This document discusses the current approach for diagnostic imaging of small hepatocellular carcinoma (HCC) in patients with liver cirrhosis using gadoxetic acid (Primovist) enhanced magnetic resonance imaging (MRI) and diffusion weighted imaging (DWI). It summarizes that the combined use of Primovist MRI and DWI provides high diagnostic accuracy and sensitivity, particularly for detecting HCC lesions smaller than 2cm. This approach can image the step-wise progression of hepatocarcinogenesis by assessing vascularity, hepatocyte function, and cell density. The summary highlights that while Primovist MRI alone has high sensitivity, it has compromised specificity, but combining it with DWI increases specificity by differentiating HCC from

1. Diagnostic imaging of small HCC
in liver cirrhosis : The current
approach
Dr.Tony Loke
United Christian Hospital

2. Disclosure
No conflict of interest

3. Contents
• Step wise progression of hepatocarcinogenesis
• MR I can image this histhopathological process - Gadoxetic
enhanced MRI + DWI
• Definite Criteria for HCC – Low sensitivity
• Gadoxetic enhanced MRI alone - ↑sensitivity
• Diffusion weighted MRI -↑ specificity
• Combined Ga MRI and DWI – current approach
• Proposed Algorithm – Best diagnostic accuracy

4. Multistep progression of Hepatocarcinogenesis
Kudo M Oncology 2010;78:87-93

5. Intranodular blood supplyunpaired arteries / portal tracts
Tajima T AJR 2002:178;885

6. Histopathological pathway of Carcinogenesis
• Replacement of normal liver cells by abnormal liver
malignant cells -↓ OATP8 expression
• Hypointensity – HBP Primovist enhanced MRI
• Unpaired arteries ↑ + Portal tract ↓
• Wash in/ Wash out – Dynamic CT and MRI
• Increased cell density with progressive undifferentiated
nodules
• Hyperintensity - DWI

7. Process of Hepatocarcinogenesis
Kudo M J gastroenterology and hepatology 2010;25:439-452

8. Diagnostic criteria for typical HCC
Bruix J, Sherman M Hepatology 2011;53:1020-1022

9. Definite HCC
• Wash in (arterial phase hyperenhancement) /
Wash out– AASLD, EASL criteria (>1 cm) and
JSH (any size)
• Dynamic MDCT, MRI (Primovist enhanced MRI),
CEUS
• Only 71% - have ‘wash in’ and ‘wash out’ on
more than one test
Marrero JA et al Liver Transpl 2005;11:281-289

10. ↑ Sensitivity - Malignant liver nodules
• Absence of OATP8 expression / kupffer
cells
• Liver specific contrast agents
• Primovist MRI, SPIO-MRI, Sonazoid- CEUS
• Restricted diffusion
• Diffusion weighted MRI

11. Ancillary MRI criteria -
Malignant liver nodules
• T2W Hyperintensity
• Capsular enhancement – LI RADS
• T1 hypointensity
• Lesion size
• Lesional fat
• Lesion growth on follow up
• ‘Nodule in nodule’ pattern

12. PRIMOVIST alone has higher sensitivity
-? Compromised specificity
• Comparing primovist and magnevist,
• Significant increased sensitivity with primovist.
• Hepatocellular phase imaging with Primovist improves HCC detection.
• Primovist MRI has higher diagnostic accuracy (0.88 vs 0.74, p<.
001) and higher sensitivity (0.85 vs0.69, p<.001) than triple phase
MDCT
• Particularly in smaller lesions (<2cm)
• MRI with primovist has equal accuracy as dual-contrast MRI for
HCC detection
• Combined use of extracellular gadolinium and SPIO.
Marin D et al Radiology 2009;251:85-95
Park G et al BJR 2010;83:1010-1016
Kim YK et al Invest Radiol 2010:45:740-746
Martino et al Radiology 2010;256:806

14. Hypervascular nodule
Size does not matter!
Hypervascular nodule without washout
Primovist
Negative uptake
HCC
Positive uptake
Biopsy or Follow up
Kudo M Oncology 2010;78:87-93

15. Hypovascular Nodule- size matters!
Hypovascular nodule
Primovist
Negative uptake
≥1.5cm
HCC – 98%
LGDN – 2%
Positive uptake
<1.5cm
≥1.5cm
<1.5cm
Follow up
Biopsy or
FU
Follow up
17% progress to
HCC in 1 year
HCC
LGDN

16. PRIMOVIST (GD-EOB-DTPA)
GADOXETIC ACID
HCC-Hypointense at 20 min HBP (e)

17. DIFFUSION WEIGHTED MRI - ↑ Specificity
• SI ratio significantly differentiates malignant and benign
lesions at all b-values.
• Optimal threshold b=600
• SI ratio 1.25.
• For detection of HCC, DWI with b=600 has
• sensitivity of 95.2% compared to 80.6% for
conventional MRI (p=0.023)
• specificity of 82.7% compared to 65.4% (p=0.064%).
• The improved accuracy was most beneficial for differentiating
lesions smaller than 2cm.
Vandecaveye V Eur Radiol 2009;may:1431

18. Classic HCC -SI ratio 1.83 for b600 (g)

19. Small HCC -No lesion seen on T1, T2 or enhanced MRI.
SI ratio=1.5 at b600

20. CURRENT APPROACH
• Combined Primovist enhanced MRI
and Diffusion weighted imaging – able
to image the step wise pathogenesis of
HCC
• Dynamic Primovist MRI - Wash in / Wash out
• Hepatobiliary phase Primovist enhanced MRI - OATP8
expression
• DWI - cell density
• Ancillary features

21. COMBINED PRIMOVIST MRI AND DWI
• Criteria for HCC
• Hypervascular nodules with washout
• Hypervascular nodules without washout, hypointense on HBP
phase (irrespective of DWI)
• Hypervascular nodules without wash out, iso/hyperintense on
HBP phase + Hyperintense DWI
• Hypovascular nodules, hypointense on HBP phase +
Hyperintensity DWI
• Combined Primovist and DWI has better diagnostic
accuracy and sensitivity (93.3%) in detection of HCC <
2cm
• False positive – HGDN
Park MJ et al Radiology 2012:264;761-770

22. COMBINED PRIMOVIST AND DWI
Park MJ et al Radiology 2012:264;761-770

23. COMBINED PRIMOVIST AND DWI
Park MJ et al Radiology 2012:264;761-770

24. Typical small HCC
Small hypervascular HCC –DWI b=100, 800

25. Segment 6 hypervascular HCC < 1.5 CM

26. Atypical small HCC
Hypervascular HCC- hypointense on HBP but DWI normal

27. Hypervascular HCC
– and DN (DWI and HBP-Normal)

28. CT hypervascular lesionSegment 5 pseudolesion

29. Hypovascular HCC < 1.5 CM
Hypointense HBP + Hypertintense DWI

30. SEGMENT 2/3 HCC
Hypo/hypervascular component (HBP+DWI = +ve)

31. HGDN simulating HCC
Hypovascular DN- hypointense on HBP and DWI hyperintense

32. Segment 5 HGDN – Hypointense HBP (DWI-Normal)

33. DN > 1.5CM
HBP + DWI = negative

34. Summary - Current Approach
• Nodule detected on USG
• Dynamic CT
• Hypervascular with washout (>1cm)
• Atypical features
• Combined Primovist enhanced MRI and DWI
• Hypervascular without washout, hypointense on HBP phase
• Hypervascular without washout, isointense on HBP phase + Hyperintense DWI
• Hypovascular nodules, hypointense on HBP phase + Hyperintensity DWI
• Ancillary features
• Higher diagnostic accuracy and sensitivity
• Particularly for HCC < 2cm
• False positive for HGDN

36. Conclusion - Current Approach
• Combined Primovist and DWI
• Higher diagnostic accuracy and sensitivity
• Particularly for HCC < 2cm
• false positive include HGDN
• Combined Primovist enhanced MRI and DWI
• Hypervascular without washout, hypointense on HBP phase
• Hypervascular nodules, isointense on HBP phase +
Hyperintense DWI
• Hypovascular nodules, hypointense on HBP phase +
Hyperintensity DWI

37. SEG 2/3 HYPERVASCULAR HCC WITH
PARADOXICAL UPTAKE (DWI-hyperintense)

38. SEGMENT 6 HCC WITH NORMAL ARTERIAL,
T2W AND DWI

39. FOCAL NODULAR LIVER LESIONS:
1994 INTERNATIONAL CLASSIFICATION
• Regenerative nodule
• Cirrhotic nodule
• Low grade dysplastic nodule (adenomatous
hyperplasia)
• High grade dysplastic nodule (adenomatous
hyperplasia with atypia)
• Dysplastic nodule with subfoci of HCC (early HCC)
• HCC (overt HCC)
International Working Party. Hepatology
1995;22:983-993

40. DEVELOPMENT OF HCC IN
CIRRHOTIC LIVER
• Temporal progression from regenerative nodules
to dysplastic nodules to well differentiated HCC.
• HCC may develop independently of RN and DN.

41. PRIMOVIST MRI MOST SENSITIVE TECHNIQUE IN
DETECTING EARLY HEPATOCARCINOGENESIS

42. CONSENSUS STATEMENTSJAPAN SOCIETY OF HEPATOLOGY
• Typical HCC can be diagnosed by imaging regardless
of the size if a typical vascular pattern is obtained on
dynamic CT, dynamic MRI, CEUS or a combination of
CTHA and CTAP.
• Different from Western guidelines, only one dynamic
study showing the typical pattern is sufficient to
diagnose HCC.
• The typical imaging pattern include hypervascularity in
the arterial phase and washes-out in the portal venous
phase.

43. CONSENSUS STATEMENTSJAPAN SOCIETY OF HEPATOLOGY
• Sonazoid-enhanced ultrasound is more sensitive for
detection of intranodular hypervascularity than MDCT
or dynamic MRI. Therefore to confirm true
hypovascularity, sonazoid-enhanced CEUS is
recommended.
• Nodules with hypovascularity and negative findings on
SPIO-MRI, Kupffer imaging of Sonazoid CEUS,
primovist MRI are likely to be benign. They can be
followed up without treatment.

44. JAPAN SOCIETY OF HEPATOLOGY

45. JAPAN SOCIETY OF HEPATOLOGY

46. 2ND INTERNATIONAL FORUM LIVER MRI

47. PRIMOVIST CONTRAST ENHANCED
MRI- PROTOCOL OPTIMIZATION AND
EVALUATION OF HEPATIC NODULES IN
LIVER CIRRHOSIS
Dr. Tony Loke
Consultant Radiologist
United Christian Hospital

48. PRIMOVIST - GADOLINIUM-ETHOXYBENZYLDIETHYLENETRIAMINEPENTAACETIC ACID (GDEOB-DTPA)
• Combined extracellular hepatobiliary gadolinium
based contrast agents with liver specific properties
• Multihance and Primovist
• These agents able to assess both vascularity and
hepatocellular function.

49. PROTOCOL OPTIMIZATION - PHARMACOKINETIC
AND PHARMACODYNAMIC PROPERTIES OF
PRIMOVIST IN LIVER CIRRHOSIS.
• Patients with advanced cirrhosis, three important differences are
present.
• Diminished and delayed liver parenchyma enhancement
• diminished parenchymal enhancement in the hepatocyte phase
• time to peak enhancement may be delayed.
• Diminished and delayed biliary excretion.
• In the noncirrhotic liver, primovist produces intense biliary tree
enhancement beginning as early as 5 minutes after contrast injection.
• Enhancement of bile ducts in the cirrhotic liver is delayed and of limited
intensity

50. PROTOCOL OPTIMIZATION - PHARMACOKINETIC
AND PHARMACODYNAMIC PROPERTIES OF
PRIMOVIST IN LIVER CIRRHOSIS.
• Patients with advanced cirrhosis, three important differences are
present.
• Prolonged blood pool enhancement.
• 50% of primovist is cleared by the liver and 50% via the kidneys.
• Patients with advanced cirrhosis, the hepatic elimination pathway is
impaired and the blood vessels appear hyperintense for a longer
duration.
• The relatively low contrast enhancement in portal and hepatic veins
is relevant because it reduces the sensitivity for detecting venous
obstruction and invasion.

51. PROTOCOL OPTIMIZATION – PRIMOVIST
ENHANCED MRI IN CIRRHOTIC LIVER
• Problems with on-label approved dose Gd-EOB-DTPA of
0.025mmol/kg.
• Selecting the appropriate scan delay is difficult from low dose and
small amount of on-label approved dose
• Studies have shown the signal intensity of vessels in the arterial phase
is less with primovist than extracellular gadolinium-based agents using
on-label approved dose.
• Standard dose provide low sensitivity for detection of hypervascular
HCC/lesions despite its higher T1 relaxivitiy
Cruite I et al AJR 2010;195:29-41

52. PROTOCOL OPTIMIZATION-SOLUTION FOR
ACHIEVING OPTIMAL ARTERIAL PHASE
• Optimal arterial phase increases sensitivity for detection of hypervascular
lesions
• Perform consecutive arterial phase data sets.
• Administer the agent at a higher off-label dose (0.0375 – 0.05 mmol/kg).
• This is 50%-100% higher than the approved dose.
• Injecting all 10ml (20ml if patient exceeds the dose rate calculation) -rounded
up to the nearest bottle
• For patients with estimated GFR of less than 60mL/min, a weightadjusted dose is administered without rounding.
• Inject contrast at slower rate of 1cc/second followed by 20ml of saline chaser
at 2cc/second.
• 2cc/sec with higher dose

53. PRIMOVIST PROTOCOL - DIFFERENCE WITH
CONVENTIONAL GD

54. UCH PROTOCOL - PRIMOVIST ENHANCED MRI
LIVER
•
MRCP performed before contrast injection
•
Bolus timing method is used
• Contrast seen at LVOT or ascending aorta, patient asked to take 2 breath holds (8 -10
seconds) and 2 consecutive arterial phases imaging is acquired using 3D VIBE (15
seconds).
• late arterial phase is performed after 2 breath holds.
•
Hepatic phase is performed after another 2 breath holds
•
Equilibrium phase is performed at 120 minutes.
•
Diffusion weighted imaging and 2D axial SPAIR is performed while waiting for delayed 20
mins hepatocyte phase.
•
Hepatocyte phase - 20 minutes delay.
•
Hepatocyte phase - 40 minutes delay
• if hepatic veins and portal veins not cleared
• contrast not visible in biliary tree.
•
Hepatocyte phase - 60 minutes delay may be necessary.

55. UCH PROTOCOL –
PRIMOVIST ENHANCED MRI LIVER

56. WHEN TO USE PRIMOVIST IN PATIENTS WITH
LIVER CIRRHOSIS
• Primovist is routinely use in cirrhosis except for:
• Assessment of ablated lesions for residual or recurrent
disease.
• Reduced vascularity
• Patients whose bilirubin is above 3 mg/dL.
• Sensitivity for lesion detection reduced from diminished liver
enhancement
• Evaluation of vascular patency
• PV and HV remains hyperintense from prolonged blood pool
• Evaluation of hemangiomas
• Appearance same as HCC

57. IS PRIMOVIST BETTER FOR DETECTING HCC?
-COMPARED WITH OTHER AGENTS /IMAGING
MODALITIES
• Combined dynamic and hepatocyte phase of Primovist has greater
diagnoctic accuracy for HCC detection than either dynamic or
MDCT alone
• Comparing primovist and magnevist, significant increase in
sensitivity was achieved with primovist.
• Hepatocellular phase imaging with Primovist improves HCC detection
compared with conventional extracellular gadolinium chelates.
• MRI with primovist has equal accuracy as dual-contrast MRI for HCC
detection (simultaneous use of conventional extracellular gadolinium and
superparamagnetic iron oxide agent).
Marin D et al Radiology 2009;251:85-95
Park G et al BJR 2010;83:1010-1016
Kim YK et al Invest Radiol 2010:45:740-746

58. PRIMOVIST UPTAKE BY HEPATOCYTE BY OATP1
EXCRETION TO BILE JUICE REGULATED BY MRP2
Kudo M J gastroenterology and hepatology 2010;25:439-452

59. Kudo M Oncology 2010;78:87-93

60. HCC (87 lesions)
Hepatobiliary phase
In one study
Hypointense
92%
Isointense
6%
Hyperintense
2%
WDHCC (39 lesions)
Hepatobiliary phase
Another study
Hypointense
35
Isointense
2
Hyperintense
2
DN (8)
Hypointense
3
Isointense
3
Hyperintense
2
Kudo M J gastroenterology and hepatology 2010;25:439-452

61. CRITERIA FOR HCC
• Reading Hepatobiliary phase alone insufficient
• Result in false positives and negatives
• Hepatocyte phase
• Post contrast EOB ratio
• Read the whole exam
• T1 (hypointense)
• T2 (hyperintense)
• Dynamic contrast (hypervascularity +/- washout)
• DWI (restricted diffusion)

62. CRITERIA FOR HCC DIAGNOSIS
• A nodule with increased enhancement on arterial phase and washout on late venous or equilibrium phase
• A nodule with arterial enhancement and hyperintensity on T2WI
(and/or DWI)
• A nodule with isointensity during contrast enhanced arterial phase,
hyperintensity on T2WI (and/or DWI) and no uptake of contrast on
hepatobiliary phase (even < 1.5cm)
• Nodule > 1.5cm with no uptake of contrast agent on hepatobiliary
phase images.

63. HYPOVASCULAR NODULE – SIZE MATTERS
• Hypovascular nodules (on arterial phase) with
negative uptake on hepatobiliary phase are thought to
represent DN or WDHCC
• Lesion > 1.5 will progress to hypervascular nodules in 80%
in 1 year
• Lesion < 1.5 will progress to hypervascular nodules in 17%
in 1 year
Kudo M Oncology 2010;78:87-93

64. DN > 1.5CM - POSITIVE UPTAKE

65. DN > 1.5 CM - POSITIVE UPTAKE

66. DN - NEGATIVE UPTAKE

67. SEGMENT 3 HCC
– LESION DEMARCATION BETTER

68. HYPERVASCULAR RECURRENT HCC
– OTHER DN +VE UPTAKE

69. SEG 2/3 HYPERVASCULAR HCC WITH
PARADOXICAL UPTAKE

70. HYPERVASCULAR SEG 7 DN - POSITIVE UPTAKE

71. HYPOVASCULAR < 1.5 CM HCC – T2/DWI BRIGHT

72. DN –NOT VISIBLE IN OTHER SEQUENCES EXCEPT
HEPATOCYTE PHASE

73. SEGMENT 6 HCC WITH NORMAL ARTERIAL,
T2W AND DWI

74. CT SEGMENT 6 HYPOENHANCING NODULE
- HCC IN HEPATOCYTE PHASE > 1.5 CM

75. CT HYPERVASCULAR LESIONS – SEG 5 HCC

76. SEGMENT 6 HCC < 1.5 CM

77. CT HYPERVASCULAR LESION – SEG 5
PSEUDOLESION

78. SMALL HYPERVASCULAR HCC

79. SEGMENT 6 HCC & 2 HEMANGIOMAS

80. SEGMENT 5 – DN FEATURES ON T1/2

81. SEGMENT 5 HCC – NEGATIVE UPTAKE

82. COMBINED PRIMOVIST AND DWI
Park MJ et al Radiology 2012:264;761-770

83. FALSE POSITIVE – HEPATOBILIARY PHASE
• Hypointense lesions seen only on hepatobiliary phase (without arterial
enhancement and T2 hyperintensity) can either be WDHCC or Cirrhotic
nodules
• HCC tend to be larger(>1.5cm) than benign nodules(0.5-1.2cm)
• Lesions <1.5cm seen on hepatobiliary phase can still be well-differentiated
HCC and should not be ignored.
• Close monitoring, biopsy or resected in patients with coexisting overt HCC.
• Hypervascular pseudolesions
• 15% shows negative uptake on hepatobiliary phase
• 13% showed T2 hyperintensity
• DWI normal

84. FALSE NEGATIVE - HEPATOBILIARY PHASE
• HCC which are T1 hyperintense may be isointense on
hepatobiliary phase.
• Look for hypervascularity on arterial phase, T2/DWI
hyperintensity
• Hepatic dysfunction or hyperbilirubinemia reduces hepatic
uptake of the contrast agent
• lesion conspicuity on hepatobiliary phase images is
decreased, although false-negative cases can occur in
patients with normal bilirubin level.
• Lesions are less conspicuous in fatty liver
• do 20 min delay without fat sat
• Paradoxical uptake – Green hepatomas
• 2.5 to 8.5% of HCC appear iso/hyperintense
• Altered transporter mechanism

85. SUMMARY- CRITERIA FOR HCC
• Hypervascular nodules with washout – irrespective of delayed
phase
• Hypervascular nodules and hyperintensity on T2WI (and/or DWI) –
irrespective of delayed phase
• Isointense nodule (arterial phase) with hyperintensity on T2WI
(and/or DWI) and negative uptake of contrast on hepatobiliary
phase (even < 1.5cm)
• Nodule > 1.5cm with no uptake of Primovist on hepatobiliary
phase images– irrespective of vascularity

86. SUMMARY
• Negative uptake in hypovascular lesions <1.5 cm can still be
HCC
• FU required as 17% becomes hypervascular in 1 year
• Positive uptake can still be HCC
• DWI (+/- hepatocyte SI ratio) to exclude pseudolesion
• Hypointense rim and/or focal defect on delayed phase
• Nodule in nodule and internal septation helps
• FU(+/-biopsy) necessary