Nowhere does the burden of disease weigh as heavily on people as it does in the low-resource settings of the developing world. Residents in parts of Asia, Africa and South America are highly vulnerable to communicable diseases such as malaria and tuberculosis. They also face increasing risk of developing noncommunicable “lifestyle” diseases common in the developed world such as obesity, high blood pressure and cancer. But people in low-resource regions don’t have access to the level of medical testing residents of wealthier countries enjoy and assume. One study from 2013, for example, estimates that pathology coverage in Sub-Saharan Africa is about one-tenth of that in the developed world. The consequences are stark: Diseases that are treatable or preventable in the developed world kill people in less developed regions. Built around a cellphone camera, the portable EVA System enables nurses to visualize the cervix at optical specifications that are functionally equivalent to colposcopes in the developed world. Here, nurses in a Baja, Calif., clinic go over results. Courtesy of MobileODT. Photonics research and innovation, particularly in the fields of microscopy and imaging, are making incredible strides to alter this picture through low-cost portable optical imaging devices capable of bringing improved medical diagnostics to some of the poorest regions in the world. One such device is the EVA (Enhanced Visual Assessment) System, an integrated, cloud-supported digital tool that screens for cervical cancer. Brought out by MobileODT of Tel Aviv, Israel, the handheld EVA System, which weighs a mere 450 grams, makes use of a smartphone camera to function as a mobile colposcope. As a light source, the system uses a compact high-power LED with extremely low thermal resistance and cross-polarization to reduce glare. It also comes with an external lens capable of up to 17× magnification. “Adding external optics to the smartphone lens is an optomechanical challenge and has several effects on imaging,” said David Levitz, co-founder and chief technical officer of MobileODT. “First, less light reaches the camera, which impacts your SNR [signal-to-noise ratio]. Second, with every piece of hardware added, there are higher chances of misalignment. Third, each smartphone camera has unknown Bayer filters and an IR-blocking filter that alter the light reaching the sensor and closed camera-control algorithms that further manipulate the images.” The EVA System is specifically designed around the optics of a Samsung J5 phone, which is sold with it. The company chose the Samsung, said Levitz, because of its price and technical specs and because it is available in markets where the company is active. The device works on a rechargeable battery for up to 10 hours and comes with a hard-shell water-resistant carrying case. In remote regions — where patients have little or no access to electricity and running water, let alone Pap smears, HPV tests or biopsies — visual screening for cervical cancer is the norm. The system can enable nurses to visualize the cervix at optical specifications that are functionally equivalent to colposcopes in the developed world. The EVA System capable of 17× magnification, for example, has a field of view of 54 mm, a depth of field of 17 mm, and a working distance of 250 mm. Another EVA System that is capable of 16× magnification has a 106-mm field of view, 34-mm depth of field and a working distance of 450 mm. The device most often finds cervical dysplasia, which can be treated by cryoablation. Patients found to have full-blown cervical cancer are referred to higher-level medical facilities. To date the device has been used in more than 20 countries. MobileODT, which recently opened an office in Nairobi, Kenya, primarily focuses its activities in Africa. But the company is also active in Latin America, the Caribbean, Southeast Asia and underserved areas in the U.S. “We are launching a couple of pilots in India this year, which is the country with the highest death toll for cervical cancer worldwide,” said Levitz. “But for me, personally, the most meaningful place where we’ve sold our technology is Afghanistan, where, to our knowledge, they don’t even have gynecologists, only obstetricians, and there were no cervical cancer screening programs in the country.” One of the frequent challenges to health care delivery in low-resource settings is that health care workers in the field often don’t have the training to consistently identify the diseases they encounter. To address this challenge, the EVA System, like many others seeking to improve point-of-care (PoC) diagnostics in the developing world, relies on the internet. The system’s CervDX app allows nurses to upload images to a HIPAA-secure image portal through which they can document what they see, access decision-making tools and consult remotely with medical experts. Training on the device depends on how familiar a user is with a cellphone, but typically takes a couple hours. The cost of the EVA System is $2400, with an annual subscription for additional software. MobileODT is also working with the Global Good Fund, a collaborative venture between Bill Gates and patent licensing company Intellectual Ventures, on developing a machine-learning algorithm for automated detection of cervical pathology in the EVA System’s images. It is among the Global Good Fund’s many projects fostering the design of PoC devices for low-resource settings. Another project is for an automated microscope designed to improve the diagnosis of malaria. The Autoscope, said David Bell, director of the Global Health Technologies group at Global Good, combines a digital microscope with a laptop that runs diagnostic software able to identify malarial parasites. The Autoscope integrates a microscope with a computer that can read a stained slide and calculate the probability that any given object is a malaria parasite. Highly portable, it measures 15 × 7 in. Courtesy of Intellectual Ventures. Typically, diagnosis of malaria relies on microscopic examination of a blood sample by trained microscopists or on a Rapid Diagnostic Test (RDT) much like that for pregnancy, where changes in color define the presence of disease. RDTs, while inexpensive and easy to use, can only detect the presence or absence of malaria. They are less than ideal for cases of severe or drug-resistant malaria, where knowing the concentration of malaria parasites is important. Traditionally for that you need a microscope and someone who knows how to use it. In low-resource settings where there aren’t a lot of well-trained microscopists, automated diagnostic microscopy could make a critical difference in the case management of patients with malaria. The Autoscope’s software, said Bell, reads a routinely stained sample on a slide and calculates the probability that any given object is a malaria parasite. It can diagnose the presence of these parasites within minutes, without relying an assessment by an expert. Field tests in Thailand and Peru are showing high accuracy, without requiring health care workers to have microscopy training. The results, said Bell, indicate that the Autoscope is as reliable as most trained human microscopists — and in some cases even more so. “It’s not subject to human beings being not trained enough or being tired or stressed,” he said. “And it will give good consistency across laboratories and across time.” Because of this consistency, he added, the device will help researchers who study malaria around the globe, as well as helping health care providers in low-resource settings. The device is enclosed in a 15 × 17-in. box and is highly portable. The current prototype is attached to a laptop. But the final version, which is due out in 2018 when it will be in 20 clinics worldwide, will have the computer integrated into the device. The Autoscope is a benchtop device that requires electricity to run, so it can’t operate out in the field. But, said Bell, “It will go wherever microscopes are now, such as a small clinic.” One of Global Good’s strategies is to drive down the cost of technologies for low-resource settings by absorbing R&D expenses. So far, so good for the Autoscope. The next step, said Bell, is to find an industrial partner who can productize and make the device commercially available in low- and middle-income markets. As with portable microscopes that are connected to the web or make use of algorithms, microfluidic labs-on-chip circumvent the need for pathologists and allow for quick bedside diagnosis by nurses and other health care workers in remote areas. These miniaturized and disposable integrated laser imaging systems require just a drop of blood or other fluid to make an analysis via a computer or smartphone. They are cheap, fast and easy to use. Component developments On the component side, Hamamatsu, like many other players in the field, has developed new technologies that can enable this sort of point-of-care diagnosis. One is its MPPC silicon photomultiplier (SiPM), which is built around avalanche photodiode technology. “In simple terms, the SiPM can measure incredibly weak light signals, down to the single photon, in a device that is very compact, low cost and easy to manufacture,” said Richard Harvey, who is employed by Hamamatsu in the U.K. as the medical group leader. Hamamatsu’s high-sensitivity micro-spectrometer can be used to run multiple assays within a single instrument and are manufactured with high volume and low cost in mind. Courtesy of Hamamatsu. Because of these qualities and the detector’s simple low-voltage electronics, Harvey added, the technology is the detector of choice for many portable or handheld medical diagnostic systems. These systems, he said, use either discrete detectors or plug-and-play OEM modules that can further help push down costs because of rapid prototyping and simplified production assembly. Another enabling technology for PoC diagnostics is Hamamatsu’s micro-spectrometer, winner of a 2015 Prism Award. The high-sensitivity fingertip-sized spectrometers can run multiple assays on a single instrument and are manufactured in high volumes and with low costs in mind. The micro-spectrometer consists of a grating chip and CMOS image sensor chip facing each other across an air gap. It is fabricated with MOEMS technology and methods, to be integrated into equipment or connected to handheld mobile devices designed to perform spectral measurements such as PoC testing. It also is hermetically sealed, so it is usable in adverse conditions. “The SiPM and the micro-spectrometer have broken down some of the traditional barriers of low-resource settings,” said Harvey. “Both technologies represent a big shift in the cost of using photonics in the point-of-care sector. “Photonics itself is the key enabling technology,” he added. In Jena, Germany, Microfluidic ChipShop has developed a universal point-of-care platform that allows the system to run different kinds of diagnostic assays, including molecular assays, immunoassays and clinical chemistry assays. A universal point-of-care platform developed by Microfluidic ChipShop allows the system to run different kinds of diagnostic assays. So far the company has created cartridges for the diagnosis of (from left) tuberculosis, HIV and for the analysis of liver function. Courtesy of Microfluidic ChipShop. “The idea is — similar to electronics 60 years ago — to integrate a complete diagnostic or analytical workflow, which can contain main protocol steps into an integrated device,” said Holger Becker, Microfluidic ChipShop’s chief scientific officer. “This allows for a so-called ‘sample-in result-out’ operation, which means such a system does not require the skills and knowledge of a Ph.D.-level analytical chemist to operate.” By miniaturizing the dimensions and volumes, he added, these microfluidic devices significantly increase the speed an analysis takes and allows for an equally significant decrease in the required sample and reagent volumes. A test requires a drop of blood instead of a venous draw. This miniaturization, in turn, reduces costs and waste, and makes overall handling of a sample more convenient. So far, the company has created prototype cartridges for the diagnosis of tuberculosis and HIV and the analysis of liver function. The cases, said Becker, represent the different assay types the system can run. They also have clinical relevance. According to a 2011 paper from the World Economic Forum, for example, about 5 million people in the developing world die each year from tuberculosis and AIDS. A liver function test, Becker added, is widely used and can serve as an indicator for hepatitis C, which is a common co-infection with HIV. As with other PoC diagnostic systems, Becker said, the cartridges will generate diagnostic information “where the patient is — for example, bedside in hospital or doctor’s office, which means that the sample does not have to be transported and the result obtained typically within the time the patient sees a doctor and not days later.” For all the technical advances, costs remain an obstacle to the adoption of point-of-care devices in low-resource settings. Even a device costing just a few dollars can constitute a serious expense in the poorest regions of the world. Such systems, said Becker, “can only be rolled out in conjunction/support with national or international organizations.” In the case of Microfluidic ChipShop, this external support came from the Bill and Melinda Gates Foundation. The company’s first cartridge based on this platform came on the market late last year. “In general, the number of microfluidics-based PoC systems is rapidly increasing — there are at least a dozen systems out on the market already for a broad range of diseases,” said Becker. “This is a fast-moving field with many new technologies involved.” At the Ozcan Research Group at UCLA, researchers are looking into one of those new technologies: lens-free on-chip imaging for the diagnosis of many types of diseases, including cancer. The technology makes use of an LED or a laser diode for illumination and CMOS/CCD image sensors for image acquisition. A system being developed in the Ozcan Research Group uses computation to reconstruct the image of a pathology specimen that is captured by an optoelectronic sensor array. “Such a computational imaging approach does not require lenses to form an image, enabling its setup to be simple, compact and cost-effective,” said Yair Rivenson, a Marie Sklodowska-Curie Fellow in the group. This optical setup of a lens-free microscope makes use of a partially coherent light source and a CMOS image sensor. Courtesy of Yibo Zhang. The system, he added, also will allow for digital correction of the resulting hologram and 3D digital refocusing. In addition, the research group has recently developed a new numerical approach for the reconstruction of an object’s image, which reduces by at least half the number of measurements required to create a clinically relevant image. This reduction in turn cuts the cost of image creation and the amount of time it takes to measure it. Improving systems There are still challenges to be overcome before the imaging system can be adopted. “One of the obstacles is to provide rapid ‘preview mode,’” said Rivenson. “Today, when a technician or a doctor examines a pathology slide, he or she would like to see with their own eyes what they are imaging. When looking at holograms acquired by our system, the result is not clinician intelligible.” What’s needed is real-time or at least rapid reconstruction that can provide the clinician with an image preview. However, Rivenson added, graphics processing units that are commonly used in computers, even laptops, can provide more than an order of magnitude speedup in image reconstruction. Another challenge lies in making the diagnostic systems more compact and cost-effective. Both further advances of backend algorithms and the ever-improving quality of consumer-based image sensors will help push this hardware miniaturization. A 3D-printed attachment containing cost-effective optical and mechanical components can turn a smartphone into a portable microscope. Courtesy of Yibo Zhang. The challenge in the field as a whole, added Rivenson, is to create high-quality data inputs that can allow machine-learning algorithms to yield the most meaningful results. “This is where we as the engineers and scientists that design photonics-based PoC devices should shift our attention in creating sensing systems that will maximize the throughput and accuracy of diagnostics,” he said. Also at the Ozcan Lab, Yibo Zhang is looking into a combination of lens-free imaging and smartphone microscopy that can turn cellphones into imaging platforms. Zhang recently reported the results of attempts to combine the two techniques to achieve accurate-color imaging of pathology slides in low-resource settings. Since the lab achieved pathology slide imaging through the melding of these techniques a couple years ago, Zhang has focused on improving the imaging system along three avenues: color accuracy, the amount of time it takes to compute a result and the amount of data to capture a high-quality image. The ambitions reach beyond PoC diagnosis for the less-wealthy regions of the world. Down the road Zhang envisions “utilizing our unique technique to achieve more than what is achievable by a regular microscope and provide a better tool for diagnosis.” Back at MobileODT, Levitz is also seeking ways to deliver improved PoC diagnostics. As with the PoC research at Global Good, the Ozcan Research Group and elsewhere, the company is seeking to replace qualitative assessments with quantitative measures. Typically, clinicians using colposcopes assess color changes in patterns of blood vessels to identify suspicious sites. Multispectral imaging under development for the EVA System would deliver actual values of blood content at each pixel in the image. “We believe this increased accuracy will eventually result in better treatment outcomes,” Levitz said. “It will offer new information that is currently unavailable to clinicians outside of select university hospital centers that are doing academic research in this topic.