], (d) [79], (e) [94].

At the moment, PCR and LAMP are the most sensitive technique for identification of asymptomatic individuals, for example, in 130 clinical samples presenting no parasites based on microscopy, as low as 3.6 × 10⁻⁴ parasite/μL could be identified in 117 samples by a highly sensitive genus-specific quantitative reverse transcriptase real-time PCR (qPCR) [118]. This low LOD was achieved by amplifying and detecting the total nucleic acids of the 18S rRNA genes, which increased the analytical sensitivity of the assay by more than 1 log unit compared to DNA only. However, current applications of PCR and LAMP are still restricted to well-equipped laboratories and thus not suitable for LRS [119]. Miniaturized PCR and/or LAMP is desirable, but developing such devices is a more challenging task than that for biomarkers detection for three reasons: (1) sample pre-treatment is essential for extracting DNA of parasites for downstream analysis, (2) the critical signal amplification step highly depends on temperature control, and (3) robust, low cost, and portable detection techniques are required for remote settings [120].

Paper-based microfluidics was proposed by Whitesides and colleagues [127]. Since then, this technology has been growing fast with great promises for global health applications [128]. Unlike its sister products of paper test strips, paper-based microfluidic analytical devices offer well-defined, millimetre-sized microchannels to transport liquids in a controlled manner, yet with low cost for production (< $0.01) [129]. Using hydrophobic “inks” to define areas on hydrophilic paper, it is possible to perform multiple immunodiagnostic assays on the same test strip. To illustrate how complex analytical processes can be simplified and transformed into a paper-based microfluidic device, Pereira et al. integrated concentration and detection steps into a single step assay [130]. The analyte PfpLDH in low abundance was first accumulated using a micellar aqueous two-phase system (ATPS). The micellar ATPS consisted in a nonionic Triton X-114 surfactant, which was used to concentrate biomarkers in a sample and enhance the LOD. In this system, a tenfold improved LOD of 10 ng/μL PfpLDH was achieved. In an alternative development of a foldable, card-like test device, PfHRP2 could be detected and quantified [131]. The generated signal in presence of PfHRP2 was amplified by gold nanoparticles, yielding a LOD of 1.2 ng/mL PfHRP2, which is four times higher than that of the unamplified case. These studies serves as excellent examples for low cost, non-instrumented analysis systems without compromised performance. Many other innovative approaches to control liquid flows such as selective hydrophobic rendering or origami in which folding of multiple paper layers to trigger reactions were also investigated successfully [132‐134].

Mobile health applications have rapidly been growing in recent years and there is a trend in interfacing consumer electronics such as smartphones with lateral flow RDTs or microfluidic-based devices [135, 136]. Such combination is expected to deliver increased objectivity of test result interpretation and improved connectivity of the entire healthcare systems. The automation and digitized test results can be more easily combined with other health related parameters and combined with medical decision support systems. User-friendly interfaces, automated result analyses, remote-monitoring and data aggregation, increased storage conditions, and active quality assurance are just a few additional benefits of this approach [137].

In 2008, paper-based microfluidics were integrated with a smartphone camera to perform immunoassays [128]. The camera of the phone was used to take a photograph of the detection zone before and after the deposition of specimen. Since then, many groups have started to develop and enhance capabilities of phone-based low cost diagnostic readers [136]. Table 6 presents an overview of recent work in developing phone-based prototypes that can be used to detect variety of biomarkers for a wide range of diseases with clinically relevant performance. Devices are designed for a broad spectrum of applications, from genetic testing, cancer detection to personalized food allergen monitoring [136, 138‐140]. A wide range of strategies are also derived to enhance signal strength, for instance, using Quantum dots, Rayleigh/Mie scatter or gold nanoparticles [141‐143]. At present, applications of smartphone-based diagnostics for malaria detection can be divided into two categories: phone-based RDT readers, which provides automatic interpretation of results, and phone-based brightfield microscopes, which allow simple and portable means to visualize parasites in blood samples [138‐149].

A smartphone was used for quantitative reading of the Optimal-IT test, a commercially available malaria RDT with a snap-on unit as reader that is suitable for both Android and iPhone [145]. Images of RDTs were acquired, in either transmission or reflection, and then processed in real time to deliver test results within 10 min. The spatio-temporal information collected by this device can document prevalence of many infectious diseases and would allow efficient tracking of epidemics. Another approach to integrate a custom microfluidic-based immunoassay detecting PfHPR2 with phone-based detection was the development of a microfluidic chip, which can be connected to a phone camera to analyze signals and deliver results in 10 min. The opto-mechanical unit in this case consisted of optical fibers, microfluidic chips and mirrors, and could be easily removed from the back camera of the phone. The principle was to quantify changes in fluorescent intensity upon capturing of PfHPR2 on the sensing region, yielding a LOD of 1 pg/mL of PfHRP2 in 10% diluted blood [144].

Accurate and consistent blood smear reading is challenging to attain in health centres or small clinics in remote regions. A phone-based microscope is a low cost option that can offers enhanced image quality, improved accuracy and user comfort [146, 150]. There are two simplified imaging techniques suitable for smartphone apps: (1) lens-free holographic imaging, and (2) on-lens devices.

Holography is an image-constructing technique using scattering and interference of light and pixel super-resolution to enhance optical images [151]. An automated lens-less holography was developed with a sufficient field of view of 24 mm² to visualize and capture images of P. falciparum in blood smears [152].

Phone-based microscopy can also be engineered to be a field-ready polarized light microscope without compromised fidelity and resolution [153]. The principle was to detect light birefringence caused by the crystallization of haemozoin. This field-based, modular microscope could magnify Plasmodium chabaudi parasites up to 50 times, gaining a comparable performance compared to conventional polarized microscope. Additional benefits of this prototype are simple operations and low cost per test. Further work using clinical samples could confirm the full potential of this novel phone-based polarized light microscope.

Accurate and effective diagnosis is the first step to further pursue efforts to eliminate and reduce the global burden of malaria by 90% in 2030. Current diagnostic methods can detect malaria symptomatic infections, but often miss out asymptomatic cases. The rise in proportion of asymptomatic infections in low transmission areas calls for a new generation of rapid diagnostic tests that can detect the hidden parasite reservoir. Technology is advanced nowadays to (at least theoretically) be able to track down the last parasite carriers. While malaria case management has improved, other causes of fever need to be detected and treated accordingly. Therefore, the ideal RDT should come in as a complete package with ultra-high sensitivity and specificity, meet the ASSURED standards for LRS, and also provide additional diagnostic capabilities. Microfluidic devices coupled to phone-based readouts offer a unique opportunity to not only reduce the burden of infectious diseases, such as malaria, but also could provide tools for monitoring epidemics and elimination progress on very large scales.