Non-invasive methods to assess hepatic fibrosis

1
Non-invasive methods to assess hepatic fibrosis

• MIRELLA FRAQUELLI MD PhD - DARIO CONTE MD
• Gastroenterology and Endoscopy Unit
• Fondazione IRCCS Cà Granda
• Ospedale Maggiore Policlinico Milano
• Department of Pathophysiology and Transplantation
• Università degli Studi Milano, Italy
• Rome June 19, 2015

2
Non - invasive methods to assess hepatic
fibrosis

• Serum markers and clinical scores
• Transient elastography
• Acoustic Radiation Force (ARFI),
• Shear Wave Elastography (SWE)

3
Serum markers to assess hepatic fibrosis
DIRECT They are directly linked to changes
in extracellular matrix turnover
occurring during fibrogenesis
INDIRECT They reflect the derangement of hepatic
function, but not
the extracellular matrix metabolism
4
Serum biomarkers TO ASSESS fibrosis in CLD HCV ( I )
FIBROTEST a-2-macroglobulin gGT, apolipoprotein A1, haptoglobin, bilirubin, age, gender
FORNS INDEX age, platelet count, cholesterol, GGT
AST TO PLATELET RATIO (APRI) AST , platelet count
FIBROSPECT a-2-macroglobulin, hyaluronate, TIMP-1
LOK INDEX platelet count, AST/ALT ratio, INR
MP3 MMP-3, TIMP-1
ELF (Enhanced Liver Fibrosis score) age, hyaluronate, MMP-3, TIMP-1
5
Serum biomarkers TO ASSESS fibrosis in CLD HCV ( II )
GUCI (Gotebörg University Cirrhosis Index) AST, INR , platelet count
VIRAHEP C MODEL AST, platelet count, alkaline phosphatase, age
FIBROINDEX platelet count, AST, gamma-globulin
FIB-4 platelet count, ALT , AST
HALT-C MODEL hyaluronic acid, TIMP-1, platelet count
HEPASCORE bilirubin, GGT, hyaluronate, a-2-macroglobulin, age, gender
FIBROMETERS platelet count, prothrombin index, AST, a-2-macroglobulin, hyaluronate, urea, age
6
Serum biomarkers TO ASSESS fibrosis in CLD HBV

Fibrotest a-2-macroglobulin gGT, apolipoprotein A1, haptoglobin, bilirubin, age, gender
Hui score BMI, bilirubin, platelet count and albumin
Zeng score age, a-2-macroglobulin, hyaluronate and GGT
AST to Platelet Ratio (APRI) AST , platelet count
7
Serum biomarkers in chronic hepatitis Cfor
diagnosing F gt2
Serum Panel Fgt2 () Cut off Sens Spec LR LR- AUROC
Fibrotest 80 gt0.48 75 85 5.0 0.3 0.87
Forns index 26 lt 4.2 gt 6.9 94-31 51-95 1.9-6 0.1-0.7 0.81
APRI 50 lt0.5 gt 1.5 91-41 47-95 1.7-8.2 0.2-0.6 0.80
Fibrospect 52 gt0.36 77 73 2.9 0.3 0.83
Hepascore 57 gt0.5 63 89 5.7 0.4 0.82
Fibrometer 56 NA 80 84 5.0 0.2 0.89
Fibroindex 50 lt1.25 gt2.25 40-30 97-97 10-13.3 0.6-0.2 0.83
Optimal diagnostic test LR gt 10 and LR- lt
0.1
8
Serum biomarkers in chronic hepatitis Cfor
diagnosing F4
Serum Panel Fgt2 () Cut off Sens Spec LR LR- AUROC
Fibrotest 14 gt0.74 63 84 4.0 0.4 0.82
APRI 17 lt 1.0 gt 2.0 89-57 75-93 3.6-8.1 0.1-0.5 0.89
FiB-4 17 lt1.45 gt 3.35 74-38 81-98 3.9-19 0.3-0.6 0.85
Lok index 38 lt 0.2 gt 0.5 98-40 53-99 2.1- 40 0.04-0.6 0.81
NFS 27 1.455 gt 0.676 77-43 97-97 10-13.3 0.6-0.7 0.82
Hepascore 16 gt0.84 71 8.9 6.5 0.3 0.89
HALT-C 38 lt 0.2 gt 0.5 88-47 45-92 1.6-5.9 0.3-0.6 0.81
Optimal diagnostic test LR gt 10 and LR- lt
0.1
9
Serum biomarkers in HBV or HCV CLD
Pros Cons
Good reproducibility Not specific for the liver
High applicability Unable to discriminate intermediate stages of fibrosis
Well validated Cost and limited availability (patented)
Can be performed in outpatients Limitations (hemolysis , inflammation, Gilberts syndrome..)
10
Elastography - Key features
Elastography
Manual compression
Acoustic Radiation Force (ARFI)
Mechanical stimulation No imaging
Absolute quantification (shear wave technology)
Absolute quantification
Qualitative / Relative quantification
STRAIN IMAGING
POINT QUANTIFICATION SHEARWAVE QUANTIFICATION KPa,
m/s
2D SHEARWAVE IMAGING (SWE)
TRANSIENT ELASTOGRAPHY KPa, m/s
11
Elastography - Key features
Elastography
Manual compression
Acoustic Radiation Force (ARFI)
Mechanical stimulation No imaging
Absolute quantification (shear wave technology)
Absolute quantification
Qualitative / Relative quantification
STRAIN IMAGING
POINT QUANTIFICATION SHEARWAVE QUANTIFICATION KPa,
m/s
2D SHEARWAVE IMAGING (SWE)
TRANSIENT ELASTOGRAPHY KPa, m/s
12
TE Physical principles and functioning
Transient elastography (TE) a rapid,
non-invasive technique conceived to evaluate
hepatic fibrosis by measuring liver stiffness.
US transducer 3.5 MHz Vibrator mild amplitude
and low frequency (50 Hz) elastic
waves Propagation speed of elastic waves
directly related to tissue stiffness
13
TE Results
The stiffer the tissue, the faster the shear wave
propagates
14
TE Examination
Explored volume TE 1/500 of the liver total
mass Liver biopsy 1/50000 of the liver total mass
15
TE Reproducibility and normal values
16
Transient elastography - Reproducibility

• 200 consecutive patients with CLD of various
  etiologies were concurrently assessed by TE and
  liver biopsy (800 measurements performed )

Inter-observer agreement 0.98 (95 CI
0.9770.987)
Intra-observer agreement 0.98
Intra-observer agreement 0.98
Fraquelli et al, Gut. 2007
17
TE Normal values
Coperchot Roulot Colombo
Subjects, n 71 Healthy volunteers 429 Medical check-up 327 Blood donors
Mean stiffness (Kpa) 4.8 (2.5-6.9) 5.41.5 4.91.7
95th centile - 8.6 7.8
Age No effect No effect No effect
Gender Mgt F Mgt F Mgt F
High BMI Increased Increased Increased
Metabolic syndrome - Increased -
Fatty liver - - Increased
Del Poggio et al, WJG 2009
18
TE - Applicability and limitations
19
TE - Applicability
Contraindications Pregnant women Patients
with implantable devices (pacemaker or
defibrillator)
Failure of examination (3-5) Ascites
Narrow intercostal spaces
Independent factor for failure
BMI gt28 kg/m2 in 2114 patients
Overweight / obesity
BMI gt30 kg/m2 in 13369 examinations
Foucher et al, Eur J Gastroenterol Hepatol 2006
Castera et al, Hepatology 2010
20
TE - Unreliable results
13369 EXAMINATIONS
Castera et al, Hepatology 2010
21
Features influencing TE

• LIVER CONGESTION
• INFLAMMATION
• EXTRA HEPATIC CHOLESTASIS
• STEATOSIS
• PRIMARY LYMPHANGECTASIA

22
TE Diagnostic performances in the assessment
of liver fibrosis
23
TE - Correlation with liver fibrosis

• TE results significantly correlate with liver
  fibrosis stage correlation coefficients ranging
  from 0.55 to 0.84.

24
TE performance in diagnosing F2 in HCV
Author, yr Etiology Patient Cut-off kPa Sensitivity Specificity -LR LR AUROC
Ziol, 2005 HCV 251 8.6 56 91 0.48 6.6 0.79
Castera, 2005 HCV 183 7.1 67 89 0.37 6.0 0.83
Coco, 2007 HCV-HBV 228 8.3 85 91 0.16 9.4 0.93
Fraquelli, 2007 mixed 200 7.9 72 84 0.30 4.6 0.86
Arena, 2008 HCV 150 7.8 83 82 0.20 4.6 0.91
Lupsor, 2008 HCV 324 7.4 76 84 0.30 4.6 0.86
Degos, 2010 mixed 1307 5.2 90 32 0.30 1.3 0.75
Zarski, 2012 HCV 382 5.2 97 35 0.08 1.4 0.82
Optimal diagnostic test LR gt 10 and LR- lt
0.1
25
TE in diagnosing HCV - cirrhosis
Author, yr Etiology Disease prev () Patient Cut-off, kPa Sens () Spec () -LR LR AUROC
Ziol, 2005 HCV 19 251 14.6 86 96 0.14 23.0 0.87
Castera, 2005 HCV 25 183 12.5 87 91 0.14 9.7 0.95
Ganne-Carrié 2006 Mixed 15 775 14.6 79 95 0.11 15.8 0.95
Coco, 2007 HCV-HBV 20 159 14.0 78 98 0.22 39 0.96
Fraquelli, 2007 Mixed 18 200 12.0 91 89 0.10 8.2 0.90
Arena, 2008 HCV 19 150 14.8 94 92 0.07 11.3 0.98
Lupsor, 2008 HCV 21 324 11.8 87 91 0.14 9.4 0.94
Zarski, 2012 HCV 15 382 12.9 77 90 0.25 7.7 0.93
Optimal diagnostic test LR gt 10 and LR- lt
0.1
26
TE performance in diagnosing F2 in HBV
Author, yr Patient Cut-off kPa Sensitivity Specificity -LR LR AUROC
Coco 2007 228 8.3 85 91 0.2 9.4 0.93
Olivieri 2008 188 7.5 94 88 0.06 7.8 0.97
Marcellin 2009 173 7.2 70 83 0.3 3.4 0.81
Wang 2009 88 8.0 80 77 0.2 3.4 0.86
Degos 2010 284 5.2 89 38 0.5 1.4 0.78
Sporea 2010 140 7.0 65 59 0.6 5.9 0.65
Cordoso 2012 202 7.2 74 88 0.2 6.1 0.87
Goyal 2013 357 6.0 82 67 0.2 3.5 0.84
Optimal diagnostic test LR gt 10 and LR- lt
0.1
27
TE performance in diagnosing F4 in HBV
Author, yr Patient Cut-off kPa Sensitivity Specificity -LR LR AUROC
Coco 2007 228 14.0 78 98 0.2 39 0.96
Olivieri 2008 188 11.8 86 96 0.04 21 0.97
Marcellin 2009 173 11.0 93 87 0.08 7.1 0.93
Wang 2009 88 10.0 85 88 0.1 7.0 0.89
Degos 2010 284 12.9 52 93 0.5 7.4 0.85
Sporea 2010 140 13.6 86 99 0.1 86 0.97
Cordoso 2012 202 11.0 75 90 0.1 7.5 0.93
Goyal 2013 357 9.0 81 90 0.2 8.1 0.93
Optimal diagnostic test LR gt 10 and LR- lt
0.1
28
FIBROSTIC study (France, 20062008). ROC curves
of non-invasive liver fibrosis tests to diagnose
F2 and F4 in 1307 patients
F2
F4
Degos et al, J Hepatol 2010
29
BORDEAUX ALGORITHM
Sequential diagnostic algorithms
AUROC Fgt2 0.91 (0.86-0.96) F4 0.93
(0.90-0.96)
Castera et al. Gastroenterology 2005 Castera et
al. J Hepatol 2010
30
TE Monitoring of disease progression
31
Monitoring disease progression by TE
Foucher et al, Gut 2006
32
Does TE and Fibrotest have a prognostic value in
the context of cirrhosis?
MULTIVARIATE ANALYSIS HR
(95 CI) P value LIVER BIOPSY 2.6
(1.83.8) .0001 TE 44 (15127)
.0001 FIBROTEST 90 (15549)
.0001 APRI 2.8 (1.6-4.7) .0002
Vergniol et al. Gastro 2012
33
Does TE have a prognostic value in the context
of CLD?
All
Cirrhotics
Performances of LS and HVPG for the prediction of
PH-related complications
Robic et al. J Hepatol 2011
34
Elastography - Key features
Elastography
Manual compression
Acoustic Radiation Force (ARFI)
Meccanic stimulation No imaging
Absolute quantification (shear wave technology)
Absolute quantification
Qualitative / Relative quantification
STRAIN IMAGING
POINT QUANTIFICATION SHEARWAVE QUANTIFICATION KPa,
m/s
2D SHEARWAVE IMAGING (SWE)
TRANSIENT ELASTOGRAPHY KPa, m/s
35
PSWE (ElastPQ) - Reproducibility
150 CLD patients ICC 0.87 (95 CI 0.82-
0.90) Learning curve
Baccarin et al. FISMAD 2015
36
Acoustic Radiation Force Impulse (ARFI- PSWE)
Volume measured 10 mm long and 6 wide
Diagnosis of F2
Author, yr Etiology Patient Cut-off m/sec Sensitivity Specificity -LR LR AUROC
Lupsor 2009 HCV 112 1.34 67 92 0.3 9.4 0.86
Friedrich Rust 2012 HBV 92 1.39 50 90 0.5 5.1 0.73
Yoon 2012 mixed 204 1.13 58 84 0.5 3.6 0.74
Sporea 2013 mixed 332 1.34 75 69 0.3 2.4 -
Cassinotto 2014 mixed 349 1.51 72 81 0.3 3.7 0.81
37
Acoustic Radiation Force Impulse (ARFI-PSWE)
Volume measured 10 mm long and 6 wide
Diagnosis of F4
Author, yr Etiology Patient Cut-off m/sec Sensitivity Specificity -LR LR AUROC
Lupsor 2009 HCV 112 2.0 80 95 0.2 17 0.93
Piscaglia 2011 mixed 70 1.7 81 88 0.2 7.0 0.91
Xe 2012 HBV 204 1.88 95 91 0.05 12 0.97
Sporea 2013 mixed 332 1.8 86 77 0.1 3.7 -
Cassinotto 2014 mixed 349 2.2 81 77 0.2 3.5 0.84
38
ARFI vs TE meta-analysis
Cumulative diagnostic estimates
Technique Fibrosis stage Sens Spec LR LR- AUROC
TE Fgt2 0.78 (9.72-0.83) 0.84 (0.75-0.90) 4.8 (2.9-7.8) 0.26 (0.1-0.3) 0.87 (0.83-0.89)
ARFI Fgt2 0.74 (0.66-0.80) 0.83 (0.85-0.89) 4.2 (2.8-6.3) 0.31 (0.2-0.4) 0.85 (0.82-0.88)
TE F4 0.89 (0.800.94) 0.87 (0.82-0.91) 6.7 (4.79.8) 0.13 (0.070.2) 0.93 (0.91-0.95
ARFI F4 0.87 (0.790.92) 0.87 (0.810.91) 6.4 (4.49.4) 0.15 (0.090.2) 0.93 (0.910.95)
Bota et all. Liver Int 2013
39
Shear wave elastography (SWE/PSWE)
Diagnosis of F2
Study Cut off Sens Spec LR LR- AUROC
Ferraioli 2012 (SWE) 7.1 90 97 7.2 0.1 0.92 (0.85-0.96)
Sporea 2014 (SWE) 7.8 76 82 4.2 0.2 0.86
Ferraioli 2014 (PSWE) 5.7 62 92 7.4 0.4 0.80 (0.71-0.87)
40
Shear wave elastography (SWE/PSWE)
Diagnosis of F4
Study Cut off Sens Spec LR LR- AUROC
Ferraioli 2012 (SWE) 10.4 87 97 27 0.1 0.98 (0.93-1)
Sporea 2014 (SWE) 11.5 80 93 11.4 0.2 0.91
Ferraioli 2014 (PSWE) 7.2 90 88 7.9 0.1 0.95 (0.89-0.99)
41
PROSTE vs ARFI vs SWE
TE ARFI SWE
- Short learning curve - Good reproducibility - High range of value (2- 75 KPa) - Quality criteria well defined -Prognostic value for cirrhosis Implemented on regular US devices ROI smaller than TE but chosen by the operator Higher applicability than TE (ascites, obesity) Performances equivalent to that of TE - Implemented on regular US devices ROI can be adjusted in size and chosen by the operator High range of value (2-150 KPa) Performances may be higher than TE for Fgt2
42
CONSTE vs ARFI vs SWE
TE ARFI SWE
- Dedicated device ROI cannot chosen Unable to discriminate between intermediate stages of fibrosis Low applicability (80) (obesity, ascites..) FP in case of acute hepatitis, extra hepatic cholestasis, vascular congestion - Ongoing validation Unit different from that of TE (m/sec) Unable to discriminate between intermediate stages of fibrosis Narrow range of values (0.5-4 m/sec) Influence of inflammation? Prognostic value in cirrhosis ? - Ongoing validation Limited data on reproducibility Unable to discriminate between intermediate stages of fibrosis Learning curve ? Quality criteria ? Influence of inflammation? Prognostic value in cirrhosis ?
43
TE Assessment of spleen stiffness
44
TE Assessment of spleen stiffness (SS)
and portal hypertension (PH)
113 patients with HCV related cirrhosis
underwent TE, HVPG and LB
Colecchia et al Gastro 2012
45
SS and the diagnosis of EV
Study Cut off Sens Spec LR LR- AUROC
Stefaneuscu 2011 46 83 71 2.8 0.2 0.78
Colecchia 2012 41 55 98 71 66 96 2.9 16.8 0.02 0.2 0.94 (0.90-0.98)
Calvaruso 2013 50 65 61 1.7 0.6 0.70
Fraquelli 2014 48 65 100 91 60 80 2.5 4.5 0.01 0.1 0.90 (0.79-100)
Sharma 2014 41 94 76 3.9 0.08 0.89 (0.84-0.95)
46
SS and LS for clinical complications in
compensated cirrhotics - A prospective study.
Events occurred during a 2 years f-up period 80
HCV patients
No events (n43 54)
Clinical decompensation (n26 32)
Other events without previous decompensation (n11
14)
Colecchia et al. J Hepatol 2014
47
Risk of clinical decompensation Cox regression
analysis
Variable Univariate Univariate Multivariate Multivariate Multivariate Multivariate
HR (95CI) p HR (95CI) p ß Std error ß
HVPG 1.21 (1.11-1.32) lt0.0001 - - - -
LS 1.06 (1.03-1.08) lt0.0001 - - - -
SS 1.09 (1.05-1.13) lt0.0001 1.08 (1.04-1.13) lt0.0001 0.081 0.020
MELD 1.63 (1.26-2.10) 0.0002 1.43 (1.07-1.91) 0.0160 0.357 0.148
LS liver stiffness SS spleen stiffness
Colecchia et al. J Hepatol 2014
48
TE - Conclusions ( I )

• TE is a user-friendly and reproducible technique
  that accurately predicts cirrhosis in patients
  with CLD
• It is less accurate in diagnosing significant
  fibrosis when used as single diagnostic
  non-invasive tool
• Newly developed US-based elastographic techniques
  (ARFI etc.) showed similar diagnostic accuracy in
  fibrosis staging but further validation is needed

49
TE - Conclusions ( II )

• Given its prognostic value TE could be used to
  better categorize patients with cirrhosis and
  assign them to different classes of risk for
  clinically relevant outcomes
• Spleen stiffness measurement allow a reliable
  prediction of the presence of EV and of hepatic
  decompensation.

50
Acknowledgements

• I would thank Francesco Di Mario, MD and Ludovico
  Abenavoli, MD for the kind invitation and all of
  you for having attended this lecture, which is
  dedicated to all my Colleagues, daily involved in
  taking care
• of difficult patients.

51
Thank you for your attention!