- Main navigation!
- Patricia Untermans San Francisco Food Lovers Pocket Guide, Second Edition: Includes the East Bay, Marin, Wine Country.
- Supplemental Content.
- A Divine and Supernatural Light, Immediately Imparted to the Soul By the Spirit of God, Shown to Both, A Scriptural and Rational Doctrine, Revised Edition (With Active Table of Contents).
Fetching bibliography My Bibliography Add to Bibliography. Generate a file for use with external citation management software. Create File. Clin Infect Dis. Free PMC Article. Images from this publication. See all images 4 Free text. Figure 1. Unsigned sums of 3 to 5 transcripts met the specificity threshold for a screening test on the training data. This process resulted in a set of nested signatures. Only these signatures were used to determine our final set of signatures. Figure 2. Performance of candidate 3—5 transcript signatures on the independent test dataset. We examined the performance of each of these signatures on an independent test dataset by generating receiver operator characteristic curves.
A novel whole-blood miRNA signature for a rapid diagnosis of pulmonary tuberculosis
Figure 3. Comparison of performance of novel, 5-transcript TB screening signature with published signatures and CRP. We calculated receiver operator characteristic curves for our novel, 5-transcript signature; 4 previously-published TB gene-expression signatures; and the non-specific marker of inflammation, CRP, on our independent test dataset.
Figure 4. Performance of novel, 5-transcript, TB screening signature on 2 external TB diagnosis cohorts. Since investigation for recurrent TB entailed 3-monthly induced sputum it is likely that many cases were diagnosed during subclinical stages of disease. To put these findings into perspective, we benchmarked them against the WHO target product profile TPP for a TB risk test World Health Organization [WHO], a , which should ideally be able to predict progression to TB within 2 years and provide a quantitative result that correlates with risk of progression.
- Green Fashion: Ökologische Nachhaltigkeit in der Bekleidungsindustrie (German Edition).
- Blood signature could improve early TB diagnosis -- ScienceDaily!
- Key Points.
- CRASH COURSE: A Head of School, His Sons Addiction, and Lessons for Schools and Families;
- The Future of Drug Discovery: Who Decides Which Diseases to Treat?.
Signature scores at TB diagnosis were high in the majority of cases and decreased significantly at the end of the intensive phase, at month two Figure 3A. This was also reflected by better discrimination between TB treatment initiation and end of treatment samples AUC 0. We next determined if patients could be classified into early and late converters based on their culture conversion status at 2 months of treatment.
No marked differences were observed in kinetics of median signature scores during treatment between early and late converters, suggesting that the observed decrease in COR signature scores after treatment initiation was not an accurate predictor of early culture conversion in this study cohort Figure 3D. B Differences in signature scores at 2 months or end of treatment relative to treatment baseline.
Red dots represent individuals with sputum culture conversion after 2 months late converters and blue dots individuals with culture conversion before 2 months early converters. Horizontal lines represent medians and error bars the inter-quartile ranges IQR.
D Longitudinal kinetics of median signature scores during TB treatment in early and late converters. Error bars represent the IQR.
AI Analyzes Blood Sample "Bio-signature" To Identify Brain Cancer | Technology Networks
We first determined if time to sputum culture positivity at treatment baseline, a measure of bacterial load, was associated with early culture conversion at 2 months. No difference in time to culture positivity between early and late converters was observed at TB diagnosis Figure 4A. However, at the end of the intensive phase of treatment month two the signature could differentiate between early and late converters AUC 0.
Furthermore, time to culture conversion was not correlated with signature scores at treatment baseline Figure 4D and month two of TB treatment Figure 4E. C gene ACS COR signature scores for early and late sputum culture converters at baseline, month two or end of treatment. D,E Association between signature scores and time to liquid culture positivity at baseline D and 2 months after treatment start E.
It was noted that signature scores remained relatively high at the end of treatment, when all individuals had achieved clinical cure. We therefore sought to determine if gene expression would decrease further after clinical cure on the basis that residual inflammation may resolve over time. A comparison of signature scores at the end of treatment and a time point at least 6 months later showed no significant changes Supplementary Figures S2A,B ; late and early converters were also not different at these late follow-up time points Supplementary Figures S2C,D.
To address this, we compared the diagnostic performance of the signature for discriminating between TB cases and Mtb -infected controls in HIV-infected 20 cases and 20 controls and HIV-uninfected 30 cases and 30 controls adults from the CTBC. These figures suggest that fewer HIV-infected people would have to be screened with the COR signature to detect a TB case compared with the current methods of symptom screening or symptom screening plus chest radiography Getahun et al. Horizontal lines represent medians, boxes represent the IQR and whiskers represent ranges.
Performance of the COR signature as a triage test in HIV-infected and uninfected persons based on results from our diagnostic cohort. In light of our results above, we hypothesized that signature scores are associated with pVL. Horizontal lines depict medians and error bars IQRs.
Only 27 samples had detectable pVL, precluding analysis of signature performance in this subset.
We therefore compared prognostic performance in all samples within 6 months of recurrent TB diagnosis or only those in this time window with undetectable pVL. Prognostic performance of the signature for recurrent TB disease was equivalent when all samples or only those with undetectable pVL were included 0—3 months before TB, AUC of 0. Blood transcriptomic signatures of TB risk show promise as non-sputum triage tests for TB, identification of individuals with incipient or subclinical disease who are at high risk of developing active TB disease prognostic performance and as tools for monitoring treatment outcome.
However, prognostic performance of risk signatures in HIV-infected individuals who are at very high risk of TB, has not been specifically and comprehensively assessed. We found that among HIV-infected individuals on ART the transcriptomic COR signature had limited prognostic value and only significantly differentiated between recurrent TB progressors and non-progressing controls within 3 months of disease diagnosis. This is markedly inferior to prognostic performance in the HIV-uninfected adolescent study, where the ACS COR signature differentiated significantly between progressors and non-progressors 1—1.
These results, as well as broader transcriptomic, proteomic and cellular analyses of the adolescent progressors Scriba et al. Since investigation for TB in the TRuTH cohort was based on sputum induction at every visit on every participant, we propose that the resultant early detection of disease during subclinical stages may have impeded the prognostic performance within the confines of the TRuTH study. Detection of subclinical disease is not unexpected since prevalence of asymptomatic TB, defined as disease without clinical TB related symptoms but with abnormalities that can be detected with radiologic or microbiological tests, is generally high in studies that perform active case finding in high-risk populations Drain et al.
One other study, by Sloot et al. Unfortunately this 2-gene signature was not externally validated and thus the true performance of this highly parsimonious signature in an external cohort is currently unknown. Provision of isoniazid prophylaxis to HIV-infected individuals typically decreases the risk of progression to active TB disease, but studies in endemic settings have shown that TB incidence increases shortly after cessation of isoniazid prophylaxis Golub et al. In the TRuTH study 20 progressors and 37 non-progressors received 6 months of isoniazid during follow-up Maharaj et al.
We did not adjust the time-at-risk in our analyses to account for effects on INH prophylaxis, but it was noteworthy that most progressors who developed recurrent TB disease were diagnosed within a few months median of days of stopping INH prophylaxis Maharaj et al. The effect of isoniazid prophylaxis may also have contributed to the poor prognostic performance of the ACS COR signature.
Taken together, these findings highlight the importance of clear guidelines on appropriate screening and treatment strategies for latent infection and subclinical disease in HIV-infected persons. We and others have shown that transcriptomic signatures may be useful to monitor treatment outcome in HIV-uninfected persons Berry et al.
However, signature discrimination between those with sputum culture conversion at 2 months of treatment was only possible at the 2-month time point, not at treatment initiation or end of treatment. This finding is consistent with the observation of high transcriptomic COR signature scores at the end of TB treatment and up to a year after successful cure in the HIV-uninfected Catalysis for Health Foundation study participants Thompson et al. These data, supported by persistent pulmonary inflammation detected on PET-CT and detection of mycobacterial RNA in bronchoalveolar lavage samples, are suggestive of residual disease or mycobacterial survival after successful cure Malherbe et al.
Since the IMPRESS trial was performed in a high Mtb -transmission setting we cannot rule out that reinfection after cure may also have contributed to elevation of signature scores. Taken together, our findings show that diagnostic, prognostic and treatment monitoring utility of the gene transcriptomic COR signature were reduced in these HIV-infected cohorts compared with studies in HIV-uninfected individuals. Nevertheless, this is in line with other studies, which also report reduced diagnostic performance of largely ISG-based transcriptomic signatures for TB disease in HIV-infected populations, relative to HIV-uninfected persons Kaforou et al.
Our results suggest that the reduction in signature performance in the HIV-infected cohort is attributable to higher signature scores in those without incipient, subclinical or active disease.
Regardless, diagnostic performance was still superior to symptom screening alone or symptom screening in parallel with chest radiography in HIV-infected persons Getahun et al. Several other studies have also reported diagnostic performance of transcriptomic signatures in HIV-infected cohorts, including a large signature based on transcripts Dawany et al. We were not able to perform direct comparisons of the performance of the our gene transcriptomic COR signature with these because this would have to be done in the same cohort.
This highlights the importance of directly comparing the multiple published transcriptomic signatures directly in diverse cohorts. Notably, a very recent study performed such a systematic comparison of the diagnostic performance of 16 published transcriptomic signatures in 24 publicly available microarray or RNA-sequencing datasets and reported that a 3-gene signature performed best Warsinske et al.
It will be important to perform prospective clinical validation of such signatures in future studies and also to assess prognostic performance of these signatures for incident TB. Our data suggest that transcriptomic signatures for TB that are discovered or parameterized in HIV-infected cohorts may yield better diagnostic performance in HIV-infected individuals than signatures developed in HIV-uninfected populations. Signatures developed by multi cohort analyses that included HIV-infected and uninfected individuals provide clues that such signatures may be more tolerant to the effects of underlying HIV infection [11, Duffy et al.
Alternatively, transcripts known to be elevated in HIV-infected individuals, such as ISGs, may have to be explicitly excluded from transcriptomic signatures for HIV-infected individuals. Such a strategy was employed by Esmail et al. Such a signature might equally perform well in HIV-uninfected persons. It is not well understood if geographic, epidemiological or population genetics differences are likely to impact overall performance and utility of such a signature.
Notably, no modification of the COR signature was made in our study when applying it as a diagnostic, prognostic or treatment monitoring tool. The demonstrated application to cryopreserved PBMC allows broader opportunities for validation in non-African cohorts in the future, if such biobanks are available.
Our analyses were limited by relatively small cohort sizes and were not sufficiently powered to investigate signature performance as a function of pVL quantitative levels. Regardless, our stratification by detectable and undetectable pVL clearly showed a strong effect on signature scores and represents, to the best of our knowledge, a first attempt at delineating effects of HIV load on TB transcriptomic signatures. We also acknowledge that healthy individuals with latent infection are not the ideal control group for establishing diagnostic performance of a triage test.
Future studies that assess diagnostic performance should also include control groups of symptomatic individuals who do not have TB but present with other respiratory diseases. Our results suggest that underlying HIV infection has a marked effect on performance of ISG-based transcriptomic signatures, which requires further investigation in larger studies.
All authors have read and approved the manuscript. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. FD was supported by the Margaret McNamara educational grant for women in developing countries. Horizontal lines represent medians and error bars the IQR. D Signature scores for early and late converters 6—8 months after end of TB treatment.
Horizontal lines represent medians and error bars the IQR and p -values were calculated using the Mann-Whitney U test.