50
Frank Cobelens European Advanced Course in Clinical TB Rotterdam 11-13 November 2019 CAN WE PREDICT TUBERCULOSIS CURE? CURRENT TOOLS AVAILABLE

CAN WE PREDICT TUBERCULOSIS CURE? CURRENT TOOLS … · relapse or re-infection Relapse Recurrent TB disease due to re-emergence of the original infection Re-infection Recurrent TB

  • Upload
    others

  • View
    9

  • Download
    0

Embed Size (px)

Citation preview

Frank CobelensEuropean Advanced Course in Clinical TBRotterdam11-13 November 2019

C A N W E P R E D I C T T U B E RC U LOS I S C U R E?

C U R RE NT TO O L S AVA I L A BL E

www.aighd.org

Conflict of interest

I have the following, real or perceived direct or indirect conflicts of interest that relate to this presentation:

I have been awarded a grant by EDCTP (RIA-2018D-2509) for evaluating prediction tests for TB.

www.aighd.org

Content

Treatment monitoring

Establishing cure

Predicting cure

Individualized treatment

www.aighd.org

Establishing cure

WHY?

To establish that a patient has been cured of TB disease

→ no residual disease/pathology

→ no relapse expected

Idenify patients who need:

→ prolonged treatment

→ posttreatment vaccination

WHEN?

At end of treatment

→ ≈ 6 months post-initiation for first-line treatment

→ depending on cure definition for second-line treatment

www.aighd.org

Predicting cure

WHY?

To predict that a patient will (not) relapse after having completed treatment (relapse-free cure)

→ for the full treatment duration

→ for shortened treatment

Identify patients who:

→ need additional treatment (e.g. adjunct host-directed therapy)

→ can be cured with a shortened treatment regimen

WHEN

At treatment initiation or within the first 2 months of treatment

www.aighd.org

Establishing cure

www.aighd.org

Definitions

Cure Smear- or culture-negative sputum specimens in the last month of treatment and on >=1 previous occasion

Recurrence Repeat occurrence of TB disease in a patient declared cured, as a result of eitherrelapse or re-infection

Relapse Recurrent TB disease due to re-emergence of the original infection

Re-infection Recurrent TB disease as a result of exogenous infection with a M. tuberculosis strainthat is different from the organism that caused the original infection

www.aighd.org

Methods for establishing cure

Sputum culture – limitations (turnaround times, logistics, false-negative rates, BSL requirements )

Sputum smear microscopy (non-viable bacilli) - improve by viability staining?

? Xpert MTB/RIF, Xpert Ultra

Bacterial rRNA (MBLA)

www.aighd.org

Sputum smear microscopy

Standard AFB microscopy detects both viable and nonviable mycobacterial cells

→ use selective dyes to detect only metabolically active and potentially replicating bacilli?

31 non-MDR patients, calculated concentration viable bacilli over first 9 days of treatment

flurorescein diacetate (FDA) viability staining

Datta et al. Clin Infect Dis 2014

www.aighd.org

MTB activity measured by FDA sputum microscopy is inversely associated with infectiousness

Index patients with low flurorescein diacetate (FDA) smear-positive concentrations had highersecondary attack rates than index patients with low FDA smear-positive concentrations

TB incidence among

household contacts, by FDA

status (below or above

median) of index case (B)

Datta et al. J Infect Dis 2017

www.aighd.org

Xpert MTB/RIF is poorly associated with culture results for treatment monitoring…

Friedrich et al. Lancet Respir Med 2013

2741 sputum specimens obtained from 221 African patients, weeks 0–26 after initiation of first-line TB treatment

www.aighd.org

.. But Xpert cycle threshold (Ct) values provide stronger correlation

Shenai et al. PLoS One 2016

Direct Ct value (A), change in Ct value from pretreatment Xpert (B) and percentage Xpert conversion (“closing”, C)Reference standard = culture 96 drug-susceptible HIV-negative patients on standard first-line treatment

www.aighd.org

Molecular bacterial load assay (MBLA)

Molecular test for detection of viable Mtb

RT-qPCR that quantifies Mtb load frompatient sputum using the16S ribosomal RNA (16S rRNA) as a reference gene

148 sputum samples from 20 patients in Tanzania collected pretreatment andlongitudinally during standard first-line treatment

Comparator: solid culture CFUs and TTP

Honeyborne et al. J Clin Microbiol 2014

www.aighd.org

Issues with current definition of cure

1. Does it adequately reflect that there is no residual disease/pathology?

2. Is it relevant with regard to relapse – i.e. does it reflect relapse-free cure?

Are 1 and 2 different?

More specifically:

Does 2 imply that the test must exclude the presence of residual persistent/latent infection?

Or is relapse “flare up” of residual active disease?

www.aighd.org

Conceptual models of relapse

15

Relapse is reactivation of incompletely sterilized “latency” following cure

Relapse is residual “active” disease

www.aighd.org

TB infection and disease: changing paradigm

Barry et al. Nat Rev Microbiol 2009

www.aighd.org

TB infection shows a continuum from elimination to overt “clinical” disease

Drain et al. Clin Microbiol Rev 2018

www.aighd.org

Schematic of continuum LTBI – TB disease

Esmail et al. Phil Trans Roy Soc 2015

www.aighd.org

In TB prevalence surveys 40% or more of pulmonary TB cases are asymptomatic

Proportion screened wo had no TB symptoms -- out of all smear-positive cases and -- out of all bacteriologically positive cases identified in TB prevalence surveys from Asia

Onozaki et al. Trop Med Internat Health 2015

www.aighd.org

PET-scan activity after experimental infection predicts disease in non-human primates

Total Lung FDG Avidity Measurements at Six Months Post

Infection Highlighting Differences in FDG Uptake in Animals

Which Reactivate from Latent Infection

Serial FDG PET/CT Images Showing a

Disseminating and Stable Pattern of Granuloma

Evolution During the Course of Early Infection

White et al. J Vis Exp 2017

www.aighd.org

Inflammatory lesions may exist well before onset of disease in humans

35 patients with LTBI (QFN-GIT+, culture -), HIV infected, ART naive (CD4>350)PET-scans, 6 months follow-up

→ 10 patients with subclinical disease more likely to progress to active disease

2-deoxy-2-[18F]fluoro-d-glucose positron emission and computed tomography

Esmail et al. Nature Med 2016

www.aighd.org

Transcriptomic and immunological changes mayexist up to 12 months before onset of disease

16-gene signature (ACS-COR)

Zak et al. Lancet 2016; Suliman et al. Am J Resp Crit Care Med 2018Scriba et al. Plos Pathogens 2017

Changes in proportions of peripheralblood cell subsets during progressionfrom infection to TB disease in ACS cohort

www.aighd.org

So…

❑ Overt clinical TB disease may be preceded by a prolonged subclinical period of “incipient” TB in which there

are no/few symptoms and sputum cultures may be positive

❑ This subclinical phase may last several months

❑ Progression of subclinical to clinical disease may be halted spontaneously

Would this also apply to relapse?

www.aighd.org

Almost all relapse occurs within 12 months post-cure

Of all relapses occurring within 2 years after completion of treatment, 78% occur within the first 6 months, and 91% within the first 12 months

15 clinical trials of first-line TB regimens (British MRC)24 months active follow-up for recurrent TB (no genotyping)

Nunn. IJTLD 2010

www.aighd.org

Inflammatory lesions often still exist despite ‘bacteriological cure’

99 HIV- pulmonary TB patients on first-line treatment

PET-scans at start and end of treatment (6 months)

Patterns:

a. Resolved (n=14)

b. Improved but still abnormal (n=51)

c. Mixed – new lesions or increased intensity (n=34)

2-deoxy-2-[18F]fluoro-d-glucose positron emission and computed tomography

Malherbe et al. Nature Med 2016

www.aighd.org

Patients with bacteriological cure had mycobacterial mRNA in sputum and BAL

mRNA on BAL 19 TB specific targets – PCA

15 TB patients <3 months after end of Tx

+ 9 contol patients (other lung disease)

+ 5 patients with newly diagnosed PTB

mRNA on sputum

13 TB specific targets – PCA

75 TB patients at end of treatment (6M)

+ 20 community controls

+ 5 patients with other lung diseases

Malherbe et al. Nature Med 2016

www.aighd.org

Implication

... but here (incipient TB)……are not here…

At treatment completion, the majority of patients who are “cured” but eventually relapse …

www.aighd.org

Predicting cure

www.aighd.org

Only a minority of patients needs 6 months of treatment

Data from British MRC TB treatment trials in East and central Africa

Fox. IJTLD 1999

www.aighd.org

4-month fluoroquinolone-based regimens

OFLOTUB REMoxTB RIFAQUIN

Short

regimen

2HRGZ/2HRG 2HRMZ/2HRM

2MREZ/2MR 2MREZ/2MR

Control 2HREZ/4HR 2HREZ/4HR 2HREZ/4HR

N 1836 1674 827

Merle et al.

NEJM 2014

Gillespie et al.

NEJM 2014

Jindani et al.

NEJM 2014

Can we identify who needs only 2-4 months of treatment?

www.aighd.org

Approaches to identifying low/high relapse-risk patients

2-months smear/culture conversion

Clinical risk stratification

Immunological biomarkers:

◦ T-cell markers

◦ T-cell responses

◦ Direct TB antigen detection

◦ Monocytic and lymphocytic cell populations

◦ Other

Inflammation and acute phase response markers

Blood transcriptomic signatures

Radiology

Keren et al. Mbio 2011

www.aighd.org

Approaches to identifying low/high relapse-risk patients

2-months smear/culture conversion

Clinical risk stratification

Immunological biomarkers:

◦ T-cell markers

◦ T-cell responses

◦ Direct TB antigen detection

◦ Monocytic and lymphocytic cell populations

◦ Other

Inflammation and acute phase response markers

Blood transcriptomic signatures

Radiology

Keren et al. Mbio 2011

www.aighd.org

2 months culture conversion predicts relapse

2mo:Conversion

6mo:Failure, death

First-line treatment

18mo:Relapse

Correlation between 2-month conversion rate and relapse rate, forvarious treatment regimens and geographic regionsClinical trials only, log scales

Wallis et al. Lancet Infect Dis 2010

Does this depend on the duration of Tx after 2 months?

www.aighd.org

2 months culture conversion predicts relapse independently of duration

By duration or treatment By % positive at 2 months

Meta-regression predicting relapse rate from duration of treatment and 2-month

conversion rate, based on published data from 24 studies of 58 regimens (7793 patients)

Wallis et al. PLoS One 2013

However, not sufficiently predictive to be used for treatment decisions

www.aighd.org

6- but not 2-months sputum conversion predicts treatment outcome in MDR-TB

Time to sputum conversion among MDR-TB patients, N = 1712

18-24 month second-line treatment regimens

ROC curve for predicting treatment success vs

failure or death among MDR-TB patients, N = 1712

Kurbatova et al. Lancet Respir Med 2015

www.aighd.org

Risk stratification: pretreatment clinical information

IPD meta-analysis of OFLOTUB, REMoxTB and RIFAQUINE trial dataPrediction of unfavourable outcomes among the 4-month treatment arms (N = 2001) Pretreatment predictors only

Imperial et al. Nat Med 2018

www.aighd.org

Risk stratification: pretreatment and early treatment clinical information combined

Imperial et al. Nat Med 2018

www.aighd.org

Risk stratification: non-inferiority of 4-month FQ-based regimens

Imperial et al. Nat Med 2018

Application of the prediction model to a validation set from the FQ-based Tx shortening trialsValidation patients were HIV-negative, had non-cavitary disease and were culture-negative at 2 months

Red squares: subgroups in which 4-month treatment is non-inferior to 6-month treatment

www.aighd.org

Potential immune markers for relapse-free cure

Goletti et al. Eur Respir J 2018

Most biomarkers in experimental stage

Prediction endpoint often mixed – not only relapse-free cure

Few independent validations

www.aighd.org

CD8 response as potential treatment monitoring marker

QuantiFERON-TB Gold Plus (QFT-Plus) testing among 38 patients on first-line TB treatment

QFT-Plus Tube 2 (TB2) designed specifically to stimulate both CD8+ T-cells and CD4+ T-cells.

Kamada & Amishima. Eur Respir J 2017

www.aighd.org

Monocyte/lymphocyte (M/L) ratio

Very high and very low M/L ratios predict onset of TB disease among HIV-infected adults starting on cART

Naranbhai et al. J Infect Dis 2014

www.aighd.org

Monocyte/lymphocyte ratio as potential treatment monitoring marker

La Manna et al. PLoS One 2017

M/L ratios in 71 patients with activepulmonary TB and 34 cured TB patients

www.aighd.org

Levels of multiple inflammation/acute phase markers change during successful TB treatment

www.aighd.org

IP-10 in serum and urine

Longitudinal analysis of IP-10 levels in 23 patients with active TB stratified by risk of relapse

Longitudinal (a) serum, (b) urine IP-10, and (c) urine IP-10/creatinine ratio obtained throughout treatment (T0, T2, and T6)

Kim et al. BMC Infect Dis 2018

www.aighd.org

Host transcriptomic (mRNA) blood signatures

Host transcriptomic signatures predict onset of TB over a short (0-6 months) period, probably identifying incipient TB -- inflammatory response

Various signatures behave more or less similar

Gupta et al. BioXriv 2019

www.aighd.org

Host blood RNA signatures show ‘mirror’ pattern for TB prediction and outcome monitoring

16-gene “Correlate of Risk” signature 9-gene optimized “DISEASE” signature

Thompson et al. Tuberculosis 2017

www.aighd.org

Host blood RNA signatures for prediction of treatment failure

DISEASE signature to predict failure

at week 1 (dotted), 4 (dashed) and

24 (solid) of Tx

13-gene FAILURE signature

predicts failure at wk 1 of treatment

No overlap with COR or DISEASE

FAILURE predicts progression to

disease, but less than COR

Thompson et al. Tuberculosis 2017

Promising field

Validation needed in independent cohorts

Geographic differences need to be taken into account

Translation to POC platforms needed

www.aighd.org

Conclusion

Current tools for establishing cure (smear, culture) do not exclude ongoing pathology that may result in relapse

Predicting relapse-free cure in early stages of treatment is important for stratifying patients:

-- those who need less than 6 months treatment

-- those who need more than 6 months treatment or adjunct treatment

Promising options for predicting relapse-free cure include:

◦ Risk-stratification based on pretreatment and early-treatment clinical information

◦ Immunological response marker signatures

◦ Inflammation/acute phase marker signatures

◦ Transcriptional signatures

Validation studies in various patient populations are needed prior to clinical application

Individualized treatment

www.aighd.org

Acknowledgements

Delia Goletti, Tonino Alonzi National Institute for Infectious Diseases “L. Spallanzani”, Rome , Italy

Cecilia Lindestam Arlehamn La Jolla Institute for Allergy and Immunology (LJI), La Jolla CA, USA

Tom Scriba South African Tuberculosis Vaccine Initiative, University of Cape Town, Cape Town, South Africa

Richard Anthony National Institute for Public Health and the Environment (RIVM), Utrecht, The Netherlands

Daniela Cirillo San Raffaele Scientific Institute, HSR, Milan, Italy

Claudia Denkinger FIND, currently University of Heidelberg, Heidelberg, Germany

THANK YOU