Coronary Magnetic
Resonance Imaging
Matthias Stuber, PhD
Associate
Associate Professor
Professor
Division
Division of
of MRI
MRI Research
Research
Johns
Johns Hopkins
Hopkins University
University
Baltimore,
Baltimore, MD
MD
• The Need for MRI
Background
• X-ray coronary angiograpy (gold standard)
–
–
–
–
Invasive
Radiation exposure for operator and patient
Small, but significant risk of complications
Substantial minority of patients are found to have no
significant coronary stenosis (~30-50%)11
– High procedural costs ($3000-$6000)22
– Inability to identify early atherosclerotic disease
• Both in the USA and in Germany ~1 million x-ray coronary
angiograms are performed each year22
1Budoff
et al. Circulation 1996; 93: 898
22002 AHA Heart and Stroke Statistical Update
Motivation
• Need for an alternative, non-invasive, more cost
efficient & patient friendlier technique,
which.…
– Accurately detects significant (≥50%) CAD
– Rules out non-significant CAD
•
Coronary Magnetic Resonance
Angiography (MRA)
• Challenges
Technical Challenges
• Small caliber & geometry of the coronary arteries
– Necessitates a high spatial resolution & sufficient
volumetric coverage.
• Contrast
– Contrast enhancement between coronary blood-pool, and
surrounding tissue (epicardial fat, myocardium).
• Myocardial motion
– Effective suppression of intrinsic (RR-interval) and
extrinsic (respiration) myocardial motion.
Technical Challenges
Intrinsic: Cardiac cycle: ~60/min; ~2cm
Extrinsic: Respiratory cycle: ~12/min; ~2cm
Expiration
Inspiration
32cm
• Solutions
Technical Challenges & Solutions
• No motion
suppression
• No contrast
enhancement
Technical Challenges & Solutions
• Suppression of intrinsic myocardial motion
– ECG triggering, segmentation of data acquisition1,
late diastolic image acquisition2
EKG
Image data acquisition
FFT
1)
2)
Edelman RR, Manning WJ, Burstein D. et al.: Radiology 181(3); 641-643 (1991).
Kim WY, Stuber M, Kissinger KV. et al.: J Magn Reson Imaging 14(4); 683-690 (2001).
Technical Challenges & Solutions
• No motion
suppression
• No contrast
enhancement
• ECG triggering
• No resp. mot.
suppression
• No contrast
enhancement
Technical Challenges & Solutions
•
Suppression of extrinsic myocardial motion
–
–
–
–
–
–
1)
2)
3)
4)
5)
Breath-holding1
Serial averaging2
Bellows gating2
Navigators2,3
‘Self’ Navigation4
Hybrid (Navigators & Breath-hold)5
Edelman RR, Manning WJ, Burstein D. et al.: Radiology 181(3); 641-643 (1991).
Oshinski JN, Hofland L, Mukundan S. et al.: Radiology 201(3); (1996).
Li D, Kaushikkar S, Haacke, EM. Et al.: Radiology 201(3); (1996).
Hardy CJ, Saranathan M, Zhu Y. et al.: Magn Reson Med. 2000 Dec;44(6):940-946.
Huber M, Oelhafen M, Kozerke S. et al.: J Magn Reson Imaging 15(2); 210-214 (2002).
Technical Challenges & Solutions
•
Suppression of extrinsic myocardial motion
–
Breath-holding
•
•
•
Diaphragmatic drift/registration errors in serial breathholds (~1cm)1
Major operator and patient involvement
Patient compliance
→ Applicability to patients with coronary disease is limited
→ Removed flexibility for enhanced spatial resolution
Free-breathing approaches
1) Danias PG, Stuber M, Botnar, RM et al.: Am J Roentgenol, 171(2):395-397, (1998).
Technical Challenges & Solutions
•
Suppression of extrinsic myocardial motion
–
MR navigator technology: Navigator gating & tracking1
Lung
Liver
Diaphragm
time
1.) McConnell et al.: Magn Reson Med 37(1); 148-152 (1997).
Technical Challenges & Solutions
• No motion
suppression
• No contrast
enhancement
• ECG triggering
• Navigator &
free breathing
• No contrast
enhancement
• ECG triggering
• No resp. mot.
suppression
• No contrast
enhancement
Contrast Generation
•
Contrast enhancement (lumen blood-pool and
surrounding epicardial fat, myocardium).
T1 [ms]
T2 [ms]
∆ω0 [Hz]
flow
Blood
1200
250
0
yes
Muscle
850
50
0
no
Fat
250
100
220
no
Technical Challenges & Solutions
•
Contrast enhancement
–
Endogenous contrast enhancement (T2Prep)
RF
90
TE
T2Myo: 50ms
180 180 180 180
T2Blood: 250ms
-90
time
z
y
x
1)
2)
GA Wright, DG Nishimura, A Macovski, Magn Reson Med 17:126-140 (1991).
JH Brittain, et al., Magn Reson Med 33:689-696 (1995).
Technical Challenges & Solutions
• No motion
suppression
• No contrast
enhancement
• ECG triggering
• No resp. mot.
suppression
• No contrast
enhancement
• ECG triggering
• Navigator &
free breathing
• No contrast
enhancement
• ECG triggering
• Navigator
• T2Prep
• State-of-the-Art & Comparison
to Gold Standard
Technical Challenges & Solutions
TTRIGGERDELAY
RIGGERDELAY
ECG
ECG
50m
50mss
FATSAT
FATSAT
TT2Prep
2Prep
NAVIGATOR
OR
NAVIGAT
Mo
Motion
tion TTrac
racking
king
30ms
15ms
30ms 15ms
3D
3D TTFE-EPI
FE-EPI
3D
3D TTFE
FE
(Sc
(Scout
out Scan)
Scan)
(HR
(HR-Scan)
-Scan)
70m
70mss
1)
1) Stuber
Stuber M,
M, Botnar
Botnar RM,
RM, Danias
Danias PG
PG et
et al.:
al.: JJ Am
Am Coll
Coll Cardiol;
Cardiol; 34(2):524-531
34(2):524-531 (1999).
(1999).
Single Center Coronary MRA Results*
* Navigators & free breathing
Study (n)
Sensitivity
Specificity
Sandstede’99 (23)
81
89
Huber’99 (?)
73
50
Sardinelli’00 (39)
90
90
Lethimonnier’99 (20)
65
93
Ikonen’00 (14)
84
70
Sommer’02 (77)
81
97
Adapted from: Sommer et al.: Rofo Fortschr Geb Rontgenstr N 2002; 459-466
Multicenter Coronary MRA Study
•
Purpose & Methods
–
–
–
1,2, to
Using uniform hardware, software & methodology1,2
examine the sensitivity, specificity, PPV, NPV of coronary
MRA for the diagnosis of significant disease of the proximal
coronary arteries.
Prospective comparison with gold standard (MR prior to XRay coronary angiography, independent core lab)
109 patients from 8 international centers in Philips Cardiac
MR Users network.
1)
1) Stuber
Stuber M,
M, Botnar
Botnar RM,
RM, Danias
Danias PG
PG et
et al.:
al.: JJ Am
Am Coll
Coll Cardiol;
Cardiol; 34(2):524-531
34(2):524-531 (1999).
(1999).
2)
Botnar
RM,
Stuber
M,
Danias
PG
et
al.:
Circulation;
99(24):3139-3148
(1999).
2) Botnar RM, Stuber M, Danias PG et al.: Circulation; 99(24):3139-3148 (1999).
Multicenter Coronary MRA Results*
* Navigators & free breathing
•
Results (Detection of >50% stenosis, n=109)
Any CAD [%]
LM/3VD [%]
Sensitivity
93
100
Specificity
42
85
PPV
70
54
NPV
81
100
1)
1) Kim
Kim WY,
WY, Danias
Danias PG,
PG, Stuber
Stuber M.
M. et
et al.:
al.: N
N Engl
Engl JJ Med;345(26):1863-1869
Med;345(26):1863-1869 (2001).
(2001).
Multicenter Coronary MRA Study
Aarhus
Berlin
Boston
Leiden
Köln
Texas
Leeds
Zürich
1)
1) Kim
Kim WY,
WY, Danias
Danias PG,
PG, Stuber
Stuber M.
M. et
et al.:
al.: N
N Engl
Engl JJ Med;345(26):1863-1869
Med;345(26):1863-1869 (2001).
(2001).
Multicenter Coronary MRA Study
Patient with LM/LAD &
LCX disease
Patient with 2 lesions in
proximal RCA
1)
1) Kim
Kim WY,
WY, Danias
Danias PG,
PG, Stuber
Stuber M.
M. et
et al.:
al.: N
N Engl
Engl JJ Med;345(26):1863-1869
Med;345(26):1863-1869 (2001).
(2001).
Multicenter Coronary MRA Study
• Conclusions
• Among patients referred for elective coronary
angiography, coronary MRA with real-time
navigator technology and T2Prep
– Accurately detects significant (≥50% lumen ∅)
coronary artery disease11.
– Reliably rules out non-significant coronary artery
disease11
1)
1) Kim
Kim WY,
WY, Danias
Danias PG,
PG, Stuber
Stuber M.
M. et
et al.:
al.: N
N Engl
Engl JJ Med;345(26):1863-1869
Med;345(26):1863-1869 (2001).
(2001).
Multicenter Coronary MRA Study
• Conclusions
There is a need for alternative methods which
• ultimately improve specificity of coronary MRA
in general
• provide additional/different information
• support access to more distal/branching vessels
• Enables visualization of atherosclerotic disease
that precedes lumen-narrowing
1)
1) Kim
Kim WY,
WY, Danias
Danias PG,
PG, Stuber
Stuber M.
M. et
et al.:
al.: N
N Engl
Engl JJ Med;345(26):1863-1869
Med;345(26):1863-1869 (2001).
(2001).
• Outlook and Promise
Works in Progress and Outlook
•
Alternative methods include
–
–
–
–
–
–
–
Contrast agents for coronary MRA
SSFP coronary MRA
‘Whole heart’ imaging
Arterial spin labeling
Interventional coronary MRI
Coronary vessel wall imaging, atherosclerosis &
molecular imaging
High field (3T) coronary MRI
• CONTRAST AGENTS
Contrast Generation
•
Contrast agents (Blood-Pool Agents)
T1 [ms]
T2 [ms]
∆ω0 [Hz]
flow
Blood
1200
~100
250
0
yes
Muscle
850
~300
50
0
no
Fat
250
50
250
no
Contrast Generation
Inversion Pre-Pulse
LV Navigator
Fat Sat (SPIR)
1
0.5
0
-0.5
Imaging
Blood
-1
Muscle
00
10
90
0
80
0
70
0
60
0
0
40
0
30
0
20
0
10
50
Time [ms]
0
Fat
-1.5
0
Magnetization M z [%M0]
1.5
Contrast Generation
TRIGGER DELAY
Ti
ECG
NAVIGATOR
FATSAT
INVERSION
INVERSION
Motion Tracking
30ms
35ms
15ms
(Sc out Sc an)
(HR-Sc an)
70ms
1)
1) Stuber
Stuber M,
M, Botnar
Botnar RM,
RM, Danias
Danias PG
PG et
et al.:
al.: JJ Magn
Magn Reson
Reson Imaging;
Imaging; 10(5):790-799
10(5):790-799 (1999).
(1999).
Intravascular Contrast Agent
Huber et al.: Magn Reson Med; Magn Reson Med. 2003 Jan;49(1):115-21
Courtesy: Paetsch I, Nagel E, Fleck E; Deutsches Herzzentrum Berlin
• Coronary Vessel Wall Imaging
Background
0.5..1mm
angiographically invisible!!
lumen
25%
50%
macrophages,
LDL,...
wall
macrophages
tissue factor
foam cells
lipid core
plaque disruption
with thrombosis
Wall inflammation
1) Glagov S, Weisenberg E, Zarins CK et al.: N Engl J Med 316(22); 1371-1375 (1987).
Introduction
• MRI has demonstrated ability to visualize the
vessel wall and its components (carotids,
aorta)1-3
• Coronary vessel wall imaging is technically
very challenging
– Small dimensions
– Constant motion
– Contrast ?
<1mm
<3mm
1) Toussaint JF, LaMuraglia GM, Southern JF et al.: Circulation 94(5);932-938 (1996).
2) Wasserman BA, Smith WI, Trout HH et al.: Radiology 223; 566-573 (2002).
3) Yuan C, Kerwin WS, Ferguson MS et al.: J Magn Reson Imaging 15; 62-67 (2002).
Methods: Contrast Generation
•
Contrast enhancement concept: ‘Dual inversion’
RF
Non-selective (180°)
Selective (180°)
} Dual-inversion
Mz
1
Vessel Wall
time
Blood
TI
1) Edelman RR, Chien, D, Kim D: Radiology; 181(3); 655-660 (1991).
Imaging
Methods: Motion Suppression
r Delay (TD)
Trigger Delay
Trigger Delay (TD)
ECG
Labeling
Delay
(TL)
Inversion
Delay
(TI)
Gx
IMAGING
Gy
Gz
RF
Non-Selective
Inversion
2D Selective
Re-Inversion
Navigator
45°
90°
Spiral
Imaging (<60ms)
t
Coronary Vessel Wall Imaging
RCA wall
3D SSFP
4th Order Local Inversion &
3D Spiral
79y old patient with ”luminal irregularities” in
RCA
luminal irregularities
thickened wall
RCA lumen
thickened wall
X-ray
MR wall image
* Botnar RM et al.: Magn Reson Med 2001 Nov;46(5):848-54 2002
* Kim W, Stuber M, Manning WJ et. Al.: Circulation 2002
RCA Vessel Wall Thickness
Wall thickness [mm]
2.50
P<0.01
2.00
1.7±0.3mm
1.50
1.0±0.2 mm
1.00
0.50
0.00
0
Healthy
subjects
Patients with
non-significant CAD
* Kim W, Stuber M, Manning WJ et. Al.: Circulation 2002
3
RCA Lumen Diameter
P=0.53
P<0.01
5.00
Lumen Diameter [mm]
4.50
4.00
3.4±0.5 mm
3.50
1.7±0.3mm
3.6±0.7 mm
3.00
2.50
1.0±0.2 mm
2.00
1.50
1.00
0.50
0.00
0
Healthy
subjects
Patients with
non-significant CAD
* Kim W, Stuber M, Manning WJ et. Al.: Circulation 2002
3
• High-Field Coronary MRA
Coronary MRA at 3 Tesla
(0.34x0.35x1.5mm voxel size)
→ MR System
→ Philips 3T Achieva
→ Dual Quasar Gradient System
→ 6-Element Cardiac SENSE Coil
→ Imaging Sequence
→
→
→
→
→
→
3D TFE
TE/TR: 2.3/7.6ms
Matrix/FOV: 800/270mm
Acquired Voxel Size: 0.34x0.35x1.5mm
Reconstructed Voxel Size: 0.26x0.26x0.75mm
Fat Saturation
→ Motion Suppression
→ FREEZE (automated prescription of diastolic
rest period)
→ VECG
→ Free-Breathing & Real-Time Navigator
M. Stuber, A. Ustun, R.G. Weiss, JHU