Smart optics against smart parasites

Abstract

Malaria remains an important cause of high morbidity and mortality worldwide. According to World Health Organisation (WHO) malaria report for 2017, malaria accounted for the death of 435,000 people. It is the leading cause of death among pregnant women and little children. 11% of maternal and 20% of under–five deaths are attributed to malaria every year. Malaria transmission is currently active in 95 countries putting the lives of 3.2 billion people at risk. 40% of the malaria related deaths are linked to Nigeria and the Democratic republic of the Congo. Since malaria symptoms are generally non-specific and usually overlap with the symptoms of other febrile illnesses, clinical diagnosis are typically presumptive and often results into high number of false positives which potentially lead to the abuse of antimalarial drugs. The consistent abuse of antimalarial drugs has produced the consequent effect of drug resistance which is a major concern in the current global malaria control and elimination efforts. The WHO therefore recommends that an effective malaria case management plan must be predicated on a standard parasite-based confirmatory diagnostic test. Conventional light microscopy is the recommended reference diagnostic standard prescribed by the World Health Organisation. This method is particularly of interest because it allows parasite specie differentiation, quantification of the parasite density in a given blood smear, high accuracy (although this depends on the expertise of the microscopist), low direct cost, visualization of different stages of the parasite development etc. While well-equipped laboratories for malaria diagnosis are commonly available in developed urban and peri-urban areas, low-resource settings of malaria endemicity usually have very limited options. The recommended standard microscopy is less accessible in resource-limited settings because of the following: lack of required technical skills, incessant power outages, lack of efficient maintenance capability, delayed diagnosis due to intense workload, inaccuracies due to manual counting of the parasites detected in the blood film etc. The inaccuracies of parasite density estimation eventually affects the accuracy and efficiency of the prescribed treatment which could have fatal consequences. A diagnostic process is termed inconclusive by the WHO until and unless a minimum of 100 measurement (microscopy examination of 100 high powered-fields) has been done on a prepared thick blood film. For a thin blood film which provides more details about the morphology of the parasite, an average of 800 measurement is required. This is an easy task for laboratory technologist in malaria non-endemic countries where an average of 120 malaria cases occur yearly. But for malaria endemic country where several thousand cases are reported daily, this is by no means a mean task as it demands full concentration, time, high expertise and experience. To realize current global effort to reduce the heavy malaria burden, the need for a reliable, efficient, accurate and automated point-of-care diagnostic tool cannot be overemphasized. The focus of this thesis work therefore, is to develop smart optical methods that alleviate the burden of manual microscopy by researching methods to optimise existing imaging modalities which can be integrated with smart algorithms for quick malaria parasite detection in infected patients. Aside malaria, schistosomiasis is the second most common parasitic diseases. Although it falls into the category of a Neglected Tropical Disease (NTD), 220.8 million people required preventive treatment in the year 2017 according to the World Health Organisation report. It is a disease of the poor and it is prevalent in tropical and subtropical areas and particularly common in communities where there is no access to clean drinking water and proper sanitation. 779 million people are at risk of contracting this disease which results into impaired growth and development, diminished physical fitness, bladder cancer and decreased neurocognitive abilities. Although safe and effective medication is widely available for treatment, accurate diagnostic techniques for schistosomiasis is hugely underdeveloped and remains a critical challenge. Intestinal and urogenital schistosomiasis are the two variants of this Neglected tropical disease but in this research, we focus on urogenital schistosomiasis (caused by S. haemtobium) because it is most prevalent among the population we worked with and also because it is easier to detect in urine. The diagnostic protocol for S. haemtobium prescribes urine filtration with WHO recommended standard membrane filters (with 12 μl pore size). Several critical measurements by an expert must be done to detect the targeted foreign bodies (parasite eggs) in the urine samples before a reliable conclusion can be made. Also for a confirmatory diagnosis, it is standard practice to examine different samples collected from the patient at different specific intervals. This is particularly recommended to increase the amount of sample analysis per patient thereby increasing the sensitivity of the test. Since this process involves the microscopy examination of filtered urine samples, it is also limited by the challenges already described for standard malaria microscopy. Although several antigen and antibody based rapid diagnostic test kits have been developed for both malaria and schistosomiasis, the reliability of the performance of these diagnostic test is still a major concern. This thesis is aimed at the development of reliable, robust, accurate, cost effective and easy-to-use point-of-care optical devices for quick diagnosis of malaria and urogenital disease in human samples. This thesis begins by looking at light microscopy with extended depth of field. Wavefront coding with adaptive optics and digital inline holography have been considered in this work. An optimal configuration that guarantees maximum resolution based on the coherence property of illuminating source and the specification of the imaging sensor is prescribed. In this system, interference of a plane and object wave at the detector plane generates a hologram from which the complex amplitude of the field in the object plane can be numerically reconstructed by solving an inverse source problem. This method is of practical interest particularly because unlike the conventional microscope, details in transparent biological samples can be retrieved since both amplitude and the phase of the field is reconstructed. It provides potential solution towards label-free diagnosis of parasitic diseases. Combined with flow cytometry and data-driven algorithms we applied this methodology to the development of rapid detection of S. haemtobium. A working prototype device with the potential to map the diseases has been developed and tested on the field. The system design takes into consideration practical field conditions such as ease-of-use, cost, harsh environmental conditions, erratic power outages, system robustness against dust and other artifacts. Feedbacks and results from the field are very promising. Leveraging on recent advances in cellphone and 3-D printing technologies we developed an automated cell-phone based microscope towards the realization of a rapid point-of-care diagnosis of malaria. The challenge here is to optimise the optical train of a low-cost commonly available cell-phone to detect malaria parasite with sufficient resolution. It was found that existing cell-phone based microscope could not resolve the 1 µm size malaria parasites because of the system optical aberration and the numerical aperture limit of the phone objectives. Although this method demonstrate the capability of the cell phone based microscope to image malaria parasite, however the achievable field of view is limited to 150 × 150 µm. This implies that over 600 measurement is needed for a conclusive diagnosis. We circumvent this limitation by the novel implementation of computer-assisted dry fluorescent microscopy. Using computational analysis of image containing large number of blood cells, we establish a robust statistics which provides reliable diagnostic recommendation. The technique was tested with in vitro and in vivo samples and has demonstrated its suitability for highly sensitive, robust and automated diagnostics of malaria. It requires minimal human intervention, uses simple sample preparation, provides high degree of independence of expert judgement, and has a potential for massive community screening for malaria control and elimination programs. The design specifications for the development of working prototypes presented in this thesis took into account feedbacks from diagnostic experts from the following non-governmental organisations: Doctors without Borders, Malaria Consortium, AMREF, Save the Children and Christian Aid (Nigeria). Also, our methodology was thoroughly validated by discussions and interactions with experts on the field (in Nigeria, Ivory Coast, Gabon, Uganda and Ghana) and with parasitologists, researchers and vaccine developers in the Netherlands, Spain, Ireland and Germany, leading to valuable new insights.”
It is our goal that the diagnostic methods and prototype presented in this thesis will be used to compliment the limitations of the existing diagnostic techniques.