Low-cost, point-of-care biomarker quantification

Abstract

Low-cost, point-of-care (POC) devices that allow fast, on-site disease diagnosis could have a major global health impact, particularly if they can provide quantitative measurement of molecules indicative of a diseased state (biomarkers). Accurate quantification of biomarkers in patient samples is already challenging when research-grade, sophisticated equipment is available; it is even more difficult when constrained to simple, cost-effective POC platforms. Here, we summarize the main challenges to accurate, low-cost POC biomarker quantification. We also review recent efforts to develop and implement POC tools beyond qualitative readouts, and we conclude by identifying important future research directions.

Graphical Abstract

Introduction

Diagnostics are a vital part of our healthcare infrastructure, with substantial impacts on individual patient treatment as well as public health policy. The COVID-19 pandemic brought this fact sharply into focus for society at large: limited access to diagnostic testing, especially in the beginning of the pandemic, precluded early diagnosis and isolation of infected people, likely costing thousands of lives. Even as testing capacity began to be built up, it was for many months restricted to costly, time-consuming laboratory tests with availability that often struggled to meet demand. Early access to minimal-equipment approaches and devices for the population to determine their COVID status at a doctor’s office or at home by detection of disease biomarkers—that is, point-of-care (POC) diagnostics—could have allowed better-informed individual health and behavior decisions, with wide-ranging public health impacts. These impacts would be further amplified in resource-limited areas where clinical laboratories and sophisticated analytical instrumentation are not readily available and thus little knowledge of disease spread is possible.

Similar to COVID-19, the diagnosis and monitoring of many diseases rely on the detection of biomarkers, which are substances or molecules present in our body that are indicative of a diseased state. In clinical settings, biomarkers are typically detected via complicated methods that require expensive equipment operated by highly trained personnel, all infeasible in low-resource settings and thus creating a need for the development of low-cost POC devices. The gold standard for low-cost, deployable POC devices is widely viewed to be defined by the World Health Organization’s ASSURED criteria[1]: Affordable, Sensitive, Specific, User-friendly, Rapid and Robust, Equipment-free and Deliverable to end-users. Critically, however, the diagnosis of many diseases depends not just on the presence or absence of a biomarker, but on its quantitative level; measuring these quantitative levels is a major challenge for low-cost POC assays.

A few POC technologies have seen broad use in clinical and public health applications around the world (Table 1). Perhaps the most prominent of these technologies is the lateral flow test (LFT), with the prototypical example being a home pregnancy test. LFTs have also been implemented in the detection of soluble transferrin receptors[2] and alpha-1-acid glycoprotein[3] to diagnose iron deficiency and inflammation, respectively. In their most common form, LFTs contain a capture reagent, such as an antibody[4–6] or aptamer[7], bound to a cellulose membrane. Biological samples are added to the membrane, flow through it via capillary action, and are ultimately captured at a specific line and visualized via a detection reagent, often colloidal gold nanoparticles[6] or colored latex spheres[5]. While this is probably the most widely-used POC platform, it has significant limitations. Most notably, LFTs are strictly qualitative, yielding only presence-or-absence results. While such binary outputs are typically sufficient for health conditions that are characterized by significant and acute biomarker buildup (e.g., pregnancy) or by the mere presence of the biomarker (e.g., viral infections) and can also be used as preliminary screening tests before a patient is referred to a clinic, many diseases and medical conditions require biomarker quantification for diagnosis, precluding the use of binary output LFTs. Achieving low-cost, reliable, sensitive POC quantification has remained an elusive goal in the field. In this article, we review recent advances and the continued challenges for POC biomarker quantification, with a focus on developments in the past three years.

Table 1.

Commonly used POC devices. While lateral flow tests (LFTs) and microfluidic paper-based devices (μPADs) are fast and inexpensive, they provide only presence/absence biomarker readouts. Glucose monitors, although more expensive, provide quantitative measurements.

Device	Advantages	Disadvantages
Lateral Flow Test (LFT)	Fast Inexpensive Stable Versatile	Binary output Poor sensitivity
Microfluidic Paper-based Device (μPAD)	Fast Inexpensive Very small sample volume	Binary output Sample can evaporate Non-uniform sample distribution
Glucose Monitor	Numerical output Minimally invasive	Expensive test strips

Challenges to developing quantitative POC diagnostics

Biomarkers are often found in complex biofluids such as blood, urine, and saliva that contain many small molecules, ions, and proteins—a characteristic that makes them valuable sources of biomarkers but also poses one of the most serious challenges to accurate quantification. Complexity and variability in biofluid composition create confounding “matrix effects”[8–10], where the readout of the analyte being measured is affected by the levels of other analytes, even in research-grade equipment. These matrix effects can ultimately limit diagnostic accuracy. Recently, a generalizable parallel calibration scheme[11] that accounts for sample matrix variability was described in the context of a low-cost nutritional diagnostic: in this scheme, a calibration curve is created from standard assays run in the patient’s sample, such that both standard and test assays have the same sample matrix. Similar sample-specific calibration schemes will likely be necessary to enable other devices and assays to yield quantitative results in complex biofluids.

Beyond matrix effects, another prominent challenge in POC diagnostic development is in creating high-compliance, user-friendly devices. One way to meet this goal is to decrease or eliminate the invasiveness of biofluid sample collection. Promisingly, electronic sensors have been incorporated into wearable formats, but these systems are primarily limited to the detection of small molecules[12,13]. However, recent progress has been made in incorporating cell-based sensor mechanisms into wearable materials via lyophilization. For example, a CRISPR-based sensor was coupled with a face mask, allowing for the quantification of nucleic acids in exhaled breath[14]. Such approaches had previously been hindered by difficulty in maintaining cell viability in wearable configurations.

However, since non-invasive detection of biomarkers is not always possible, minimization of required assay volumes is another common goal of POC diagnostic development. Using smaller samples can reduce invasiveness (requiring only a finger-stick blood sample rather than a venous blood draw, for example) and simplify sample processing. While conducive to patient adherence, minimal sample volume is an obstacle to accurate quantification, as blood droplets are more susceptible to variation in composition than the milliliters of blood used in clinical laboratory diagnostics, potentially leading to less accurate measurements[15,16]. Even when biomarker readouts are accurate, using a small volume of blood can come at the expense of the limit of detection (LOD) of the assay[17].

The LOD of an assay is a crucial figure of merit in the development of quantitative, low-cost POC diagnostics. Assays must be tuned to enable detection of potentially low levels of biomarker concentrations to limit false-negative results, but must be sufficiently robust to avoid returning false-positive results. Since physiologically relevant levels of biomarkers are typically quite low, detecting them is difficult enough when sophisticated instrumentation is used; it poses an even greater hurdle to simpler, point-of-use devices. Several strategies have been implemented to reduce the LOD of POC devices. For example, to improve the sensitivity of RNA-based biosensors, RNA amplification steps have been integrated into workflows. Although isothermal amplification methods such as loop mediated isothermal amplification[18] and recombinase polymerase amplification (RPA)[19] can in principle be used in low-resource settings since they don’t require a thermocycler, they add significant complexity to the assay, hinder accurate analyte quantification, and do not have specificity for single base pair differences. To increase the specificity of amplification-based detecting, CRISPR-Cas13a has been coupled with RPA[20,21], enabling base pair-level specificity in the detection and (to some extent) quantification of nucleic acid sequences.

One approach that has been developed to try to improve the LODs of LFTs is the microfluidic paper-based analytical device (μPAD)[22]. μPADs use microfluidics and capillary action in paper substrates to channel biological samples, allowing for directed and controlled flow without use of pumps or valves. This flow control can be used to decrease the flow rate of the sample through the detection zone, allowing for increased interactions between the biomarker and capture reagent, resulting in lower LODs[23] when compared to traditional LFTs. The controlled flow also opens the door for multiplexed tests and integrated sample preparation[24] via the use of separate channels to ensure that samples and reagents interact in designated locations on the μPAD. Additionally, μPADs are inexpensive, easy to use, and environmentally friendly. As a result, they have already been applied in multiple contexts for detecting biomarkers, including for traumatic brain injury[25] and acute myocardial infarction[26]. Although μPADs have some advantages compared to traditional LFTs, they still encounter many of the same challenges, including sensitivity, specificity, and low-resource, accurate quantification[27].

To try to address quantification challenges for low-cost and minimal-equipment POC diagnostics, semi-quantitative[28] and colorimetric readouts[11,29,30] have been explored. A common reporter used for this purpose is the enzyme β-galactosidase, which can cleave a yellow molecule (CPRG) to a purple product (CPR). This reaction has multiple distinguishable intermediate colors visible to the naked eye, allowing for equipment-free visual interpretation of test results. While color perception can be somewhat subjective, coupling these approaches with POC quantification devices such as smartphones can further enable quantification.

Emerging approaches towards accurate POC biomarker quantification

Smartphones have great potential as a large-scale enabler of quantitative POC diagnostics. Smartphones have become ubiquitous in society even in low-income populations and in low-resource areas, thus representing a piece of “equipment” that can be assumed to be available for many doctors or patients. Their multitude of functions—notably including imaging, communications, and data processing—can be leveraged to avoid the need for research grade equipment, making quantification in the field more feasible. For example, smartphones have been coupled with LFTs to enable quantification of alpha-1-acid glycoprotein[3] and ferritin[31], and coupled with an optical device to measure the myocardial infarction biomarker cTnI[32].

Cell-free expression (CFE) systems have also emerged as a promising platform for the development of POC diagnostics. CFE systems use the natural sensing machinery of cells, such as riboswitches[33] and transcription factors[34–37], combined with its expression machinery to enable detection and reporting of analyte concentrations. They are inexpensive, easy to use, portable, and can be lyophilized and rehydrated on demand, much more so than their whole-cell biosensor counterparts[35]. Cell-free biosensors have been used to detect a variety of analytes, including viruses, water contaminants, vitamins, and other essential small molecules indicative of disease[33,34,36–38]. Biomarker detection can be robust even in serum[11,39] though the addition of RNase inhibitors to prevent mRNA and tRNA degradation by RNases in the biofluid is typically necessary. Enriching CFE systems with RNase inhibitors and their folding chaperones can avoid the costs associated with adding exogenous inhibitors, but it requires changes in cell growth conditions that reduce system productivity[40]. There is, therefore, room to improve these implementation approaches as the diagnostic use of CFE systems expands.

Notably, however, CFE systems still face the same quantification challenges as other low-cost POC devices. To meet the need for biomarker quantification, CFE systems have been coupled with portable electronic devices to enable simple quantification[34,41,42], as well as with the CPRG reporter molecule described above. Recently, a cell-free zinc biosensor was coupled with a glucose monitor[43]—a robust, well-established, and widely-used POC device capable of providing numerical outputs—as proof of principle for a modular, digital platform for quantitative cell-free biosensing.

Future directions

The modularity that can be achieved, for example, by coupling a cell-free biosensor with a glucose monitor is a key feature that will be central to the future large-scale success of quantitative POC devices. Most POC diagnostics to date leverage some target-specific property or characteristic that is not easily translatable to detecting different analytes. Significant time and funds are spent developing devices that, in the end, have only one specific application. Even the development of new LFTs with different antibodies takes significant investment and expertise. A platform that could be easily modified for new biomarkers (Figure 1A), where general detection machinery is paired with an analyte-specific attachment or cartridge, would be of great value. Such an approach would simplify assay development and could even increase patient compliance[44,45] by providing one easy-to-operate tool for multiple measurements. However, achieving accurate quantification for each target analyte within the clinically relevant concentration range in such a modular platform is likely to be a significant challenge[21,44–46].

Figure 1.

Promising platforms for POC diagnostics. Adapted from Voyvodic et al.[51] (A) Modular sensing platforms. A single output module can be combined with different, biomarker-specific detection modules. (B) Multiplex assays. A single detection module can be used to detect different biomarkers within the patient sample.

Another promising frontier in POC diagnostics is using one device to measure different biomarkers from a single sample, in “multiplex assays” (Figure 1B). Many diseases have multiple biomarkers, and often a patient may be tested for multiple potential diagnoses using different biomarkers when the patient’s symptoms are non-specific. Multiplex assays have the potential to accomplish these tasks while minimizing the total volume of sample being used, but they are difficult to develop and implement, especially when constrained to ASSURED criteria. Recently, a cell-free platform for multiplexing biosensor reactions in an aqueous two-phase scheme was reported[47], where the biosensors are partitioned in separate spatially separated compartments immiscible with the bulk fluid sample. This spatial separation strategy is a promising one for enabling accurate quantification in multiplex assays to minimize crosstalk and resultant false-positive results. LFTs can move towards accomplishing this goal by segregating reaction “zones” along the test strip[48] or by implementing a “multilabel” approach for each analyte[46,49]. These strategies help preserve selectivity and could potentially even be used in conjunction with parallel calibration approaches to enable analyte quantification, but they do increase the complexity and cost of the detection platform. Microfluidic devices can overcome some of these issues and thus may serve as an alternative platform for multiplex assays if they are sufficiently inexpensive, robust, and field-deployable[50].

One final factor impacting both quantitative and qualitative POC diagnostics and thus warranting significant further effort is the development of sample collection and processing approaches. Tests must include easy-to-follow steps to prevent user error (e.g., collection of inadequate sample volume) that could result in false negative results. Addition of a positive control to the testing framework can help identify errors in sample collection and thus minimize false negative results[52]. Following sample collection, sample processing such as removing red blood cells from whole blood or lysing immune cells to measure their contents typically requires extensive steps, expensive equipment such as centrifuges, or both. POC diagnostics would benefit from additional efforts to develop easy-to-operate, low-technology equipment to complement the testing pipeline, including handheld centrifuges[53] and other portable materials. However, it remains critical that these low-technology processing approaches yield consistent and robust results, otherwise biomarker quantification will remain elusive.

Many challenges remain in the implementation of ASSURED POC diagnostics. While progress has been made in robustness to matrix effects and broader use of equipment-free approaches, accurate quantification remains a primary challenge even when sophisticated analytical equipment is available (let alone in POC contexts). Ultimately, POC devices cannot become a ubiquitous, accessible standard for diagnostics until simple yet accurate quantification methods are more widely integrated into POC platforms.