High-sensitivity, imaging-based immunoassay analysis for mobile applications
Rapid medical diagnostics (RMD) and point-of-care testing (POCT) offer the benefits of convenience, simplicity, and low cost. These tools are increasingly in demand in the developed world, where preliminary screenings can relieve the burden on healthcare systems. In developing regions and disease-endemic countries, RMD and POCT are not only ancillary alternatives to clinical examination and laboratory testing, but are also necessary solutions to expedite response to outbreaks, improving treatment outcomes. The 2013–2015 West African Ebola outbreak demonstrated the importance of rapid POCT technologies.1, 2
Immunoassay-based rapid diagnostic tests (RDTs) are key components in some POCT applications, having the advantages of fast turnaround times, low cost, and simplicity. Lateral flow- and flow-through immunoassays (LFIs and FTIs, respectively), rely on biochemistry to measure the presence (or absence) and concentration of a specific marker for infection (Ebola, HIV, or malaria, for example), and/or a condition (pregnancy, drug abuse, or cardiac problems). Forming colored lines and/or spots on the assay substrates, the target markers (such as proteins, molecules, or other analytes) are labeled with latex and colloidal gold particles, carbon, liposomes, or fluorescent labels.3 The efficacy of the immunoassay technology depends on the accurate and sensitive interpretation of these macroscopic spatial features. Ideally, an interpreter (a reader) should conduct objective and repeatable measurement of the contrast/color signal generated by the immunoassay. However, immunoassays come in various formats, and their instrumentation includes associated hardware (mechanical and optical) and software modules (user interface and algorithm). These often require fundamental modification and customization to address the technology's evolving needs.
To address these requirements for accurate and sensitive immunoassay applications, we introduced a novel reader platform.4, 5 This offers an imaging-based analysis toolset to improve the performance of the immunoassays by increasing the limit of detection while minimally affecting the coefficient of variation (CV) of measurements. This toolset includes a smartphone-based reader for data acquisition and interpretation (see Figure 1), test developer software (TDS) for reader configuration and calibration, and a cloud database for spatiotemporal tracking of testing results. With its custom-developed universal mechanical interface and image-processing algorithm, this universal reader platform can work with any LFI or FTI, achieving high sensitivity and accuracy beyond the ability of human vision and other reader solutions.
The mechanical interface of the reader unit provides the flexibility to accommodate several assay types and formats with varying physical dimensions, as shown in Figure 1(c, d). The user slides the assay into the mechanical holder and closes the reader tray for the measurement. To complement this mechanical universality, the user (or test manufacturer/developer) can configure and calibrate the reader to work with new assay types using TDS (see Figure 2). End-users can define the specifications of their assays under development (qualitative or quantitative/semi-quantitative, chromatographic or fluorescent, reflection or transmission, direct or competitive) and create new test types for the reader application. After this one-time new test implementation, users can run the manufacturing lot calibration and start using the reader units as stand-alone analyzers to interpret the assay results, presented in calibrated or raw units (concentration or arbitrary pixel intensity, for example). The user determines the output values (multivariate) and labels (yes/no, positive/negative, negative/borderline/positive) as a result of assay analysis in TDS, and delivers a single final configuration file to the reader unit(s) through the cloud interface.
Once the assay is loaded, the smartphone application communicates with the reader hardware to switch on the appropriate illumination scheme (fluorescent, reflection, transmission) based on the test/assay type selected. Running the digital image processing steps, the reader application acquires the optically enhanced and filtered image of the assay, processed in real time for the final decision. The application represents the results of the assay analysis in the format determined by the user in test configuration. The smartphone application controls the entire operation (see Figure 3).
The reader platform offers a universal tool for imaging and digital analysis of assays, but also complements the assay performance by increasing the sensitivity (limit of detection) and accuracy (low CV) of measurements. To achieve these, the optoelectronic hardware design and the image processing algorithm—with dynamic feature detection and background estimation—play critical roles in the analysis performance. The transmission and reflection imaging modes enable chromatographic and colorimetric analysis, while the fluorescent imaging mode makes it possible to evaluate fluorescent immunoassays.
Immuno-chromatographic lateral flow and flow-through devices are among the most widely used immunoassays. However, their performance often suffers from very weak signal levels, limiting their use in POCT applications. Our transmission mode imaging approach overcomes this limitation, achieving ‘trans-visual’ sensitivity and accuracy. We place an illumination array to the back of the immunoassay, such that the narrow-band LED illumination passes through the assay membrane, and then we acquire and process the image of the assay using the smartphone application. This brings an important advantage when compared to the other approaches. Most lateral flow and flow-through technologies use semitransparent material (nitrocellulose membranes, for example) as a substrate, dispensing the chemical components into the membrane. The significant penetration depth in the highly porous substrates6 means that most of the information (contrast or signal, for example) is actually embedded inside immunoassays after the assay activation is completed. Therefore, our transmission mode analysis may detect signals many orders of magnitude greater than for reflection-mode imaging and scanning-based systems, or visual inspection by individuals. Figure 4 shows our experimental results.
In summary, we have developed an imaging-based immunoassay analysis toolset that includes stand-alone smart-phone-based reader units, TDS, and cloud services. The reader units can digitally analyze a variety of assays without requiring modification of mechanical and optical interfaces or software customizations. Providing trans-visual sensitivity with very low instrument CV (down to <1%), it enables a digital imaging system but also a universal solution that can improve the sensitivity of immunoassays.
In the near future, we are planning to enable the analysis of any immunoassay type and achieve even higher reading performance with new image processing algorithms and software under development. Moreover, we are working on innovative hardware architectures to further enhance the user experience and address the requirements of the immunoassay market.
Holomic LLC gratefully acknowledges the support of the Department of Defense (US Army Research Office, ARO, Small Business Innovation Research, contract W911NF-14-C-0077).
Onur Mudanyali is director of R&D and engineering, responsible for research and development of photonics-based medical imaging, sensing, and diagnostic technologies that can address global health challenges. Previously, he was a research assistant and teaching fellow at UCLA's Electrical Engineering Department, where he received his PhD in 2013.
Justin White is the principal software engineer. He received his BSc in engineering from Harvey Mudd College and his MSc in electrical engineering from UCLA. He has more than five years of professional and research experience in the fields of image processing and computer vision.
Chieh-I Chen is a software engineer pursuing an MSc at the Department of Computer Science at the University of Southern California. He received a BA from the Department of Computer Science at National Tsing-Hua University, Taiwan. He is interested in smart computational solutions to global health problems, computer vision, and biomedical device engineering.
Neven Karlovac is CEO of Holomic LLC. He has more than 30 years' experience in technology and business management for companies ranging from early-stage startups to Global 500 corporations. Previously, he was a senior technology strategist with the UCLA Institute for Technology Advancement, a privately funded organization with a mission to foster and accelerate commercialization of UCLA research.