Researchers at Tulane University School of Medicine have developed a new highly sensitive blood test for tuberculosis (TB) that screens for DNA fragments of the Mycobacterium tuberculosis bacteria that causes the deadly disease.
The test could give doctors a new tool to both quickly identify TB and then gauge whether drug treatments are effective by monitoring levels of DNA from the pathogen circulating through the bloodstream, according to a new study published in the journal The Lancet Microbe.
Tuberculosis is now the second most deadly infectious disease in the world, behind only COVID-19. In 2020, an estimated 10 million people contracted TB and 1.5 million people died from it, according to the World Health Organization.
Most TB tests rely on screening sputum, a thick type of mucus from the lungs. But collecting sputum from patients suspected of having TB can be difficult, especially for children.
In these extrapulmonary cases, patients can have little bacteria in the sputum, which leads to false negatives using current testing methods, said lead study author Tony Hu, Ph.D., Weatherhead Presidential Chair in Biotechnology Innovation at Tulane University.
“This assay may be a game-changer for TB diagnoses that not only provides accurate diagnosis results but also has the potential to predict disease progression and monitor treatment,” Hu said. “This will help doctors rapidly intervene in treatment and reduce the risk of death, especially for children living with HIV.”
The study evaluated a CRISPR-based assay that screened for cell-free DNA from live Mycobacterium tuberculosis bacilli. The screening target is released into the bloodstream and cleared quite rapidly, providing a real-time snapshot of active infection.
Researchers tested preserved blood samples from 73 adults and children with presumptive TB and their asymptomatic household contacts in Eswatini, Africa.
The test identified adult TB with 96.4% sensitivity and 94.1% specificity and pediatric TB with 83.3% sensitivity and 95.5% specificity. (Sensitivity refers to how well a test can diagnose a positive case, while specificity is a measure of a test’s accurately determining a negative case.)
Researchers also tested 153 blood samples from a cohort of hospitalized children in Kenya. These were HIV-positive patients who were at high risk for TB and presented with at least one symptom of the disease. The new test picked up all 13 confirmed TB cases and almost 85% of unconfirmed cases, which were cases that were diagnosed due to clinical symptoms and not existing gold standard testing methods.
The CRISPR-based test uses a small blood sample and can deliver results within two hours.
“We are particularly excited that the level of Mycobacterium tuberculosis cell-free DNA in HIV-infected children began to decline within a month of treatment, and most of the children’s blood was cleared of the bacteria DNA fragments after treatment, which means that CRISPR-TB has the potential to monitor treatment and will give physicians the ability to better treat worldwide TB infections,” Hu said.
The researchers have since adapted the assay to a rapid test platform that can deliver results in 30 minutes without any special equipment. Results would be viewable on a paper strip like a rapid COVID-19 test.
“A highly accurate, rapid blood test that could be used anywhere would benefit millions of people living in resource-limited areas with a high TB burden,” Hu said.
Currently Available Point-of-Care Tests (POCT) for Childhood Tuberculosis
Detection of Lipoarabinomannan (LAM) in Urine Using Lateral Flow Assays
Tests based on detection of mycobacterial lipoarabinomannan (LAM) antigen in urine have been in use since 2015. LAM is a constituent glycolipid from Mycobacterium tuberculosis (Mtb) that is released from metabolically active or degenerating mycobacterial cells into urine. The Alere Determine™ TB LAM Ag (AlereLAM) was the first commercially available assay using lateral flow for LAM detection in TB patients. As the sensitivity of the AlereLAM is limited, the WHO, however, recommends its use mostly in inpatient settings . In addition, it is only recommended as a rule-in test for TB disease in individuals living with HIV infection including adults, adolescents, and children who are seriously ill, defined as having fever above 39 °C, being tachypneic and tachycardic and “unable to walk unaided” . Additionally, the test is recommended for those with advanced HIV disease, defined as CD4 cell count of < 200 cells/mm³. Studies determining the test performance of AlereLAM in children living with HIV have shown widely varying sensitivity ranging from 43–65% and specificity from 57–94% .
In recent years, a new LAM detection assay, Fujifilm SILVAMP TB LAM (FujiLAM, Fujifilm, Tokyo, Japan) has been introduced. The improvements include a combination of high-affinity monoclonal antibodies against M. tuberculosis-specific LAM epitopes and a silver amplification step to increase the visibility of test and control lines. Studies show better sensitivity in both HIV-infected and uninfected patients than its predecessor [10,11]. The sensitivity of FujiLAM in HIV-uninfected TB outpatients, reported in a multicenter cohort study in Peru and South Africa, was 53%, which was five times higher than that obtained by AlereLAM . In a meta-analysis including HIV-infected adults, the sensitivity in patients with confirmed TB was 71% (95%CI: 59–81%) .
A Nigerian study found similar sensitivities of FujiLAM in HIV-infected and -uninfected adults (66 and 70%, respectively), whereas a study in Zambia found a sensitivity of 75% in both HIV-infected and -uninfected groups [13,14].
The biological mechanisms for the possible better performance of LAM assays in HIV-infected adults are not fully understood. It has been speculated that the higher concentration of LAM in the urine of HIV-infected individuals may reflect an increased hematogenous spread of Mtb to the kidneys in those patients . In that context, and given the increased likelihood of disseminated TB in children, it was hypothesized that LAM assays may have increased sensitivity in children.
However, the sensitivity of FujiLAM in HIV-infected children was approximately 55–60% and therefore lower than in HIV-infected adults [9,16]. Interestingly, the sensitivity in HIV-uninfected children was higher than expected, with 60–65%, and was therefore similar to that reported in HIV-infected adults [16,17]. The reasons behind these observations are speculative, but Mtb bacterial load and TB severity may influence the detection of LAM in urine [11,18].
In children specifically, FujiLAM sensitivity increased with more advanced stages of the disease, and the number of positive cases was proportionally higher among underweight and stunted children . This suggests that undernutrition may facilitate bacteremia, increasing the amount of LAM in urine samples and thus the sensitivity of LAM tests [19,20]. These findings show that LAM assays have a potential as adjunctive TB diagnostic tests in children, but further improvements in sensitivity are required to enable their use as rule out tests. Specificity in FujiLAM was higher in children compared to AlereLAM, and both tests show a higher specificity in children older than 2 years, suggesting a robust “rule-in” test for older, hospitalized children living with HIV (8). As HIV-coinfection and malnourishment are risk factors for rapid deterioration, ruling-in tests in these individuals is crucial.
Loop-Mediated Isothermal Amplification (LAMP)
The TB loop-mediated isothermal amplification (LAMP) assay, developed by Eiken Chemical Company Ltd. (Tokyo, Japan), is a molecular diagnostic test that detects Mtb complex deoxyribonucleic acid (DNA) in sputum samples. The test was officially endorsed by the WHO in 2016 . The LAMP method obviates the need for thermal cyclers required for traditional polymerase chain reaction assays . Further advantages are that the assay provides a result in about one hour, that it can be read by eye under ultraviolet light, and that it requires only limited training. The main disadvantage is that it lacks the ability to detect drug resistance .
Additionally, in contrast to Xpert MTB/RIF, the assay has only been validated for use in sputum samples, so that it cannot be used for gastric aspirate specimen, which are more commonly obtained in younger children. Consequently, since the wide-spread implementation of Xpert MTB/RIF, which detects both the organism and rifampicin-resistance, the assay has only found application in a limited range of settings, mainly in environments that prevent the use of the GeneXpert instrument (e.g., with unreliable electricity, extreme temperatures and humidity or excessive dust). Crucially, the WHO recommendations highlight that implementation of TB-LAMP does not eliminate the need for smear microscopy, as the latter is currently recommended for treatment response monitoring. Similarly, the Xpert MTB/RIF assay is also not considered suitable for monitoring as it does not distinguish between viable and non-viable bacilli.
A systematic review and meta-analysis performed in 2016 found that in adults with culture-confirmed pulmonary TB the performance characteristics of the TB-LAMP assay were very similar to that of the Xpert MTB/RIF assay, with pooled sensitivity and specificity estimates of 78% and 99% versus 81% and 98%, respectively . However, subgroup analyses revealed that the sensitivity of the TB-LAMP assay appears to be substantially lower in HIV-infected TB patients, reducing the assay’s usefulness in regions with high HIV prevalence, as illustrated by an adult study in Uganda that reported an overall sensitivity of only 55% . To date, data on the performance of TB-LAMP in children remain very limited, with fewer than a thousand patients included across all published studies combined [25,26,27]. Nevertheless, the available data indicate that the assay has similar sensitivity and specificity in adult and pediatric setting including young children under the age of 5 years, which qualifies TB-LAMP as a potential tool that overcomes the hampering paucibacillary nature of TB in children.
PCR-Based Point-of-Care Tests
A large number of studies have evaluated the performance of the “classical” Xpert MTB/RIF assay across various locations and using a broad range of clinical specimens (e.g., respiratory samples, pleural fluid, cerebrospinal fluid, ascitic fluid and lymph node material). A recent, detailed meta-analysis estimated the sensitivity of the MTB/RIF assay performed on sputum in children with pulmonary TB to be 65% with a specificity estimate of 99% .
The same meta-analysis, using culture-confirmation as the comparator, estimated the sensitivity of the assay in pulmonary TB with different clinical specimens to be as follows: gastric aspirates 73%, nasopharyngeal aspirates 46%, and stool 62%. In the setting of TB meningitis (analysis of CSF) and lymph node TB (analysis of lymph node material), the sensitivity of the Xpert MTB/RIF assay was estimated to be 54% and 90%, respectively. In 2017, a new generation assay termed Xpert MTB/RIF Ultra, incorporating two different multi-copy amplification targets (IS6110 and IS1081) was introduced. To date, the published data on the performance of Xpert MTB/RIF Ultra in children remain relatively limited [29,30,31].
Compared to microbiological reference standards (that varied between studies), key pediatric studies reported sensitivities ranging from 64% to 75% in sputum samples in children with pulmonary TB; however, the study with the lowest sensitivity estimate used frozen sputum pellets instead of fresh samples, which may have impacted assay performance [29,30,31]. Nevertheless, those results suggest that one quarter of children with pulmonary TB have false-negative results, highlighting the need for further improvements of the sensitivity of this assay to be used as a rule-out test.
Additionally, as children have distinctively lower bacterial load compared to adults, “trace” results for Mtb detection are more frequent, which also results in loss of rifampicin-resistance results. Despite its limitations, Xpert testing allows for a more rapid diagnosis awaiting culture confirmation, or it can even substitute culture testing in settings where the latter is not available, which is already a considerable progress. Importantly, Xpert using stool specimen has become of interest, because its collection is easier compared to sputum collection in children. Two systematic reviews and meta-analyses showed a pooled sensitivity of 50% (95% CI 0.44–0.55) and 67% (95% CI, 52–79%) compared with the reference standard of culture and/or respiratory samples using Xpert [32,33]. Despite lower overall sensitivity compared to sputum, its child-friendly collection and preclusion of aerosol-generation is promising, especially in the context of adaptation of sample processing, which can further increase the detection yield .
The GeneXpert Omni instrument developed by Cepheid (Sunnyvale, CA, USA) is a portable system designed specifically for POCT use . The Omni runs the same PCR-based Xpert cartridges as the ‘classical’ GeneXpert platforms and was therefore expected to achieve similar accuracy as larger, previous generation instruments. However, the commercialization of GeneXpert Omni has been halted by Cepheid . Another GeneXpert instrument that was designed for the low-resource setting and has subsequently been endorsed by the WHO Global TB Program is the GeneXpert Edge system, which was launched in 2018. It comprises a single module that can be powered by an external battery pack and is operated by a touch screen tablet. In common with the Omni, the Edge instrument can process regular Xpert cartridges, although the need for air-conditioned environmental temperatures limits its use beyond a district laboratory level. Early observations from a study in Brazil, which commenced in 2020 but had to be put on hold because of the COVID-19 pandemic, suggest that the Edge instrument facilitates the implementation of TB testing in remote communities with very limited infrastructure .
In 2020, a further PCR-based assay was endorsed by the WHO. The Truenat assays, developed by MolBio Diagnostics Pvt. Ltd. (Verna, India), are based on real-time PCR chips that are analyzed on the company’s Truelab instruments. The instruments are compact, portable and can be operated with a battery-pack, making them ideal for use as a POCT. So far, all studies evaluating these assays were performed in India, with some including head-to-head comparisons between Xpert MTB/RIF and Truenat MTB indicating both assays had similar performance characteristics . A large multi-center study led by FIND, which included 1807 adult participants with suspected pulmonary TB from study sites in India, Peru, Ethiopia and Papua New Guinea, published its results recently, reporting overall pooled sensitivity estimates of 73% and 80% for the Truenat MTB and the Truenat MTB Plus assay, respectively . However, the sensitivity of both assays was far lower in smear-negative, culture-positive participants (36% and 47%, respectively), which raises concerns regarding their performance in children, who typically have paucibacillary TB disease. To date, the assays have not been evaluated in a dedicated pediatric study.
There are several other compact POCT molecular test platforms commercially available that enable the detection of a variety of pathogens, including the Alere q system (Alere Inc., Waltham, MA, USA), ID Now (Abbott Laboratories, Chicago, IL, USA), Revogene (Meridian Bioscience, Cincinnati, OH, USA), Biofire Filmarray (bioMérieux, Marcy-l’étoile, France), Q-POC (QuantuMDx, Newcastle upon Tyne, UK) and Cobas Liat (Roche, Basel, Switzerland), but at present, none of those systems have commercially available assays for the detection of Mtb. However, several of those manufacturers are currently working on diagnostic TB assays for their POCT platforms.
Pipeline and Developments of Point-of-Care Tests for the Diagnosis of Tuberculosis
Immune Response Biomarkers
The C-reactive protein (CRP), a well-established host immune biomarker of infection, received renewed attention after the WHO recently endorsed CRP as a screening marker in HIV-infected individuals . Advantageous for this approach is that reliable POCT assays detecting CRP are widely available. However, CRP is a sensitive but non-specific marker of infection and inflammation. Consequently, it will not be useful as a stand-alone test for TB evaluation but may be useful for inclusion in a combined biomarker approach. This has been shown in recent promising protein biosignature studies including adults and children [71,72].
Among newer host immune biomarkers currently evaluated for TB diagnosis, antibodies, cytokines, ribonucleic acid (RNA) signatures and host cellular markers are the most commonly investigated. A recently published systematic review  showed that among 44 biomarkers that met at least one minimal targeted product profile (TPP) defined by the WHO, 37/44 (84%) were host and 7/44 (16%) were pathogen biomarkers. Host biomarkers seem to be sensitive but often have limited specificity. Therefore, they are generally more useful as screening tests rather than confirmatory tests, for which a POCT format is highly desirable. The same review showed that the majority of host biomarkers were multiple marker biosignatures. The outperformance of multiple marker biosignatures over single markers mirrors the complexity of the underlying pathophysiology in TB.
Development of multiple marker POCT would thus be a major achievement but poses additional challenges as most POCTs are single marker based. In that context, RNA biomarkers may be particularly attractive, as the analyses of several targets can be readily multiplexed into a single POCT. However, a recent systematic review and meta-analysis showed insufficient accuracy of RNA-based blood signatures for point of care TB confirmation . At this stage, RNA signatures would therefore only reach the TPP targets for a screening test. Another option lies in the measurement of host T cell biomarkers using flow cytometry.
Currently, the detection of such markers includes surface molecules such as cluster of differentiation (CD) 27, CD38, or CD153, as well as human leukocyte antigen (HLA)-DR and cell proliferation markers such as Kiel (Ki) 67 [75,76,77,78]. These markers have repeatedly achieved sensitivity and specificity levels compatible with TPP targets for a confirmatory TB test, including in children. Substantial simplification of this approach has been successfully achieved to allow its implementation in a resource-limited setting . Nonetheless, functional T cell assays still require laboratory infrastructure and have a 24 h turn-around time, which remains incompatible with the ambitious POCT TPP target to deliver results on the same day and hampers implementation in remote areas.
Antibodies were the most studied host immune biomarkers. Despite this, older studies evaluating the diagnostic performance of TB-specific antibodies were rather discouraging . However, more recent studies showed promising results for antibody signatures [81,82]. Recently, TB-specific IgG4 has been shown to correlate with disease activity and decrease after treatment .
The potential advantages of antibody biomarkers include high stability of the marker of interest, exciting robust low-cost assays for other pathogens (e.g., lateral flow assays), and limited operator training. Those features highlight the potential for antibody-based tests to be translated into POCT format.
Biomarker Detection in Urine
Urine is a readily available and non-invasive specimen, and therefore particularly useful for the diagnosis of TB in patients unable to expectorate sputum, such as children, the elderly and adults without a productive cough [13,84]. In addition to the aforementioned AlereLAM and FujiLAM, other rapid urine-based tests for the diagnosis of TB are currently being developed. FLOW-TB is a lateral flow urine LAM assay that includes a concentration step using exclusion-based sample preparation technology to increase the detection of LAM. The assay has shown a sensitivity of 52% in HIV-uninfected patients with a specificity of 87%, which is below the WHO minimum TPP . Along the same lines, a recently published study reported a urine LAM concentration method in combination with a sensitive lateral flow assay to improve LAM detection and showed a sensitivity of 60% and specificity of 80%, which was comparable in HIV-infected and -uninfected patients . In addition, further studies have highlighted the importance of sample pre-treatment to improve LAM assay performance .
A recent development is PCR-based detection of TB-specific cell-free DNA fragments excreted in the urine . Using a highly sensitive sequence-specific purification method, the first results of this urine-based assay in a South African cohort showed a sensitivity of 84% and a specificity of 100% . Currently, no data is available from pediatric studies. However, the operational characteristics of this test would need to be simplified to allow its implementation in resource-limited settings.
Artificial Intelligence-Supported Interpretation Radiography
Artificial intelligence (AI)-based interpretation of medical imaging has recently been extensively researched and developed further, and multiple commercial products have become available for clinical use [89,90]. Computer-aided detection (CAD) products indicating the likelihood of TB use AI to analyze radiographs and express abnormality scores. Diagnostic performance of CAD software proved similar to interpretation of digital chest radiography by a human reader, with estimates ranging around 90–92% for sensitivity and 23–79% for specificity . CAD has recently been conditionally recommended by WHO as an alternative to human interpretation of digital chest radiography for screening and triage of TB in individuals ≥ 15 years of age .
CAD systems are continuously being improved to further optimize test accuracy, turnaround times (currently < 15 s per chest radiograph) and costs (currently around USD 8 per analysis). Prerequisites for CAD in routine care comprise digital radiography equipment and internet access, as well as funds for the CAD product license fees . Few CAD studies focusing of pneumonia and TB have been conducted in pediatric populations [91,92,93]. Despite the challenges of a higher spectrum of chest shapes and disease presentation in younger children, CAD algorithms can outperform human readers and a first software (CAD4TB v6) has recently been licensed for children ≥ 4 years of age.
reference link : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8949489/
More information: Zhen Huang et al, CRISPR detection of circulating cell-free Mycobacterium tuberculosis DNA in adults and children, including children with HIV: a molecular diagnostics study, The Lancet Microbe (2022). DOI: 10.1016/S2666-5247(22)00087-8