Positional contrastive learning for improved thigh muscle segmentation in MR images.

The accurate segmentation of individual muscles is essential for quantitative MRI analysis of thigh images. Deep learning methods have achieved state-of-the-art results in segmentation, but they require large numbers of labeled data to perform well. However, labeling individual thigh muscles slice by slice for numerous volumes is a laborious and time-consuming task, which limits the availability of annotated datasets. To address this challenge, self-supervised learning (SSL) emerges as a promising technique to enhance model performance by pretraining the model on unlabeled data. A recent approach, called positional contrastive learning, exploits the information given by the axial position of the slices to learn features transferable on the segmentation task. The aim of this work was to propose positional contrastive SSL for the segmentation of individual thigh muscles from MRI acquisitions in a population of elderly healthy subjects and to evaluate it on different levels of limited annotated data. An unlabeled dataset of 72 T1w MRI thigh acquisitions was available for SSL pretraining, while a labeled dataset of 52 volumes was employed for the final segmentation task, split into training and test sets. The effectiveness of SSL pretraining to fine-tune a U-Net architecture for thigh muscle segmentation was compared with that of a randomly initialized model (RND), considering an increasing number of annotated volumes (S = 1, 2, 5, 10, 20, 30, 40). Our results demonstrated that SSL yields substantial improvements in Dice similarity coefficient (DSC) when using a very limited number of labeled volumes (e.g., for S = 1, DSC 0.631 versus 0.530 for SSL and RND, respectively). Moreover, enhancements are achievable even when utilizing the full number of labeled subjects, with DSC = 0.927 for SSL and 0.924 for RND. In conclusion, positional contrastive SSL was effective in obtaining more accurate thigh muscle segmentation, even with a very low number of labeled data, with a potential impact of speeding up the annotation process in clinics.

Pooled sputum to optimise the efficiency and utility of rapid, point-of-care molecular SARS-CoV-2 testing.

BACKGROUND: As SARS-CoV-2 testing expands, particularly to widespread asymptomatic testing, high sensitivity point-of-care PCR platforms may optimise potential benefits from pooling multiple patients' samples. METHOD: We tested patients and asymptomatic citizens for SARS-CoV-2, exploring the efficiency and utility of CovidNudge (i) for detection in individuals' sputum (compared to nasopharyngeal swabs), (ii) for detection in pooled sputum samples, and (iii) by modelling roll out scenarios for pooled sputum testing. RESULTS: Across 295 paired samples, we find no difference (p = 0.1236) in signal strength for sputum (mean amplified replicates (MAR) 25.2, standard deviation (SD) 14.2, range 0-60) compared to nasopharyngeal swabs (MAR 27.8, SD 12.4, range 6-56). At 10-sample pool size we find some drop in absolute strength of signal (individual sputum MAR 42.1, SD 11.8, range 13-60 vs. pooled sputum MAR 25.3, SD 14.6, range 1-54; p < 0.0001), but only marginal drop in sensitivity (51/53,96%). We determine a limit of detection of 250 copies/ml for an individual test, rising only four-fold to 1000copies/ml for a 10-sample pool. We find optimal pooled testing efficiency to be a 12-3-1-sample model, yet as prevalence increases, pool size should decrease; at 5% prevalence to maintain a 75% probability of negative first test, 5-sample pools are optimal. CONCLUSION: We describe for the first time the use of sequentially dipped sputum samples for rapid pooled point of care SARS-CoV-2 PCR testing. The potential to screen asymptomatic cohorts rapidly, at the point-of-care, with PCR, offers the potential to quickly identify and isolate positive individuals within a population "bubble".

The use of whole-genome sequencing in cluster investigation of a multidrug-resistant tuberculosis outbreak.

We used whole-genome sequencing (WGS) to delineate transmission networks and investigate the benefits of WGS during cluster investigation.We included clustered cases of multidrug-resistant (MDR) tuberculosis (TB)/extensively drug-resistant (XDR) TB linked by mycobacterial interspersed repetitive unit variable tandem repeat (MIRU-VNTR) strain typing or epidemiological information in the national cluster B1006, notified between 2007 and 2013 in the UK. We excluded from further investigation cases whose isolates differed by greater than 12 single nucleotide polymorphisms (SNPs). Data relating to patients' social networks were collected.27 cases were investigated and 22 had WGS, eight of which (36%) were excluded as their isolates differed by more than 12 SNPs to other cases. 18 cases were ruled into the transmission network based on genomic and epidemiological information. Evidence of transmission was inconclusive in seven out of 18 cases (39%) in the transmission network following WGS and epidemiological investigation.This investigation of a drug-resistant TB cluster illustrates the opportunities and limitations of WGS in understanding transmission in a setting with a high proportion of migrant cases. The use of WGS should be combined with classical epidemiological methods. However, not every cluster will be solvable, regardless of the quality of genomic data.

Tracking and responding to an outbreak of tuberculosis using MIRU-VNTR genotyping and whole genome sequencing as epidemiological tools.

Background: We describe an outbreak that contributed to a near doubling of the incidence of tuberculosis in Southampton, UK. We examine the importance of 24 locus mycobacterial interspersed repetitive unit variable number tandem repeat (MIRU-VNTR) genotyping in its identification and management and the role of whole genome sequencing (WGS) in tracing the spread of the strain. Methods: Outbreak cases were defined as those diagnosed between January and December 2011 with indistinguishable 24 locus-MIRU-VNTR genotypes or, cases linked epidemiologically. A cluster questionnaire was administered by TB nurses to identify contacts and social settings. Results: Overall, 25 patients fulfilled the case definition. No cases with this MIRU-VNTR genotype had been detected in the UK previously. Connections were found between all cases through household contacts or social venues including a football club, Internet cafe and barber's shop. Public health actions included extended contact tracing, venue screening and TB awareness-raising. The outbreak resulted in a high rate of transmission and high incidence of clinical disease among contacts. Conclusions: This outbreak illustrates the value of combining active case-finding with prospective MIRU-VNTR genotyping to identify settings to undertake public health action. In addition WGS revealed that the VNTR-defined cluster was a single outbreak and that active TB transmission not reactivation was responsible for this outbreak in non-UK born individuals.

Whole Genome Sequence Analysis of a Large Isoniazid-Resistant Tuberculosis Outbreak in London: A Retrospective Observational Study.

BACKGROUND: A large isoniazid-resistant tuberculosis outbreak centred on London, United Kingdom, has been ongoing since 1995. The aim of this study was to investigate the power and value of whole genome sequencing (WGS) to resolve the transmission network compared to current molecular strain typing approaches, including analysis of intra-host diversity within a specimen, across body sites, and over time, with identification of genetic factors underlying the epidemiological success of this cluster. METHODS AND FINDINGS: We sequenced 344 outbreak isolates from individual patients collected over 14 y (2 February 1998-22 June 2012). This demonstrated that 96 (27.9%) were indistinguishable, and only one differed from this major clone by more than five single nucleotide polymorphisms (SNPs). The maximum number of SNPs between any pair of isolates was nine SNPs, and the modal distance between isolates was two SNPs. WGS was able to reveal the direction of transmission of tuberculosis in 16 cases within the outbreak (4.7%), including within a multidrug-resistant cluster that carried a rare rpoB mutation associated with rifampicin resistance. Eleven longitudinal pairs of patient pulmonary isolates collected up to 48 mo apart differed from each other by between zero and four SNPs. Extrapulmonary dissemination resulted in acquisition of a SNP in two of five cases. WGS analysis of 27 individual colonies cultured from a single patient specimen revealed ten loci differed amongst them, with a maximum distance between any pair of six SNPs. A limitation of this study, as in previous studies, is that indels and SNPs in repetitive regions were not assessed due to the difficulty in reliably determining this variation. CONCLUSIONS: Our study suggests that (1) certain paradigms need to be revised, such as the 12 SNP distance as the gold standard upper threshold to identify plausible transmissions; (2) WGS technology is helpful to rule out the possibility of direct transmission when isolates are separated by a substantial number of SNPs; (3) the concept of a transmission chain or network may not be useful in institutional or household settings; (4) the practice of isolating single colonies prior to sequencing is likely to lead to an overestimation of the number of SNPs between cases resulting from direct transmission; and (5) despite appreciable genomic diversity within a host, transmission of tuberculosis rarely results in minority variants becoming dominant. Thus, whilst WGS provided some increased resolution over variable number tandem repeat (VNTR)-based clustering, it was insufficient for inferring transmission in the majority of cases.

Correlates between models of virulence for Mycobacterium tuberculosis among isolates of the Central Asian lineage: a case for lysozyme resistance testing?

Virulence factors (VFs) contribute to the emergence of new human Mycobacterium tuberculosis strains, are lineage dependent, and are relevant to the development of M. tuberculosis drugs/vaccines. VFs were sought within M. tuberculosis lineage 3, which has the Central Asian (CAS) spoligotype. Three isolates were selected from clusters previously identified as dominant in London, United Kingdom. Strain-associated virulence was studied in guinea pig, monocyte-derived macrophage, and lysozyme resistance assays. Whole-genome sequencing, single nucleotide polymorphism (SNP) analysis, and a literature review contributed to the identification of SNPs of interest. The animal model revealed borderline differences in strain-associated pathogenicity. Ex vivo, isolate C72 exhibited statistically significant differences in intracellular growth relative to C6 and C14. SNP candidates inducing lower fitness levels included 123 unique nonsynonymous SNPs, including three located in genes (lysX, caeA, and ponA2) previously identified as VFs in the laboratory-adapted reference strain H37Rv and shown to confer lysozyme resistance. C72 growth was most affected by lysozyme in vitro. A BLAST search revealed that all three SNPs of interest (C35F, P76Q, and P780R) also occurred in Tiruvallur, India, and in Uganda. Unlike C72, however, no single isolate identified through BLAST carried all three SNPs simultaneously. CAS isolates representative of three medium-sized human clusters demonstrated differential outcomes in models commonly used to estimate strain-associated virulence, supporting the idea that virulence varies within, not just across, M. tuberculosis lineages. Three VF SNPs of interest were identified in two additional locations worldwide, which suggested independent selection and supported a role for these SNPs in virulence. The relevance of lysozyme resistance to strain virulence remains to be established.

Microevolution of extensively drug-resistant tuberculosis in Russia.

Extensively drug-resistant (XDR) tuberculosis (TB), which is resistant to both first- and second-line antibiotics, is an escalating problem, particularly in the Russian Federation. Molecular fingerprinting of 2348 Mycobacterium tuberculosis isolates collected in Samara Oblast, Russia, revealed that 72% belonged to the Beijing lineage, a genotype associated with enhanced acquisition of drug resistance and increased virulence. Whole-genome sequencing of 34 Samaran isolates, plus 25 isolates representing global M. tuberculosis complex diversity, revealed that Beijing isolates originating in Eastern Europe formed a monophyletic group. Homoplasic polymorphisms within this clade were almost invariably associated with antibiotic resistance, indicating that the evolution of this population is primarily driven by drug therapy. Resistance genotypes showed a strong correlation with drug susceptibility phenotypes. A novel homoplasic mutation in rpoC, found only in isolates carrying a common rpoB rifampicin-resistance mutation, may play a role in fitness compensation. Most multidrug-resistant (MDR) isolates also had mutations in the promoter of a virulence gene, eis, which increase its expression and confer kanamycin resistance. Kanamycin therapy may thus select for mutants with increased virulence, helping preserve bacterial fitness and promoting transmission of drug-resistant TB strains. The East European clade was dominated by two MDR clusters, each disseminated across Samara. Polymorphisms conferring fluoroquinolone resistance were independently acquired multiple times within each cluster, indicating that XDR TB is currently not widely transmitted.

Heterogeneity of phenotypic characteristics of the modern and ancestral Beijing strains of Mycobacterium tuberculosis.

BACKGROUND: The Beijing strain of Mycobacterium tuberculosis (MTB) is of great concern because this hypervirulent strain has caused numerous tuberculosis outbreaks. However, the mechanisms that allow the MTB Beijing strain to be highly pathogenic remain unclear and previous studies have revealed heterogeneity within this family. OBJECTIVE: To determine the association between some phenotypic characteristics and phylogroups of the Beijing strain of MTB. METHODS: Eight Beijing strains, 5 modern and 3 ancestral sublineages, were selected from the phylogroups of MTB. The selection was based on copy number of IS6110 at NTF, region of differences, and single nucleotide polymorphisms. The abilities of these strains to grow intracellularly in THP-1 macrophages, to induce apoptosis, necrosis, and cytokines production were examined using quantitative real-time PCR and commercially available ELISA kits, respectively. RESULTS: There were some significant differences between the two sublineages of the Beijing strain of MTB. The ancestral Beijing sublineages showed higher intracellular growth rates (p < 0.05) and necrosis induction rates (p < 0.01) than the modern Beijing sublineages. By contrast, the modern Beijing sublineages induced lower apoptosis and protective cytokine responses, i.e., TNF-α (p < 0.05) and IL-6 (p < 0.01) and higher non-protective IL-10 response. The modern Beijing sublineages may have evolved so that they have greater ability to diminish host defense mechanisms. The slower growth rate and reduced necrosis induction in host cells might allow the bacteria to cause a persistent infection. CONCLUSION: The results revealed a phylogroup-associated heterogeneity of phenotypes among MTB Beijing sublineages.

Classification of Muscular Dystrophies from MR Images Improves Using the Swin Transformer Deep Learning Model.

Muscular dystrophies present diagnostic challenges, requiring accurate classification for effective diagnosis and treatment. This study investigates the efficacy of deep learning methodologies in classifying these disorders using skeletal muscle MRI scans. Specifically, we assess the performance of the Swin Transformer (SwinT) architecture against traditional convolutional neural networks (CNNs) in distinguishing between healthy individuals, Becker muscular dystrophy (BMD), and limb-girdle muscular Dystrophy type 2 (LGMD2) patients. Moreover, 3T MRI scans from a retrospective dataset of 75 scans (from 54 subjects) were utilized, with multiparametric protocols capturing various MRI contrasts, including T1-weighted and Dixon sequences. The dataset included 17 scans from healthy volunteers, 27 from BMD patients, and 31 from LGMD2 patients. SwinT and CNNs were trained and validated using a subset of the dataset, with the performance evaluated based on accuracy and F-score. Results indicate the superior accuracy of SwinT (0.96), particularly when employing fat fraction (FF) images as input; it served as a valuable parameter for enhancing classification accuracy. Despite limitations, including a modest cohort size, this study provides valuable insights into the application of AI-driven approaches for precise neuromuscular disorder classification, with potential implications for improving patient care.

Rapid diagnostics of tuberculosis and drug resistance in the industrialized world: clinical and public health benefits and barriers to implementation.

In this article, we give an overview of new technologies for the diagnosis of tuberculosis (TB) and drug resistance, consider their advantages over existing methodologies, broad issues of cost, cost-effectiveness and programmatic implementation, and their clinical as well as public health impact, focusing on the industrialized world. Molecular nucleic-acid amplification diagnostic systems have high specificity for TB diagnosis (and rifampicin resistance) but sensitivity for TB detection is more variable. Nevertheless, it is possible to diagnose TB and rifampicin resistance within a day and commercial automated systems make this possible with minimal training. Although studies are limited, these systems appear to be cost-effective. Most of these tools are of value clinically and for public health use. For example, whole genome sequencing of Mycobacterium tuberculosis offers a powerful new approach to the identification of drug resistance and to map transmission at a community and population level.

Posttreatment reactivation of tuberculosis in mice caused by Mycobacterium tuberculosis disrupted in mce1R.

BACKGROUND: The reactivation of tuberculosis arises in persons who are latently infected and in those who have been previously treated. The mechanism of the reactivation of tuberculosis in either situation is not well understood. A 13-gene mce1 operon of Mycobacterium tuberculosis was previously shown to be associated with latent infection in mice and may also play a role in reactivation. METHODS: We tested mce1 operon M. tuberculosis mutants in a Cornell mouse model to examine disease progression and reactivation. RESULTS: In BALB/c mice, the wild-type, mce1 operon mutant, and mce1R (negative transcriptional regulator of the mce1 operon) mutant M. tuberculosis strains were equally susceptible to orally administered isoniazid and pyrazinamide. However, after cessation of the treatment, the mce1R mutant rapidly and progressively proliferated in mouse lungs and spleens, whereas the other strains remained latent. The reactivation of the mce1R mutant was associated with disease progression in the mouse lungs. CONCLUSION: This observation demonstrates that the constitutive expression of the mce1 genes by M. tuberculosis in the latent state can cause a reactivation of tuberculosis. The constitutive expression of the mce1 genes in the mce1R mutant may allow this mutant to maintain its lipid metabolism, enabling it to survive long-term and proliferate inside granulomas.

A carbapenem-susceptible Pseudomonas aeruginosa strain carrying the bla(SPM) gene.

We studied a carbapenem-susceptible Pseudomonas aeruginosa strain that does not produce carbapenemase but carries the metallo-beta-lactamase gene bla(SPM) identical in sequence to the gene of other fully carbapenem-resistant isolates. Carbapenem-susceptible isolates may be silent reservoirs of the bla(SPM) gene.

A phylogenomic analysis of the Actinomycetales mce operons.

BACKGROUND: The genome of Mycobacterium tuberculosis harbors four copies of a cluster of genes termed mce operons. Despite extensive research that has demonstrated the importance of these operons on infection outcome, their physiological function remains obscure. Expanding databases of complete microbial genome sequences facilitate a comparative genomic approach that can provide valuable insight into the role of uncharacterized proteins. RESULTS: The M. tuberculosis mce loci each include two yrbE and six mce genes, which have homology to ABC transporter permeases and substrate-binding proteins, respectively. Operons with an identical structure were identified in all Mycobacterium species examined, as well as in five other Actinomycetales genera. Some of the Actinomycetales mce operons include an mkl gene, which encodes an ATPase resembling those of ABC uptake transporters. The phylogenetic profile of Mkl orthologs exactly matched that of the Mce and YrbE proteins. Through topology and motif analyses of YrbE homologs, we identified a region within the penultimate cytoplasmic loop that may serve as the site of interaction with the putative cognate Mkl ATPase. Homologs of the exported proteins encoded adjacent to the M. tuberculosis mce operons were detected in a conserved chromosomal location downstream of the majority of Actinomycetales operons. Operons containing linked mkl, yrbE and mce genes, resembling the classic organization of an ABC importer, were found to be common in Gram-negative bacteria and appear to be associated with changes in properties of the cell surface. CONCLUSION: Evidence presented suggests that the mce operons of Actinomycetales species and related operons in Gram-negative bacteria encode a subfamily of ABC uptake transporters with a possible role in remodeling the cell envelope.

Enhanced mortality despite control of lung infection in mice aerogenically infected with a Mycobacterium tuberculosis mce1 operon mutant.

Mycobacterium tuberculosis causes a variety of host clinical outcomes. We previously showed that M. tuberculosis disrupted in an operon called mce1 proliferates unchecked in BALB/c mouse lungs. The observed outcome could be attributed either to the mutant bacterial burden or to the host immunopathologic response. To differentiate these possibilities, we studied the outcomes of infection in a mouse strain (C57BL/6) less susceptible to M. tuberculosis than BALB/c. We found that the mutant infection reached a plateau in the lungs at a rate similar to that of the wild type. All mice infected with the mutant, but only half of the groups of mice infected with the wild type or complemented strain, died by 40 weeks (p<0.05). At 12-21 weeks of infection, histological examination of the lungs of mice infected with the mutant showed a diffuse pattern of lymphocyte infiltration, while that of mice infected with the other strains exhibited a nodular cellular infiltration pattern. Surprisingly, the number of bacilli recovered from the lungs was similar in all three groups. These observations suggest that rather than the bacterial burden, products of the mce1 operon may directly or indirectly modulate the host immune response that is protective to both the tubercle bacilli and the host.

Combination of amikacin and doxycycline against multidrug-resistant and extensively drug-resistant tuberculosis.

The objective of this study was to assess the activity of amikacin in combination with doxycycline against clinical strains of Mycobacterium tuberculosis in the search for new strategies against multidrug-resistant (MDR) and extensively drug-resistant (XDR) tuberculosis. The study included 28 clinical M. tuberculosis strains, comprising 5 fully susceptible, 1 isoniazid-resistant, 17 MDR, 1 poly-resistant (streptomycin/isoniazid), 1 rifampicin-resistant and 3 XDR isolates, as well as the laboratory strain M. tuberculosis H37Rv. Minimum inhibitory concentrations (MICs) were determined using a modified chequerboard methodology in a BACTEC™ MGIT™ 960 System. Fractional inhibitory concentration indices (FICIs) were calculated, and synergy, indifference or antagonism was assessed. Whole-genome sequencing was performed to investigate the genetic basis of synergy, indifference or antagonism. The MIC50 and MIC90 values (MICs that inhibit 50% and 90% of the isolates, respectively) were, respectively, 0.5 mg/L and 1.0 mg/L for amikacin and 8 mg/L and 16 mg/L for doxycycline. The combination of amikacin and doxycycline showed a synergistic effect in 18 of the 29 strains tested and indifference in 11 strains. Antagonism was not observed. A streptomycin resistance mutation (K43R) was associated with indifference. In conclusion, the benefit of addition of doxycycline to an amikacin-containing regimen should be explored since in vitro results in this study indicate either synergy or indifference. Moreover, doxycycline also has immunomodulatory effects.

Beijing clades of Mycobacterium tuberculosis are associated with differential survival in HIV-negative Russian patients.

We conducted a prospective study to establish factors associated with survival in tuberculosis patients in Russia including social, clinical and pathogen-related genetic parameters. Specifically we wished to determine whether different strains/clades of the Beijing lineage exerted a differential effect of survival. HIV-negative culture-confirmed cases were recruited during 2008-2010 across Samara Oblast and censored in December 2011. Molecular characterization was performed by a combination of spoligotyping, multilocus VNTR typing and whole genome sequencing (WGS). We analyzed 2602 strains and detected a high prevalence of Beijing family (n=1933; 74%) represented largely by two highly homogenous dominant clades A (n=794) and B (n=402) and non-A/non-B (n=737). Multivariable analysis of 1366 patients with full clinical and genotyping data showed that multi- and extensive drug resistance (HR=1.86; 95%CI: 1.52, 2.28 and HR=2.19; 95%CI: 1.55, 3.11) had the largest impact on survival. In addition older age, extensive lung damage, shortness of breath, treatment in the past and alcohol abuse reduced survival time. After adjustment for clinical and demographic predictors there was evidence that clades A and B combined were associated with poorer survival than other Beijing strains (HR=0.48; 95%CI 0.34, 0.67). All other pathogen-related factors (polymorphisms in genes plcA, plcB, plcC, lipR, dosT and pks15/1) had no effect on survival. In conclusion, drug resistance exerted the greatest effect on survival of TB patients. Nevertheless we provide evidence for the independent biological effect on survival of different Beijing family strains even within the same defined geographical population. Better understanding of the role of different strain factors in active disease and their influence on outcome is essential.

Systematic review, meta-analysis and economic modelling of molecular diagnostic tests for antibiotic resistance in tuberculosis.

BACKGROUND: Drug-resistant tuberculosis (TB), especially multidrug-resistant (MDR, resistance to rifampicin and isoniazid) disease, is associated with a worse patient outcome. Drug resistance diagnosed using microbiological culture takes days to weeks, as TB bacteria grow slowly. Rapid molecular tests for drug resistance detection (1 day) are commercially available and may promote faster initiation of appropriate treatment. OBJECTIVES: To (1) conduct a systematic review of evidence regarding diagnostic accuracy of molecular genetic tests for drug resistance, (2) conduct a health-economic evaluation of screening and diagnostic strategies, including comparison of alternative models of service provision and assessment of the value of targeting rapid testing at high-risk subgroups, and (3) construct a transmission-dynamic mathematical model that translates the estimates of diagnostic accuracy into estimates of clinical impact. REVIEW METHODS AND DATA SOURCES: A standardised search strategy identified relevant studies from EMBASE, PubMed, MEDLINE, Bioscience Information Service (BIOSIS), System for Information on Grey Literature in Europe Social Policy & Practice (SIGLE) and Web of Science, published between 1 January 2000 and 15 August 2013. Additional 'grey' sources were included. Quality was assessed using quality assessment of diagnostic accuracy studies version 2 (QUADAS-2). For each diagnostic strategy and population subgroup, a care pathway was constructed to specify which medical treatments and health services that individuals would receive from presentation to the point where they either did or did not complete TB treatment successfully. A total cost was estimated from a health service perspective for each care pathway, and the health impact was estimated in terms of the mean discounted quality-adjusted life-years (QALYs) lost as a result of disease and treatment. Costs and QALYs were both discounted at 3.5% per year. An integrated transmission-dynamic and economic model was used to evaluate the cost-effectiveness of introducing rapid molecular testing (in addition to culture and drug sensitivity testing). Probabilistic sensitivity analysis was performed to evaluate the impact on cost-effectiveness of diagnostic and treatment time delays, diagnosis and treatment costs, and associated QALYs. RESULTS: A total of 8922 titles and abstracts were identified, with 557 papers being potentially eligible. Of these, 56 studies contained sufficient test information for analysis. All three commercial tests performed well when detecting drug resistance in clinical samples, although with evidence of heterogeneity between studies. Pooled sensitivity for GenoType® MTBDRplus (Hain Lifescience, Nehren, Germany) (isoniazid and rifampicin resistance), INNO-LiPA Rif.TB® (Fujirebio Europe, Ghent, Belgium) (rifampicin resistance) and Xpert® MTB/RIF (Cepheid Inc., Sunnyvale, CA, USA) (rifampicin resistance) was 83.4%, 94.6%, 95.4% and 96.8%, respectively; equivalent pooled specificity was 99.6%, 98.2%, 99.7% and 98.4%, respectively. Results of the transmission model suggest that all of the rapid assays considered here, if added to the current diagnostic pathway, would be cost-saving and achieve a reduction in expected QALY loss compared with current practice. GenoType MTBDRplus appeared to be the most cost-effective of the rapid tests in the South Asian population, although results were similar for GeneXpert. In all other scenarios GeneXpert appeared to be the most cost-effective strategy. CONCLUSIONS: Rapid molecular tests for rifampicin and isoniazid resistance were sensitive and specific. They may also be cost-effective when added to culture drug susceptibility testing in the UK. There is global interest in point-of-care testing and further work is needed to review the performance of emerging tests and the wider health-economic impact of decentralised testing in clinics and primary care, as well as non-health-care settings, such as shelters and prisons. STUDY REGISTRATION: This study is registered as PROSPERO CRD42011001537. FUNDING: The National Institute for Health Research Health Technology Assessment programme.

Evolution and transmission of drug-resistant tuberculosis in a Russian population.

The molecular mechanisms determining the transmissibility and prevalence of drug-resistant tuberculosis in a population were investigated through whole-genome sequencing of 1,000 prospectively obtained patient isolates from Russia. Two-thirds belonged to the Beijing lineage, which was dominated by two homogeneous clades. Multidrug-resistant (MDR) genotypes were found in 48% of isolates overall and in 87% of the major clades. The most common rpoB mutation was associated with fitness-compensatory mutations in rpoA or rpoC, and a new intragenic compensatory substitution was identified. The proportion of MDR cases with extensively drug-resistant (XDR) tuberculosis was 16% overall, with 65% of MDR isolates harboring eis mutations, selected by kanamycin therapy, which may drive the expansion of strains with enhanced virulence. The combination of drug resistance and compensatory mutations displayed by the major clades confers clinical resistance without compromising fitness and transmissibility, showing that, in addition to weaknesses in the tuberculosis control program, biological factors drive the persistence and spread of MDR and XDR tuberculosis in Russia and beyond.