Exciting Developments Towards Non-Sputum Based Diagnosis of TB · Sandlund et al DMID 2018 HIV-...
Transcript of Exciting Developments Towards Non-Sputum Based Diagnosis of TB · Sandlund et al DMID 2018 HIV-...
Exciting Developments Towards Non-Sputum Based Diagnosis of TB
Niaz Banaei MDProfessor of Pathology and Medicine
Stanford University [email protected]
Active TB
Latent Infection
10 million cases estimated
1/4 of world’s population
Global Burden of M. tuberculosis
WHO 2018
Active TB
Latent Infection
10 million cases estimated6.4 million reported3.6 million diagnostic gap
1/4 of world’s population
Global Burden of M. tuberculosis
WHO 2018
Contributors to Global Diagnostic Gap
Pediatric ExtrapulmonaryHIV/AIDSUnproductive
1 million 0.9 million
Active TB
Latent Infection
Incipient TB
10 million cases estimated6.4 million reported3.6 million global gap
1/4 of world’s population
Global Burden of M. tuberculosis
WHO 2018
Recent and Projected Trends in Global TB Incidence Cases
Dye et al Annu Rev Public Health 2013
Active TB
Latent Infection and
Incipient TB
Sputum: Culture & NAATBlood: IGRA, RNA signature, biomarkersUrine: biomarkers (LAM)Breath: biomarkers
Blood: IGRA, RNA signature
Diagnostic Tools for M. tuberculosis
Active TB
Latent Infection and
Incipient TB
Blood: IGRA
Blood: IGRA
Diagnostic Tools for M. tuberculosis
‘A 21st Century Solution for Latent TB Detection’
IGRAs entered the scene with a lot of promise
More sensitive and specific than TSTMore reproducible/objective
More predictive
Active TB
Sensitivity and Predictive Value of IGRAs
Pai et al. Clin Micro Rev 2014
Diel et al. Chest 2010
Sensitivity: 80-90s%Specificity: low due to LTBI
Latent Infection
Sensitivity and Predictive Value of IGRAs
Pai et al. Clin Micro Rev 2014
Diel et al. Chest 2010
Sensitivity: 40-70%Specificity: >90%Predictive value: <10%
Incipient TB
Sensitivity and Predictive Value of IGRAs
Pai et al. Clin Micro Rev 2014
Diel et al. Chest 2010
Sensitivity: 40-70%Specificity: low due to LTBIPredictive value: <10%
Active TB
Latent Infection
Incipient TB
Sensitivity and Predictive Value of IGRAs
Pai et al. Clin Micro Rev 2014
Diel et al. Chest 2010
Sensitivity: 80-90s%Specificity: low due to LTBI
Sensitivity: 40-70%Specificity: >90%Predictive value: <10%
Sensitivity: 40-70%Specificity: low due to LTBIPredictive value: <10%
‘A 21st Century Solution for Latent TB Detection’
IGRAs entered the scene with a lot of promise
More sensitive and specific than TSTMore reproducible/objective
More predictive
TB Ag Tube 1 (TB1): ESAT-6 and CFP-10 peptides for CD4 T Cells
TB Ag Tube 2 (TB2): ESAT-6 and CFP-10 peptides for CD4 and CD8
T Cells
QuantiFERON®-TB Gold Plus
TB1 +
TB2 +
+- +
-
-
-
• Interpretation of QFT-Plus using
manufacturer’s interpretation
Interpretation of QFT-Plus Results
• Evidence for role of CD8+ T cells in TB immunity
• IFN- positive Mtb-specific CD8+ T cells
– Associated with recent exposure to TB
– More frequently detected in active TB vs. latent infection
– Mycobacterial burden-dependent
– Detectable in active TB subjects with HIV co-infection
and young children
– Decline after anti-tuberculosis treatment
Why Target CD8 T Cells in QFT-Plus?
Does QFT-Plus have a higher sensitivity
than QFT-GIT for recent exposure to Mtb?
Barcellini et al ERJ 2016
Study Design
QFT-Plus vs. QFT-GIT
Prospective contact screening
Location: Milan, Italy
Contacts: Tested 119 adults with newly positive TST (≥5mm)
Included immunocompromised (9%)
Retested 10-12 weeks if negative
• QFT-Plus: 57.1% (68/119) vs. QFT-GIT: 47.1% (56/119)
QuantiFERON-TB Gold Plus performance in a high-risk population in the United States: a comparison with QuantiFERON-TB Gold In-tube, tuberculin skin test and T-SPOT.TBCDC Tuberculosis Epidemiologic Studies Consortium II (TBESCII)
Study Design
QFT-Plus vs. QFT-GIT
Prospective cross-sectional study
Location: 12 clinics in the U.S.
Contacts: Tested 508 individuals (adult and children) at high
risk for LTBI
Venkatappa et al JCM 2019
94%
99%
98%
Agree-
ment
QuantiFERON-TB Gold Plus performance in a high-risk population in the United States: a comparison with QuantiFERON-TB Gold In-tube, tuberculin skin test and T-SPOT.TBCDC Tuberculosis Epidemiologic Studies Consortium II (TBESCII)
Venkatappa et al JCM 2019
In 44 children with household
TB exposure, agreement
between the QFT-GIT and
QFT-Plus was 96%.
Does QFT-Plus have a higher sensitivity
than QFT-GIT for recent exposure to Mtb?
Answer: No. Differences observed are
due to antigen formulation and not
immunological response.
Does QFT-Plus have a higher sensitivity
than QFT-GIT in patients with active TB?
Does QFT-Plus have a higher sensitivity
than QFT-GIT in patients with active TB?
QFT-GIT QFT-Plus P value TB Ag TB1 TB2 QFT-GIT QFT-Plus
Petruccioli et al
Tuberculosis 2017Italy Adult ≤7 days 0% 69 88% 90% <0.05 2.6 1.9 2.5 19 100% 100%
24 96% 96% >0.05
33 clin 85% 85% >0.05
Yi et al Sci Rep 2016 Japan Adult ≤14 days 4% 162 96% 96% >0.05 4.23 2.36* 2.85* 212 99% 97%
Horne et al IJTLD 2018 US & Japan Adult ≤14 days 2% 164 94% 93% >0.05 4.45 3.07* 3.56** ND --- ---
5 100% 100% >0.05 ND --- ---
7 clin 25% 25% >0.05 ND --- ---
87%4.67
Controls
77Germany 90%3.1* 3.7*
Sensitivity Specificity Country Age Abx Tx HIV+/ICH CasesMedian or Mean (IU/ml)
Hoffmann et al
Clin Microb Inf 2016Adult
Not
provided5%
Study
Kay et al AJTMH 2019 Eswatini Pediatric 0 days 42%Not
provided
Not
provided
Not
provided
Does QFT-Plus have a higher sensitivity
than QFT-GIT in HIV+ patients with active
TB?
Study Design
Single arm
Prospective
Location: Zambia
TB Patients: Smear+ or Xpert+
68 HIV+
<3 days of anti-TB therapy
HIV-
HIV+
Prior study on QFT-GIT
Sensitivity: 63%
Active TB
Incipient TB
Blood: RNA signature
Blood: RNA signature
WHO 2018
Diagnostic Tools for M. tuberculosis
Nature 2010
IFN I/II signalling and complement transcripts up 18 months
before TB diagnosis, while changes in myeloid , lymphoid,
monocyte and neutrophil responses occurred more proximally
Scriba PLoS Pathogens 2017
GBP5 promotes inflammasome assembly
DUSP3 regulator of JNK and ERK signalling
KLF2 anti-inflammatoryLANCET RM 2016
GBP5 promotes inflammasome assembly
DUSP3 regulator of JNK and ERK signalling
KLF2 anti-inflammatoryLANCET RM 2016
Genes comprising the 8
best signatures for
incipient TB
LANCET RM 2020
Diagnostic Accuracy of the 8 Best
Signatures for Incipient TB
LANCET RM 2020
Diagnostic Accuracy of the 8 Best
Signatures for Incipient TB
LANCET RM 2020
LANCET RM 2020
Active TB
Incipient TB
Urine: biomarker (LAM)
WHO 2018
Diagnostic Tools for M. tuberculosis
Urinary LAM Detection with Lateral Flow
(LAM)
Lipoarabinomannan (LAM)
Urinary LAM Detection with Lateral Flow
Alere Determine TB LAM Ag (AlereLAM)
≤100 101-200 >200CD4
Schiller Cochrane 2019
Sensitivity Specificity
≤100 101-200 >200CD4
WHO 2020
Novel Fujifilm SILVAMP TB LAM (FujiLAM)
Lancet Infect Dis 2019
Lancet Infect Dis 2019
Inpatients with HIV
OFID 2020
Inpatients and outpatients with HIV
Sensitivity 95%
Specificity 80%
2017
Active TB
Urine: biomarkersBreath: biomarkers
WHO 2018
Diagnostic Tools for M. tuberculosis
Infectious Diseases Have Metabolic Signatures Need for Technological Innovation to Detect Them
Colorimetric Sensor Array
Urine
Breath
Colorimetric Array for Detection
of Volatile Signatures
• Digitally image before & after exposure & subtract.
After Exposure
ammonia
• Difference Map is a “molecular fingerprint”:
a unique 108-dimensional vector (36 ΔR, ΔG, ΔB).
(center avg.
300 pixels)
Difference Map
|Rafter- Rbefore|,|Gafter- Gbefore|,|Bafter- Bbefore|
• Printed array of chemically responsive dyes.
Before Exposure
Biggest color changesare boxed in gray.
9
mm
Urine collected
Gold StandardCases: Culture+/NAAT+Controls: Culture-/NAAT-
Sputum
Analysis
Difference
Diagnosis of Tuberculosis from Analysis of
Urine Volatile Organic Compounds
Highland Study
Correlation matrix indicates similarities between patients of the same category
TB
N=39
Non-TB
N=25
Sensitivity: 86.9%
Specificity: 87.6%
Lim et al ACS Sensors 2016
Urine collected
Kenya Study at KEMRI
TB: 200 Non-TB: 200
2 hr
Gold StandardCasesGeneXpert+ &or Cx+Controls: GeneXpert- & Cx-
Sputum
Analysis
Difference
Principal Component Analysis Score plots
Sensitivity: 78.3%
Specificity: 69.2%
Sensitivity: 57.9%
Specificity: 63.1%
Sandlund et al DMID 2018
HIV- Group HIV+ Group
Active TB
Blood: cfDNA Urine: cfDNA
WHO 2018
Diagnostic Tools for M. tuberculosis
Application of cf DNA in Diagnostics
Fetal aneuploidy Cancer mutations Organ rejection
Microbial cf DNA
Infectious diseases
- EBVnasopharyngeal CA (Cancer Res 1999)
- Invasive fungal infection (CID 2013)
Biology of Cell-Free DNA
Discovered in 1948
(Mandel & Metais)
Removed by liver
Half-life 10-15min
Apoptosis: 180-1000bp
Necrosis: 10,000bp
Trans-renal
150-200bp
Target Population for cfDNA TB Diagnosis
Pediatric ExtrapulmonaryHIV/AIDSUnproductive
1 million 0.9 million
Accuracy of Plasma cf DNA for TB Diagnosis
IS611065%
(21/33)
93%
(18/19)
gyrB29%
(10/33)
100%
(19/19)Click et al
Scientific
Reports 2018
>18 yo Kenya PTB 43 ND PCR IS611044%
(19/43)ND
Brazil PTB 4974%
(34/49)
USA EPTB 875%
(6/8)
Stanford/GG/
IV>18 yo Uganda PTB 54 40 PCR IS6110
46%
(25/54)
100%
(40/40)
GWiS
Target Specificity
Digital
PCR
Cases Controls
Hogan et al
In
preparationPCR
PCR69
Method Sensitivity
Ushio et al
Tuberculosis
2016
Age
19
CountryTB
Type
53%
(8/15)
100%
(69/69)
33PTB
GWiS100%
(15/15)PTB 15 15
Vietenam
& SA>18 yo
>18 yo
Japan>18 yo
Bilateral vsUnilateral PTB
PTB+EPTB vs. PTB
↑cfDNA ↑cfDNA
Day14 Day0+1484% 94%
Accuracy of Urine cf DNA for TB Diagnosis
Miliary Multifocal LAN Pleural Joint90%(9/10)
67%(16/24)
72%(18/25)
33%(1/3)
45%(5/11)
Accuracy of Urine cf DNA for TB Diagnosis
Accuracy of Urine cf DNA for TB Diagnosis
vsRadiology vsSmear+ vsTTCxP Wk1 Wk12
écfDNA None None écfDNA 9/11Neg*100% with
retesting
Cannas et al
IJTLD 2008>18 yo Italy PTB 43 23
Nested
PCRIS6110
79%
(34/43)
100%
(23/23)
EPTB 82 70%
(57/82)
PTB 2518%
(5/25)
Labugger et al
Infection 2017>18 yo Germany PTB 11 8 PCR IS6110
64%*
(7/11)
100%
(8/8)
Stanford/GG/
IV>18 yo Uganda PTB 75 59 PCR IS6110
32%
(24/75)
98%
(58/59)
Fortun et al
IJTLD 2014
Sensitivity SpecificityTB
TypeTargetCases Method
TMA16S
rRNA
Not
Done>18 yo
CountryAge
0Spain
Controls
Accuracy of Urine cf DNA for TB Diagnosis
Pediatric ExtrapulmonaryHIV/AIDS
Potential of cf DNA in Diagnosis of TB
Cell-free DNA
in Urine
Accuracy
Sensitivity 40%-70%
Specificity ≈100%
Unproductive
Accuracy of Urine cf DNA for TB Diagnosis
Hogan
Preanalytical Variables Impacting
Pathogen cfDNA in Blood and Urine
Murugesan JCM 2019 PMID:31511335
Optimization of Variables Impacting Plasma cfDNA Detection
Murugesan JCM 2019 PMID:31511335
Comparison of Blood Collection Tubes
Murugesan JCM 2019 PMID:31511335
Comparison of Urine Collection Preservatives
Murugesan JCM 2019 PMID:31511335
Comparison of Blood Collection Tubes andUrine Preservatives in Patients with TB
Murugesan JCM 2019 PMID:31511335
Optimization of Variables Impacting Plasma cfDNA Detection
Murugesan JCM 2019 PMID:31511335
Comparison of Plasma Processing Delay
Murugesan JCM 2019 PMID:31511335
Comparison of Urine Processing Delay
Murugesan JCM 2019 PMID:31511335
Optimization of Variables Impacting Plasma cfDNA Detection
Murugesan JCM 2019 PMID:31511335
Comparison of 1 spin vs 2 spin Plasma Collection
Murugesan JCM 2019 PMID:31511335
Comparison of Whole vs 1 Spin Urine Processing
Murugesan JCM 2019 PMID:31511335
Optimization of Variables Impacting Plasma cfDNA Detection
Murugesan JCM 2019 PMID:31511335
Comparison of Fresh vs. Frozen Plasma
Murugesan JCM 2019 PMID:31511335
Comparison of Fresh vs. Frozen Urine
Murugesan JCM 2019 PMID:31511335
Optimization of Variables Impacting Plasma cfDNA Detection
Murugesan JCM 2019 PMID:31511335
Comparison of Plasma and Urine Volume
Preanalytical Variables Impacting
Pathogen cfDNA in Blood and Urine
Higher volume
more sensitive
EDTA is adequate
Murugesan JCM 2019 PMID:31511335
Up to 24 hr delay
is adequate
1 spin is adequate for plasma
No spin needed for urine
Freeze thaw
has no impact
Promega
OmegaQiagen
Thermo
Plasma
ccfDNA extraction
DNA spiking
Sample collection
tube
Processing delay
Centrifugation
Sample storage
Promega &Maxwell
KingFisher &Omega
KingFisher &Thermo
QiaSymphony &Qiagen
Optimization of Variables Impacting Plasma cfDNA Detection
Mtb cfDNA Testing is Promising
Optimize extraction of cfDNA from plasma and urine
Short cfDNA fragments
At lower limit of detection
Assess sensitivity of TB cf-DNA using optimized pre-
analytics in adult and pediatric cohorts
Develop sample-to-answer assay
GeneXpert Ultra
Active TB
Latent Infection and
Incipient TB
Sputum: Culture & NAATBlood: IGRA, RNA signature, biomarkersUrine: biomarkers (LAM)Breath: biomarkers
Blood: IGRA, RNA signature
Diagnostic Tools for M. tuberculosis
Acknowledgements
Stanford University
Kanagavel Murugesan
Rajiv Gaur
Fiona Senchyna
Hee-Won Moon
Catherine Hogan
Jason Andrews
Juan Santiago
Nobuyuki Futai
Global Good/IVL
UCSF/Uganda
Adithya Cattamanchi
Uganda Team
Stellenbosch University
Grant Theron
Funding
Stanford Global Health
Global Good/IVL, ChEM-H