INTRODUCTION
While coronavirus disease 2019 (COVID-19) is not the first pandemic of the 21st century (
1), it has generated unprecedented global concern and responses. COVID-19, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is thought to have emerged from a zoonotic source (
2) and spread rapidly in humans through respiratory droplets and contact. There is some concern for airborne transmission, but the role of this transmission route outside the potential aerosolizing procedures in health care settings is unclear (
3–5). With an estimated reproductive number,
R naught (
R0), of between 1.4 and 5.6, SARS-CoV-2 rapidly spread worldwide (
6–9). Since the first cases reported in December 2019 (
10–12), there have been over 106 million confirmed cases and 2.3 million deaths reported worldwide (as of 9 February 2021) (
13).
From a disease manifestation perspective, SARS-CoV-2 infection can be asymptomatic (
14), and COVID-19 spans from a mild influenza-like illness (ILI) to life-threatening complications (
15,
16). SARS-CoV-2 not only affects the respiratory tract, resulting in pneumonia, but also can affect gastrointestinal (GI), neurological, or cardiovascular systems. Atypical presentations of COVID-19 include cutaneous manifestations such as a Kawasaki-like disease in children and ophthalmic/gustatory dysfunction (i.e., anosmia and ageusia, which are the loss of smell and taste, respectively), which may have been underestimated in initial reports (
17–20).
Despite numerous therapeutic options being explored (e.g., convalescent-phase plasma), no large-scale treatments are available. Public health interventions have evolved over time to limit viral spread (
Fig. 1) and have included the use of personal protective equipment (PPE) like masks, handwashing, and containment measures such as city lockdowns, travel restrictions, and physical distancing (
21–30). Although these strategies have been essential to reduce the virus’s spread, they have had significant adverse socioeconomic impacts, and adherence to these prevention strategies is challenging to sustain (
22). Currently, cases of COVID-19 have declined following a first pandemic wave in some areas, whereas other areas are experiencing subsequent waves of activity. Fortunately, many vaccine candidates are under development and undergoing regulatory approval processes (
31–35). Recently, COVID-19 mRNA vaccines have been the first licensed for use and are rapidly being administered as supplies are provided (
28,
36). However, given the time required for adequate immunization coverage in the population at large, subsequent pandemic waves are anticipated (
31,
37–39). Therefore, detection methods for SARS-CoV-2 remain a crucial part of containment and mitigation strategies, and lessons learned from this pandemic may help prepare against future pandemics.
In terms of testing, real-time reverse transcription-PCR (RT-PCR) remains the most common method used to identify SARS-CoV-2 (
40). While common in diagnostic laboratories worldwide, many laboratories remain faced with supply chain shortages for real-time RT-PCR reagents and consumables, all while being asked to increase testing capacity. As such, delays were common for test results, prompting the exploration of alternative testing options such as specimen pooling or laboratory testing using methods other than RT-PCR. Methods that could enhance testing capacity, streamline testing (i.e., automation), or provide more rapid results in easy-to-use formats that are amenable to point-of-care (POC) applications without complex instrumentation (e.g., isothermal technologies) were all desired (
41–47). Rigorous research escalated quickly from the academic to industry partners, and this research is ongoing to develop testing alternatives or complements to existing technologies.
While recent reviews have been published on the management of SARS-CoV-2 (
41,
47–55), recent advancements in novel diagnostic methods justify the need for a more comprehensive synthesis of the current literature. In this review, first, the biological characteristics of SARS-CoV-2 are described in order to fully understand the molecular and immunological methods for its detection. Following a brief discussion on the COVID-19 manifestations, compatible signs and symptoms, and disease biomarkers, diagnostic imaging techniques are described in relation to COVID-19 lower respiratory tract involvement, including applications such as monitoring disease severity, the progression of the illness, or complications. Next, a comprehensive review of current and recent advances in molecular, antigen (Ag), and serological immunodiagnostic methods is covered, including rapid diagnostic tests (RDTs) used in the laboratory setting and POC applications. Overall, this review expands our knowledge of current and exploratory avenues for detecting SARS-CoV-2 and COVID-19.
It should be noted that some of the references used in this review were preprints that have not been peer reviewed, and recognizing that data on the detection of SARS-CoV-2 or COVID-19 are rapidly evolving, some details on testing options and guidelines may no longer be recent and should thus be reviewed in the context of recent findings and recommendations. Nonetheless, this review provides a comprehensive synthesis of the most current data available to date, along with current recommendations for the detection of SARS-CoV-2 or the diagnosis of COVID-19.
SARS-CoV-2 GENOME AND STRUCTURE
Understanding the genetic and structural properties of SARS-CoV-2 is a prerequisite to developing effective diagnostic tools. SARS-CoV-2 was first isolated and sequenced in China in January 2020 (
10–12). Transmission electron microscopy revealed that SARS-CoV-2 has a diameter in the range of 60 to 140 nm, and its morphology was consistent with those of other members of the
Coronaviridae family (
Fig. 2A) (
12,
25). SARS-CoV-2 is an enveloped, positive-strand RNA virus, and on the genetic level, it shares 96%, 80%, and 50% sequence identities with bat coronavirus (RaTG13), SARS-CoV-1, and Middle East respiratory syndrome coronavirus (MERS-CoV), respectively (
11,
56). Based on these analyses, the International Committee on Taxonomy of Viruses named the virus SARS-CoV-2, which was formerly referred to as the 2019 novel coronavirus (2019-nCoV) or human coronavirus 2019 (hCoV-19) (
25).
Our understanding of SARS-CoV-2 structure and function has been largely derived from research on SARS-CoV-1, MERS-CoV, and seasonal coronaviruses. SARS-CoV-2 has a single-stranded positive-sense RNA genome of between 26 and 35 kb, encoding approximately 27 proteins with similarity to proteins of known functions, while others are unclear/unknown or putative (
Fig. 2B) (
21,
37,
53,
57,
58). The first open reading frame (ORF1a/b) on the 5′ end of the viral genome occupies ∼71% of the entire genome and produces two polyproteins (pp’s), pp1a and pp1ab. These two polyproteins are processed by the viral proteases into 15 nonstructural proteins (nsp’s), and these proteins are collectively involved in polyprotein processing, viral RNA replication, and mRNA synthesis (
53,
57). The remaining proteins, including the structural and accessory proteins, are expressed from several nested subgenomic mRNAs produced through a process known as discontinuous transcription by the viral RNA-dependent RNA polymerase (RdRp).
The structural proteins include the small envelope (E) protein, membrane (M) protein (also known as the matrix protein), nucleocapsid (N) protein, hemagglutinin-esterase (HE) protein, and spike (S) glycoprotein (
Fig. 2A) (
57,
59). The E and M proteins are primarily involved in viral assembly, budding, and virion morphogenesis (
60–62), while the N protein complexes with the viral genomic RNA to generate the nucleocapsid (
63). The S protein is the major surface glycoprotein on SARS-CoV-2, forming approximately 40 trimers that play an important role in both receptor binding and membrane fusion through the two functional subunits S1 and S2 (
37,
64). The S protein trimers contain a stable stalk separated from the globular heads by three flexible hinges, allowing for orientation freedom to interact with host cell receptors (
65). The S1 subunit contains the receptor-binding domain (RBD) that directly interacts with the angiotensin-converting enzyme 2 (ACE2) receptor on the host cell surface, whereas the S2 subunit contains a structural loop responsible for fusion events between the viral and host cell membranes, resulting in the release of the viral genomic RNA into the cytoplasm (
66,
67). Of note, along with engaging the ACE2 host cell receptor, the cellular serine protease TMPRSS2 is engaged for S protein priming, and this cofactor has been investigated as a possible antiviral target using viral entry inhibitors (
68,
69).
Overall, having knowledge of SARS-CoV-2 pathogenesis can help in understanding disease manifestations and help guide the development of molecular and immunological tools for the identification of this virus.
CLINICAL MANIFESTATIONS OF COVID-19
The spectrum of SARS-CoV-2 infection can vary from asymptomatic infection to life-threatening complications of COVID-19 (
37). Using modeling, it was estimated that over 59% of transmissions arise from asymptomatic individuals, with 35% from individuals in presymptomatic stages of infection and 24% from individuals who never develop symptoms (
70). These estimates are concerning but emphasize the need for the wide use of vaccines and maintaining key public health interventions like mask wearing, hand hygiene, and social distancing.
In most symptomatic cases, COVID-19 presents as a mild to moderate upper respiratory illness, with signs and symptoms compatible with those of other respiratory viruses (
71). As such, the diagnostic accuracy of any individual sign or symptom is very poor, and neither the presence nor the absence of any sign or symptom can be used to rule in or out COVID-19 (
71). With the possibility of other pathogens that could present like SARS-CoV-2 infection, case definitions based on clinical presentation are not sufficiently specific but can help support the investigation of suspect COVID-19 cases. On the other hand, given that the list of possible presentations and atypical manifestations of COVID-19 could mirror those of other diseases, identifying the etiology of illness as SARS-CoV-2 requires laboratory testing.
In a recent Cochrane review, a summary of 16 studies (7,706 patients) was presented (
71). Only six of the possible signs and symptoms of COVID-19 had sensitivities of >50%, and results were highly variable between studies and settings. The most common signs and symptoms (and their performances) are summarized as follows: cough (with sensitivity and specificity from 43 to 71% and 14 to 54%, respectively), sore throat (5 to 71% and 55 to 80%), fever (7 to 91% and 16% to 94%), musculoskeletal symptoms (e.g., arthralgias or myalgias) (19 to 86% and 45 to 91%), fatigue (10 to 57% and 60 to 94%), and headache (3 to 71% and 78 to 98%) (
71). It was noted that possible confounders were present, and the high heterogeneity between data suggested that signs and symptoms are variable between individuals (
71). Other less common clinical presentations have been documented, including alterations in smell or taste (i.e., anosmia or dysgeusia) as well as neurological or cutaneous manifestations (
17–19,
72–76). It is noteworthy that in the early stages of the pandemic, some of these symptoms may have been missed or underreported, but knowledge on possible clinical presentations of COVID-19 have evolved over time.
In some cases of COVID-19, progression to lower respiratory tract illness (e.g., pneumonia) can occur and may require hospitalization, intensive care unit (ICU) support, and mechanical ventilation, and complications can arise, which include acute respiratory distress, multiorgan dysfunction, and death (
71,
77–85). In general, adverse outcomes and deaths are more common with increasing age or in individuals with underlying medical comorbidities such as respiratory system disease, cardiovascular disease, and diabetes (
78–80). Fatality rates vary among studies and countries but are generally high in the hospital setting (e.g., 4 to 11%) compared to the overall case fatality rates (e.g., 2 to 3%) in the general population (
80,
82,
85,
86). In terms of recovery, the median duration of hospital stay is 10 to 14 days, and resolution generally occurs within 2 to 3 weeks (
85). There is a lack of evidence on whether some symptoms can persist after recovery. In one study, patients were monitored up to 60 days after recovery, with 87.4% reporting at least one symptom (
86). The most common symptoms were fatigue (53.1%), dyspnea (43.4%), joint pain (27.3%), and chest pain (21.7%).
Overall, while some signs or symptoms may be compatible with COVID-19, none are specific, and laboratory testing is required to confirm the diagnosis. Further studies are required to help identify the frequency of atypical clinical presentations, and additional studies looking at known clinical presentations of COVID-19 should consider possible confounders such as the possibility of other etiologies, host factors (e.g., comorbidities), disease severity, and the times from infection and symptom onset.
BIOMARKERS FOR COVID-19 AND ROUTINE LABORATORY INVESTIGATIONS
Apart from laboratory tests specific for detecting SARS-CoV-2 discussed throughout this review, routine laboratory testing spanning hematological, biochemical, and chemical markers is used to assess a patient’s health or identify possible clues to a disease state (
87–90). Such routine laboratory workup of individuals is used to refine a medical differential diagnosis, thereby supporting or refuting potential causes of the clinical presentation based on typical outcomes of these investigations for a defined disease. Many of these investigations can evolve through the clinical course of illness, and additional testing can be ordered by physicians based on the clinical presentation. These can include tests such as white blood cell (WBC) counts, markers for inflammatory conditions (C-reactive protein [CRP], procalcitonin [PCT], or interleukin 6 [IL-6]), tests for anticoagulation, and indicators of tissue damage (alanine aminotransferase [ALT], aspartate aminotransferase [AST], lactate dehydrogenase [LDH], and creatine kinase [CK]). While biomarkers for COVID-19 have been the subject of much investigation during the current pandemic, none of these tests are sensitive or specific for COVID-19. In a Cochrane review analyzing 67 laboratory tests from 21 studies encompassing 14,126 COVID-19 cases and 56,585 non-COVID-19 cases, only three markers showed sensitivity and specificity values of >50%: a decrease in the lymphocyte count and increases in the inflammatory markers CRP and IL-6 (
90). Overall, no individual biomarker can be used reliably to rule COVID-19 in or out, and laboratory testing should be performed. However, it should be noted that some laboratory markers have value for patient management as they can help assess the severity of the disease or progression of the illness or even act as risk factors for death. In the most recent Centers for Disease Control and Prevention (CDC) guidance documents for clinicians caring for patients with COVID-19, a summary of important laboratory tests is described, with lymphopenia being the most common laboratory finding in patients hospitalized with COVID-19 (
87). Laboratory markers associated with increased illness severity include lymphopenia, neutropenia, and elevated serum ALT, AST, LDH, CRP, and ferritin (
87,
88). Patients with critical illness have high plasma levels of inflammatory makers, and elevated levels of d-dimer and lymphopenia have been associated with an increased risk of death.
Of note, this section is not intended to be a comprehensive review of all biomarkers used in routine or exploratory investigations for COVID-19. We recognize the availability of guidelines for clinicians caring for patients with suspected or confirmed infection with SARS-CoV-2 (
87,
88) as well as the expertise of medical staff in ordering laboratory tests to help guide evolving differential diagnoses throughout the clinical course of illness. However, this section also recognizes the ongoing efforts of researchers who are dedicated to understanding the role of existing or novel biomarkers. Overall, no laboratory marker to date is diagnostic for COVID-19, but they have value in patient management over time, regardless of SARS-CoV-2 infection status. Biomarkers for COVID-19 severity or prognosis remain an active area of research that may not only lead to new diagnostic approaches but also help us understand disease progression and host responses to COVID-19 (
91–94).
DIAGNOSTIC IMAGING FOR COVID-19
While testing of specimens collected from the upper respiratory tract is common for diagnosing SARS-CoV-2 infection, the progression of the disease may involve the lower respiratory tract (e.g., pneumonia), with or without detectable SARS-CoV-2 in the upper respiratory tract (
55,
95–103). Testing of specimens from the lower respiratory tract (e.g., bronchoalveolar lavage [BAL] fluid) is possible using nucleic acid amplification tests (NAATs) like RT-PCR, but obtaining lower respiratory tract specimens is not always possible (
104–107). Along with laboratory testing, diagnostic imaging can complement investigations of COVID-19 to assess the involvement of disease in the lower respiratory tract or other anatomical sites. Diagnostic imaging techniques include chest radiography (or chest X ray [CXR]), computed tomography (CT) scan, ultrasound, magnetic resonance imaging (MRI), and positron emission tomography-CT (PET/CT) (
108–116). Among these, CT scans are the most frequently used methods for diagnosis of lower tract involvement or follow-up of COVID-19 cases (
110–112). CT scans produce cross-sectional images at different angles, thereby providing a three-dimensional (3D) look at the targeted anatomy. Chest CT scan images can be assembled and assessed by radiologists to check for possible abnormalities suggestive of lower tract disease such as viral pneumonia (
53,
112,
117). Typical features of a chest CT image in COVID-19 are ground-glass or reticular opacities (GGOs) with or without consolidations that present bilaterally, peripherally, or in posterior distributions (
113).
The utility of diagnostic imaging for routine screening for COVID-19 has been a subject of debate and has not been recommended by most radiology societies (
113,
114,
118–121). On the other hand, due to the shortage of RT-PCR supplies during the early days of the pandemic and the possibility of false-negative RT-PCR results from sampling the upper respiratory tract, some hospitals in the Hubei province of China included CT scans in the diagnosis of SARS-CoV-2 infection (
53,
117,
122,
123). While diagnostic imaging techniques like CT have merits to help assess lower respiratory tract disease involvement, to monitor disease progression, or to investigate other complications of COVID-19, it should be noted that diagnostic imaging methods are less sensitive than sampling the lower respiratory tract and testing using molecular methods, and specificity is low, given that typical features of COVID-19 are common to other respiratory viruses or illnesses (
113–116,
124–128). Initial reports of the utility of CT scans in the diagnosis of COVID-19 suggested an increased sensitivity of CT over real-time RT-PCR, but others have suggested explanations for the disparities between RT-PCR results and diagnostic imaging assessments, including poor sampling techniques, differences in the performances of testing methods, the anatomical site of RT-PCR testing (upper versus lower tract), and disease prevalence (
111,
124,
126,
129–134). High sensitivities (i.e., >90%) have been reported for CT scans in high-prevalence populations, while low sensitivities (<60%) were reported in studies with low-prevalence populations (
112–114,
118–121,
123). In a Cochrane review for confirmed cases of COVID-19, the pooled sensitivities were 93.1% (95% confidence interval [CI], 90.2% to 95.0%) for chest CT and 82.1% (95% CI, 62.5% to 92.7%) for CXR, but heterogeneity between studies was considerable (
121). Specificity for diagnostic imaging is low, at 18.1% (95% CI, 3.7 to 55.8%) (
121). In other words, approximately 80% of individuals would have received a diagnosis of COVID-19 in the absence of disease. As such, the use of diagnostic imaging techniques should be accompanied by careful consideration of factors such as disease prevalence in the study population, severity of the illness, performance and context of the methods used, differences in radiologist opinions, and possible confounding diagnoses (
112–114,
118–121,
123,
126,
132–135). On the other hand, it is also important to recognize that diagnostic imaging is a useful tool for patient management with or without a confirmed etiology through laboratory testing, as it can be used to monitor the severity of illness and disease progression and assess possible complications (
136–141).
Understanding the benefits and limitations of diagnostic imaging for COVID-19 is an active area of research, along with applications of artificial intelligence (AI) (also known as machine learning) (
142–144). AI-based methods can be used in diagnostic imaging to help recognize abnormal features in images and classify them into defined categories, thus increasing accuracy, standardization, and speed of analyses by radiologists (
50,
109,
142–146). AI approaches can be categorized into three main groups: approaches that analyze CT scan images, methods based on X ray, and those that realize diagnosis through jointly analyzing CT scan and X-ray images (
147–151). While AI-based applications have shown benefits for diagnostic imaging methodologies (
50,
109,
145,
146), more clinical investigations are needed to evaluate their possible incorporation into routine procedures for investigations of suspected cases of COVID-19, and laboratory testing is required to confirm the disease etiology. Furthermore, acquiring a reliable AI-based system requires access to a comprehensive training data set that includes all variations of COVID-19 as well as other lung diseases; providing such an all-inclusive data set is difficult and labor-intensive.
CONCLUSIONS AND OUTLOOK
Despite the recent availability of vaccines, with the ongoing and rapid spread of SARS-CoV-2 worldwide, laboratory testing remains the cornerstone of public health containment and mitigation strategies. Given the thoroughness of the data on methodologies in the body of this review, only key findings, some considerations for testing, and areas for improvement and successes are discussed below; however, a summary of testing modalities is presented in
Fig. 13.
While compatible signs and symptoms, routine laboratory testing for biomarkers of health or disease, and diagnostic imaging have roles to play in diagnostic investigations, none alone are sufficient for diagnosing SARS-CoV-2, and the reliance on specific laboratory testing is paramount. Many technological advances have been made to detect SARS-CoV-2, which include NAATs to detect viral RNA and immunoassays to detect viral antigens or virus-specific antibodies generated by the immune system in response to SARS-CoV-2. There is a growing list of commercially available methods for testing for SARS-CoV-2 (
29,
30), and while many have been validated, the results of any test should be interpreted with consideration of the context in which they are used and based on the sum of all diagnostic evidence. When using any method to rule in or out SARS-CoV-2 infection, many factors should be considered, such as the timing and type of specimen collection, the anatomical site of sampling, the method and its expected performance characteristics, host factors like compatible signs and symptoms (versus asymptomatic testing), risk factors for serious outcomes, and disease prevalence in the population (
82,
542–544).
NAATs like real-time RT-PCR quickly became the gold standard for diagnostic testing. However, apart from those listed above, other considerations for NAATs are inherent to the methodologies that are all based on the detection of SARS-CoV-2 RNA. For example, it is important to recognize that the performance of any molecular diagnostic method can be influenced by sequence mismatches between the method’s targets and the different genome permutations of circulating SARS-CoV-2 lineages and variants. A possible strategy to reduce the chance of false-negative results that could occur by target mismatches is the simultaneous use of molecular methods targeting more than one gene, as failure to detect a signal in one target gene may not preclude detection in another (
40,
164,
443). Alternative strategies could include the use of degenerate primers and probes, a strategy used with other RNA viruses that are prone to mutation (
443,
545,
546). Ongoing molecular surveillance using sequencing technologies should also be encouraged to monitor changes in SARS-CoV-2 genome sequences and COVID-19 epidemiology, particularly with an emphasis on variants linked to failures in diagnostic testing, with increased transmissibility, increased severity, or decreased susceptibility to convalescent-phase sera or responses to vaccines.
The primary reliance on NAATs for SARS-CoV-2 detection during the pandemic came with many challenges globally. With human resource strains and supply chain shortages, providing NAATs during the pandemic was challenged by factors such as PPE, human resource strains for sample collection and testing, swabs, and test reagent availability for NAATs. From a clinical perspective, alternatives to NP swabs traditionally used for respiratory specimens were rapidly validated, as was the use of various transport media or swab-free options that are amenable to self-collection (e.g., saliva or saline gargles). The application of these specimens in the laboratory added complexity to the laboratory workflow and required rapid validations to ensure compatibility with new or existing instrumentation.
To help meet capacity demands in the laboratory, laboratories were faced with the need to rapidly procure large quantities of supplies, acquire instrumentation, train additional personnel, and validate specimens, reagents, and equipment. Despite this, supply chain challenges and rapid escalation of testing demands led to the need for resource-sparing strategies for NAATs, including specimen pooling and extraction-free NAAT protocols. Such strategies come at the cost of a relatively reduced sensitivity, which is associated with the potential risk of missing detection of SARS-CoV-2 in specimens with low viral loads. It could be argued that small reductions in sensitivity may have little impact on the detection of most cases. Remnant SARS-CoV-2 at the outset of illness can persist for weeks and is unlikely to represent a period of communicability. On the other hand, missing detection of SARS-CoV-2 in a specimen with a low viral load at the early stages of illness may potentially lead to further virus spread. These testing strategies should be carefully considered prior to implementation.
SARS-CoV-2-specific NAATs evolved over time to facilitate testing and streamline the laboratory workflow. For example, to further streamline testing for respiratory viruses, some methods have now been multiplexed to simultaneously detect SARS-CoV-2, influenza A and B viruses, and other respiratory viruses like respiratory syncytial virus (RSV). In nonpandemic years, influenza and other viral etiologies of respiratory tract infections represented a leading cause of death in North America, particularly among hospitalized patients with community-acquired pneumonia (
547–553). Interestingly, while cocirculation and coinfection with other respiratory viruses were reported, there was little activity for non-SARS-CoV-2 respiratory viruses (
554). It is unclear whether public health interventions (e.g., travel restrictions, social distancing, handwashing, or PPE like masks) resulted in this decline or whether other factors inadvertently contributed to biases in data, such as fewer individuals seeking routine medical attention or the lack of testing for influenza and other respiratory viruses due to competition for resources used for SARS-CoV-2 testing (
554–558). Concomitant diagnostic testing using multiplex technologies could provide an option for syndromic testing that would ensure surveillance for SARS-CoV-2 and other important respiratory viruses like influenza virus and appropriate interventions as needed (e.g., antivirals).
The public and political pressure for laboratory testing evolved with public health indications, and laboratory testing continues to guide public health policies as restrictions ease or escalate throughout the ongoing pandemic (
47,
484). One area of significant advancement in NAAT methodologies includes the use of automation to minimize hands-on processing time and increase specimen throughput. With the high demand for laboratory testing, automation is an important consideration to avoid the possibility of staff repetitive-stress injuries. Over the course of the COVID-19 pandemic, the scale and demand for laboratory testing have been unmatched by other pandemics. Concerns over SARS-CoV-2 prompted the testing of both symptomatic and asymptomatic individuals, in the context of public health case contact tracing and surveillance purposes. This includes testing for SARS-CoV-2 in both health care settings where patients are at increased risk (e.g., hospitals and long-term-care facilities) and situations where testing would not otherwise have been performed (i.e., professional sports teams, public events, prior to or after travel, and various workplaces). Thus, innovative, dynamic, and adaptable approaches were required to meet the testing demands, which are covered extensively in this review.
Other areas are recognized as being crucial to increase laboratory testing capacity but are not covered in this review, such as supply procurement, distribution, and management and the coordination of training or recruitment of increased human resources for specimen collection, testing, processing, and registration. Also, rapid validation of laboratory tests and collection devices to meet regulatory requirements despite availability through EUA from regulatory bodies as well as the development of an information technology infrastructure to continuously improve the laboratory information management system (LIMS) in the laboratory testing workflow and dissemination of near-real-time laboratory data to various stakeholders from local to national levels and dissemination to the public through various media formats are of great value. While digitization was not an absolute requirement, the interconnectivity of data ensured transparency and up-to-date information as decision support tools for recommendations, policies, and guidelines (
53,
559–561). All of these have been identified by the WHO as key factors to control the COVID-19 pandemic (
562).
To further enhance testing capacity and provide access for rapid SARS-CoV-2 testing options, both antigen- and NAAT-based RDTs have been developed and are now readily available (
29,
30). These can support rapid laboratory and POC applications, screening of large patient populations, or rapid deployment for assessments of target areas. NAAT-based RDTs rely on real-time RT-PCR or isothermal amplification methods (e.g., RT-RPA, RT-LAMP, and NEAR) and, like antigen-based RDTs, can provide rapid results without complex instrumentation. The ease of use of these portable devices allows access to SARS-CoV-2 diagnostics in settings that may have been prohibitive for traditional laboratory NAATs. Throughout the literature, the main limitation noted for RDTs is their reduced sensitivity compared to traditional NAATs, but the evaluations of their performance do not always reflect conditions in which the RDT was licensed under EUA, and the comparator method and distribution of expected specimen results can have a large impact on assay performance (
274,
290,
341,
342,
457). Even if these issues are not considered, and all SARS-CoV-2 detection results were considered of value, a possible strategy to mitigate the relatively reduced sensitivity of antigen- or NAAT-based RDTs is to increase the frequency of testing in the patient population over time, to increase the chances of identifying individuals who fall in a period of high viral shedding (which would be less likely to be impacted by reduced sensitivity) (
288,
299,
301,
467,
468). On the other hand, the implementation of such strategies has challenges of its own due to the limited scalability of RDTs, and the balance between sensitivity and testing frequency to achieve optimal SARS-CoV-2 detection in the target population would need to be defined (
288). Moreover, if, for example, EUA defined the use of RDTs as within 7 days of symptom onset, this excluded their use for testing of asymptomatic individuals. While much development is ongoing to enhance existing RDT technologies or explore novel methodologies (
53,
465,
466,
563–565), how RDTs can effectively be used in practice is the subject of ongoing debate, and further research is needed to understand what setting they would best be of benefit.
For serology, currently available commercial assays are based on ELISAs, CLIAs, and LFIAs and are designed to detect SARS-CoV-2 antibodies. However, the performance of serological assays varies across the different technologies, the timing from disease onset, and comparator methods. Clinical validation of serological methods is ongoing, but recent meta-analyses and systematic reviews so far have shown high levels of heterogeneity and a risk of bias and applicability (
472,
502,
503). For example, with the time required to mount anti-SARS-CoV-2 antibodies during seroconversion (or possibly the absence of seroconversion in mild disease), serology has limited value in identifying SARS-CoV-2 in the early stages of illness. Serology has value for seroepidemiological studies to help with ongoing outbreak investigations, can aid in the diagnosis of suspect cases for whom NAATs were persistently negative or not performed, or can help with conditions in children and adolescents like multisystem inflammatory syndrome where NAAT results may be negative (
102,
133,
241,
566–570). Much research is needed to expand the knowledge on the use of immunological methods, particularly in the context of the recent availability of vaccines.
Key areas of research interest for serology should include the relevance, magnitude, and duration of protective antibody responses, particularly faced with a population that may have been exposed to SARS-CoV-2 or other human coronaviruses or vaccinated against SARS-CoV-2. Detection of different antibody isotypes is possible with commercial or laboratory-developed serological assays, and recent quantitative methods have now become commercially available; however, much is yet to be learned on how these methods correlate with protective immune responses. Establishing correlates of protection will require longitudinal studies with parallel assessments of quantitative levels of nAbs and cell-mediated immune responses. Various neutralization assays have been established to assess and quantify antibodies against SARS-CoV-2 in serum or plasma (e.g., PRNTs, MN tests, or surrogate assays like PBNAs and snELISAs), but data are scarce when it comes to longitudinal studies undertaken to fully understand the differences or similarities between commercial assays for the qualitative detection of antibodies, quantitative antibody analyses over time, and relevance to the level and duration of protective nAb titers. While the longevity of humoral immunity is not yet fully understood for SARS-CoV-2, some data suggest that antibody levels against SARS-CoV-2 wane over 3 months. Notwithstanding these observations, it is possible that memory T cells may still be able to mount effective responses upon reexposure to SARS-CoV-2, and the role of cell-mediated immune responses may also provide additional benefits (
508,
509). Further research is needed, as a greater understanding of the immune response to SARS-CoV-2 is fundamental in making informed recommendations for the use of immunological methods to assess current or future protection against the virus, to determine meaningful endpoints in the development and evaluation of effective vaccines, or to assess effectiveness following immunization with existing vaccines.
Overall, methods based on RNA, antigen, or antibody detection as well as diagnostic imaging all have a place in our response to SARS-CoV-2, but ongoing research is crucial to further optimize and apply all these testing modalities. The emergence of the SARS-CoV-2 pandemic has created a wave of innovative and creative thinking, and more and more creative methods and platforms are being introduced with goals to increase the armamentarium of diagnostic methods for SARS-CoV-2 detection. An understanding of the advantages and limitations of each method used for SARS-CoV-2 detection as well as the development of novel methods will help us unravel the unknowns of disease pathogenesis, epidemiology, and transmissibility and help us develop interventions to mitigate and contain its spread. There will always be a need for laboratory testing and collaboration between clinical laboratories, public health, infection prevention and control, and many others who contribute to the efforts to contain the spread of COVID-19. On the other hand, it is also recognized that maintaining large investments in the rapid deployment of translational research and such a high degree of laboratory testing for COVID-19 will likely not be sustainable from an economic perspective, and justifications for such investments will be more difficult if cases decline significantly with vaccines. However, the lessons learned from SARS-CoV-2 could potentially be used in the preparedness for potential future pandemic threats, thus strengthening global health and surveillance systems.