1

TRANSIENT ELASTOGRAPHY AND OTHER TECHNIQUES

Raj Vuppalanchi, MDAssociate Professor of Medicine

Division of Gastroenterology and HepatologyMay 5, 2017

Financial Disclosure

2

• None

Overview

3

• Liver Stiffness and Liver Elastography

• Vibration-Controlled Transient Elastography

– Liver Stiffness Measurement (LSM) and

– Optimal Cut-offs

– Controlled Attenuation Parameter (CAP)

• Role of combined LSM and CAP in NAFLD/NASH

• Other techniques of Elastography

• Optimal Strategies for use of Elastography

Liver Stiffness

4

Fibrosis

Sinusoidal pressure Heart failure

Portal blood flow Post prandial

Biliary pressure PSC

Inflammation Acute hepatitis

Glisson’s capsule

NAFLD? ALT<80 U/L

? ALT<100 U/L

Liver Elastography

5

Liver Elastography- Current Approaches

6

Techniques that image the tissue response to a mechanical stimulus

Excitation Source

Mechanical (surface of the body)

Acoustic Radiation Force (inside liver)

Physiological Motion

Imaging Modality

Ultrasound

Magnetic Resonance

Property Displayed

Strain Elastography

Transient Elastography

Harmonic Elastography

Point Shear Wave Elastography

2D Shear Wave Elastography

Liver Elastography- Current Approaches

7

Ultrasound based Shear Wave Imaging

8

SWI

VCTE

(A mode imaging)

pSWE

(B mode imaging)

2D-SWE

(B mode imaging)

Probe Size

9

Probe size Depth (mm) US Freq (MHz)M+ 25-65 3.5XL+ 35-75 2.5

LSM (VCTE) and Fibrosis in NAFLD

10

Pathik et al. Ann Gastroenterol. 2015 Apr-Jun; 28(2): 281–286

N=110

kP

a

Park et al. Gastroenterology 2017; 152:598-607

Optimal VCTE- LSM Cut-Off for ≥F2

11

Study Probes No. of Pts % with ≥ F2 LSM Cut-off AUROC Sensitivity (%) Specificity (%)

Yoneda (2008) M+ 97 52 6.6 0.86 88 74

Nobili (2008) M+ 50 48 7.4 0.99 100 92

Wong (2010) M+ 246 17 7.0 0.84 79 76

Petta (2011) M+ 169 30 7.3 0.79 69 70

Lupsor (2010) M+ 69 26 6.8 0.79 67 84

Gaia (2011) M+ 72 64 7.0 0.80 76 80

Myers (2012) M+ 75 7.8 0.86 79 64

Wong (2012) M+ 193 23 7.0 0.83 79 64

Cassinotto (2015) M+ 291 71 6.20.82

90 45

9.8 60 90

Imajo (2016) M+ 142 54 11.0 0.82 65 89

Naveau (2014) M+ and XL+ 100 22 7.6 0.81 73 78

Park (2017) M+ and XL+ 94 30 6.9 0.86 79 85

Chen (2017) M+ and XL+ 111 36 7.8 0.83 82 78

NAFLD with F0/F1 vs. F2 to F4 (Clinically Significant Fibrosis )

Optimal VCTE-LSM Cut-off for Cirrhosis

12

Study Probe No. of Pts % with F4 LSM Cut-off AUROC Sensitivity (%) Specificity (%)Yoneda (2008) M+ 97 9 17 0.99 100 97Wong (2010) M+ 246 10 10.3 0.95 92 88Gaia (2011) M+ 72 12.5 10.5 0.94 78 96Myers (2012) M+ 75 22.3 0.88 80 91Wong (2012) M+ 193 13 10.3 0.89 81 83Cassinotto (2015) M+ 291 16.8 9.5

0.8792 62

16.1 65 90Imajo (2016) M+ 140 8 14 0.92 100 76Park (2017) M+ and XL+ 94 9 6.9 0.69 63 66Chen (2017) M+ and XL+ 111 10 14.6 0.90 82 92

Current Dilemma

13

• What are the cut-offs for ≥ F2 or F4?– Optimal

– 90% sensitivity

– 90% specificity

• Do the cut-offs vary depending on the distribution of NAFLD phenotype

• Are the cut-offs study cohort specific?

• How do we address the spectrum bias?

Controlled Attenuation Parameter

14

Ultrasound Attenuation Rate : dB/M (decibels per meter)

Range: 100 to 400 dB/M

CAP and Hepatic Steatosis

15

Normal liver

Fatty liver

de Lédinghen et al. J Gastroenterol Hepatol. 2015 Oct 29.

CAP and 1H-MRS

16

• 1H-MR Spectroscopy and CAP in 50patients with biopsy proven NAFLDand healthy volunteers

• Steatosis defined by % hepatocytesaffected by steatosis

• S1 – 5 to 33%

• S2 – 34 to 66%

• S3 – ≥ 67%

• CAP cut-off of 300 dB/m fordetection of S3 steatosis

• CAP cut-off of 215 dB/m for healthy

Karlas et al. PLoS One. 2014 Mar 17;9(3):e91987

CAP and MRI-PDFF

17

MRI-based protondensity fat fraction(PDFF) and TE-basedCAP for assessment ofsteatosis in 142 liverbiopsy proven NAFLD

Imajo et al. Gastroenterology. 2016 Mar;150(3):626-637

CAP and Hepatic Fat Fraction

18

M probe XL probe P-Value

S ≥ 2% AUROC = 0.83 (0.71-0.95)Cut-off 251 dB/mSe= 0.78 Sp= 0.78Acc = 0.80

AUROC = 0.84 (0.73-0.95)Cut-off 254 dB/mSe= 0.83 Sp= 0.78Acc = 0.83

0.76

S ≥ 8% AUROC = 0.87 (0.78-0.97)Cut-off 267 dB/mSe= 0.80 Sp= 0.79Acc = 0.81

AUROC = 0.90 (0.82-0.99)Cut-off 270 dB/mSe= 0.88 Sp= 0.79Acc = 0.85

0.50

S ≥ 16% AUROC = 0.92 (0.85-0.99)Cut-off 299 dB/mSe= 0.92 Sp= 0.88Acc = 0.90

AUROC = 0.91 (0.83-0.99)Cut-off 301 dB/mSe= 0.92 Sp= 0.81Acc = 0.85

0.78

S = steatosis; AUROC = area under the receiver operating characteristic curve; Se = sensitivity; Sp = specificity; Acc = accuracy

Sasso et al. Ultrasound Med Biol. 2016 Jan;42(1):92-103

MRE vs. TE and PDFF vs. CAP

19Park et al. Gastroenterology. 2017 Feb;152(3):598-607

ARFI-based Shear Wave Elastography

20

• B mode imaging requires expertise and prone to operator bias

• Simultaneous steatosis quantification is currently lacking

• Encouraging data using SSI in NAFLD

Excitation method Modalities Technique Options

ARFI (acoustic radiation force impulse)

Point SWE Point shear wave speed measurement

Virtual Touch Quantification (Siemens Healthcare)ElastPQ (Philips)Smart-Shearwave elastography (Samsung Medison)

2D-SWE Shear wave speed imaging

SWE (Aixplorer SuperSonic Imagine -SSI)Virtual Touch Image Quant, (Siemens Healthcare) Shear Wave Elastography (GE Healthcare)Shear Wave Elastography (Toshiba Medical Systems)

Manual compression or ARFI

Strain or displacement

Real-time tissue elastography (Hitachi)Virtual Touch Imaging (Siemens Healthcare)

Feasibility in Multi-Center Study –NASH CRN Experience

21

➢A total of 1706 exams were performed in992 NAFLD patients across eight clinicalcenters

➢Cohort—➢65% women➢BMI 33.6 (± 6.8) kg/m2➢Waist circumference 107 (± 14) cm

➢XL+ probe – 60%➢Failure rate – 3.2%➢Proportion of Unreliable scans – 2.4%

Vuppalanchi et al. AASLD 2015 (Manuscript under review)

N=55Skin to capsule distance > 3.5 cm 19 35%All readings were invalid 6 11%No liver detected 1 2%Machine/technician error 13 24%Patient in too much pain 4 7%Patient refusal 11 20%Unknown reason 1 2%

Performance Characteristics

22

BMI categoryNormal /

Overweight*<30 kg/m2 (n=311)

Obese30-34.9 kg/m2

(n=311)

Extreme Obese≥35 kg/m2

(n=358)

Total†(n=992)

P-value

Probe type - % XL probe 26.7% 60.8% 87.4% 59.7% <0.001

Liver Stiffness Measurement (LSM)Average 1st and 2nd reading

Mean (SD) - kPa 9.0 (9.6) 11.0 (11.8) 13.8 (12.8) 11.4 (11.7) <0.001Median [IQR] - kPa 6.0 [4.4, 9.3] 7.1 [5.2, 11.8] 9.3 [6.6, 15.0] 7.5 [5.3, 12.2]

Difference 1st and 2nd readingMean (SD) - kPa -0.2 (4.8) 0.3 (6.3) 0.0 (5.2) 0.0 (5.5) 0.96Median [IQR] – kPa 0.0 [-0.7, 0.7] 0.0 [-0.7, 1.1] 0.1 [-1.3, 1.4] 0.0 [-0.9, 1.1]95% limits of agreement† - kPa ±2.0 ±2.6 ±3.9 ±2.9

Controlled Attenuation Parameter (CAP)

Average 1st and 2nd readingMean (SD) – dB/m 284 (56) 303 (47) 336 (42) 307 (53) <0.001Median [IQR] – dB/m 284 [254, 321] 306 [270, 337] 340 [311, 366] 312 [272, 347]

Difference 1st and 2nd readingMean (SD) – dB/m 3 (32) -2 (38) 5 (36) 2 (36) 0.07Median [IQR] – dB/m 3 [-14, 18] -3 [-21, 16] 0 [-16, 24] 0 [-16, 19]95% limits of agreement† - dB/m ±46 ±54 ±58 ±51

Vuppalanchi et al. AASLD 2015 (Manuscript under review)

Covariates of Unreliable LSM

23

Unreliable examsQuartile No. prior exams Total exams No. %

1 1 to 11 244 11 4.92 12 to 27 250 10 4.13 28 to 58 245 14 5.04 59 to 180 253 4 1.6

Proportion of operator’s experience as quartiles and Unreliable VCTE scans

Odds of unreliable liver stiffness measurementCovariate Odds Ratio 95% CI P valueBody Mass Index (per category†) 2.7 1.5, 4.8 0.001International normalized ratio (%) 0.94 0.89, 0.99 0.01Ethnicity (Hispanic vs non-Hispanic) 3.2 1.1, 9.3 0.03Age (per yr) 1.03 1.00, 1.07 0.08Prior operator-specific readings (per quartile) 0.75 0.52, 1.08 0.12ALT (per 10 U/L) 0.91 0.78, 1.05 0.19

Variables associates with Unreliable VCTE scans

Vuppalanchi et al. AASLD 2015 (Manuscript under review)

Relationship between Probe and LSM / CAP

24

XL+ probe M+ probe BMI-AdjustedDifference (XL – M)

N BMI-Adjusted

Mean

N BMI-Adjusted

Mean

Mean 95% CI P-value

LSM (kPa)Fibrosis stage0 132 5.4 112 6.5 -1.0 -1.6, -0.5 <0.0011 136 6.4 106 7.2 -0.9 -1.7, -0.1 0.032 90 7.2 57 7.8 -0.5 -1.7, 0.6 0.353 137 11.3 70 13.7 -2.4 -4.6, -0.2 0.044 57 21.3 30 26.2 -4.8 -12.6, 2.9 0.22Total 552 8.0 375 8.8 -0.8 -1.4, -0.3 0.003

CAP (dB/m)Steatosis score0 20 268 41 289 -21 -50, 9 0.171 160 313 155 300 13 0, 25 0.052 117 321 106 306 15 1, 29 0.043 80 342 73 310 32 18, 46 <0.001Total 377 319 375 303 16 8, 24 <0.001

Vuppalanchi et al. AASLD 2015 (Manuscript under review)

Accuracy of VCTE to Detect Cirrhosis

25

0.0

00.2

50.5

00.7

51.0

0

Sensitiv

ity

0.00 0.25 0.50 0.75 1.001 - Specificity

Area under ROC curve = 0.8517

Optimal LSM: 15.5 kPa• Specificity fixed at 90%: AUROC

0.93 (95% CI: 0.86-0.99) for

differentiating cirrhosis from lesser

fibrosis stage

• Sensitivity: 0.76

• Positive Predictive Value: 0.37

• Negative Predictive Value: 0.98

M. Shadab Siddiqui et al. EASL 2017

N=62

High Success Rate in Multi-Center

26

• A failure rate of 3% when FibroScan is used forestimation of LSM and CAP in patients withNAFLD

• Excellent inter- and intra-operator agreement forLSM and CAP

• Use of the XL+ was associated with• lower LSM values• higher CAP values

Vuppalanchi et al. AASLD 2015

Longitudinal Assessment

27

Change in LSM (VCTE) over time (ΔLSM/Δt )

• Understand the natural history (placebo arm)

• Assess therapeutic response to treatmentintervention

• ΔLSM/Δt : Change in fibrosis + (inflammation)

• ΔCAP/Δt : Change in hepatic fat

• ΔLSM/Δt• ΔCAP/Δt

Hard outcomes• Death or transplant

Composite endpoints• Improvement in NAS• Improvement in Fibrosis• Resolution of NASH

Elastography in Drug Development

28

VCTE with LSM and CAP

MRE with PDFF

Histology – L. Bx

Hard Outcomes

Phase 1

• Safety

Phase 2a

• Demonstrate ability to reduce liver fat content

• Trends in improvement in ALT and other blood based markers (inflammation)

• Metabolic indices

• Continued safety profile

Phase 2b

• Confirm effect on reducing liver fat content

• Reduce inflammation

• Significant improvement in metabolic syndrome

• Confirmation of clean safety profile

Phase 3

• Resolution of NASH with no worsening in fibrosis

• Reduction in fibrosis with no worsening in NASH

Screening Enrichment

Summary

29

• Liver elastography can be obtained through US or MRIbased approaches

• Simultaneous LSM and CAP by VCTE or MRE with PDFF isvery convenient.

• One-time measurement of LSM and CAP – riskstratification and screening enrichment tool

• Change over time (ΔLSM/Δt or ΔCAP/Δt or Δ PDFF%/Δt)when linked to hard outcomes or surrogate endpointsmay be a game changer

• Strategies to minimize confounders is very critical in aclinical trial setting

Liver Stiffness Revisited

30

Liver hand palpation: “The living are soft and yielding; the dead are rigid

and stiff”, Lao Tzu, (6th century BCE).

Thank youQ and A

31

32

33

NAFLD- Evaluation

34The diagnosis and management of non-alcoholic fatty liver disease: practice guideline. HEPATOLOGY 2012;55:2005-2023.

Reliability: Effect of Confounders

35

Confounders for LSM

ObesityAlcohol

Congestion

Inflammation

Inexperience Non-fasting

Cholestasis

Fast >3 hrs

Check ALP and T.Bili

Check ALT

Operator experience >100 exams

Right heart failure

Alcohol history XL+ probe

Patel et al. Semin Liver Dis. 2015 May;35(2):166-83.

Effect of Food Intake on LSM

36Arena et al. Hepatology. 2013 Jul;58(1):65-72.

Standardized liquid meal (400 mL, 600 Kcal, 16.7% protein, 53.8% carbohydrates, 29.5% fat) in HCV patients.

Echosens: 3 hours fastNASH CRN : Overnight fast

Myths and Mysteries

37

Technology Technician/Technique Tissue Tablets/Tonics

Shear wave propagation Operator experience Probe to liver distance Medication usage

TM mode Variability - Ascites - Beta Blockers

A mode - Intra-operator - Adiposity Etiology specific therapy

Algorithm - Inter-operator - Altered anatomy Significant weight loss

Software Acute hepatitis - Bariatric surgery

Probe size Cholestasis Recent alcohol use

- Medium Portal flow

- Extra-large - Postprandial state

TIPSS

Tumor

Cysts

Infiltrative liver disease

Hemangioma

Congestive hepatopathy

Hepatic steatosis

Vuppalanchi and Sanyal. Clin Gastroenterol Hepatol. 2015 Apr;13(4):780-2.

38

39

Fibroscan – Clinical Trial Setting

40

• Successful recruiting by enrichment• CAP>300 dB/m for MRI-PDFF>10%

• LSM >10.3kPa for ≥ F2 stage of fibrosis

• LSM ≥ 14kPa for NASH cirrhosis

COMBINED LSM and CAP

41

CAPLSM

Probe

NAFLD Phenotype

42

• NAFLD

• NAFL

• NASH– Without Fibrosis

– With Fibrosis

• Early (F1)

• Clinically significant (≥ F2)

• Advanced (F3-F4)

Where does my patient fall on the

spectrum?

BMS NuSirt Intercept Genfit Galectin

Phase 2 ≥10% hepatic fat ≥15% hepatic fat NASH cirrhosis with HVPG ≥ 6 mm Hg

Phase 3 NASH with F2-F3 NASH with F1-3

43

44

45

46

47

Excitation method Modalities Technique Options

ARFI (acoustic radiation force impulse)

Point SWE Point shear wave speed measurement

Virtual Touch Quantification

2D-SWE Shear wave speed imaging SWE (Aixplorer Supersonic Imagine); Virtual Touch Image Quantification, (Siemens Healthcare); Shear Wave Elastography (GE Healthcare): Shear Wave Elastography (Toshiba Medical Systems)

Manual compression or ARFI Strain or displacement Real-time tissue elastography (Hitachi); Virtual Touch Imaging (Siemens Healthcare)

MRE

48

Clinical Gastroenterology and Hepatology

Volume 5, Issue 10, Pages 1207-1213.e2

(October 2007)

SWE

49