1. Pedro R. Moreno, MD, FACC
University of Kentucky
Lexington, Kentucky
Detection of Vulnerable
Atherosclerotic Plaques
By Near Infrared Spectroscopy
2. • Biological tissues have unique absorbance in the NIR wavelength range
• NIR light has enough penetration that may obtain spectra through blood
3. Tissue Evaluation by Near-IR
Spectroscopy
Absorbance peaks are caused by:
Combinations of fundamental bonds (C-H, C=C, C=O)
Electron transitions in the heaviest atoms
4. Advantages of Near-IR Spectroscopy
For Vulnerable Plaque Research
• Analysis under 1 second
• Simultaneous, multi-component, non- destructive analysis
• Chemical, biological and molecular information
• Automated predictions using computer algorithms
• Detection limits can be very low (from picograms to planets)
• Cost per analysis is minimal (no reagents used)
Dempsey RJ & Lodder RA. Applied Spectroscopy 1996;50:18A-34A
5. Near-IR Spectroscopy to Identify Vulnerable
Plaques
Studies 1998-2001
• Human aortic plaques in-vitro
• Human coronary plaques in-vitro
• Rabbit aortic plaques in-vivo
• Coronary pig safety study
• Human coronary plaques in-vivo
6. Hypothesis
• NIRS will identify vulnerable
plaques (1)
Methods
• Spectrometer: InfraAlyzer 500 (2)
• H & E and Trichrome staining
Identification Algorithm Model (3,4)
• 50% - Training set (histology)
• 50% - Validation set
Plaque Composition by Near-IR Spectroscopy
198 Human Aortic Plaques
1. Lipid pool (>30%), thin cap (< 65 µm), and macrophages 3. Chemometric software (Math 3.0, Matlab 5.1,Speakease IV Eta)
2. (Bran+Luebbe, Elmsford, N.Y.) 4. Regression by principal component analysis
Blinded prediction
Near-IR System
7. Near Infrared in Human Aortic Plaques
Moreno PR, et al. Circulation 2002;105:923-927
Lipid
Fibrotic
Thin Cap
Thick Cap
Macrophages
No Macrophages
(n=198)
8. Correlation of Blinded Near-IR Spectroscopy
Results with Histologic Findings
99 Aortic samples
HISTOLOGY
LIPID POOL THIN CAP MACROPHAGES
+ - + - + -
+
-
35 4
4 56
13 6
4 76
37 6
7 49
NEAR-INFRARED
SPECTROSCOPY
Moreno PR, et al. Circulation 2002;105:923-927
9. • Sensitivity (%) 90 77 84
• Specificity (%) 93 93 89
• PPV (%) 90 68 86
• NPV (%) 93 95 88
PPV:Positive Predictive Value; NPV:Negative Predictive Value
99 Aortic samples
Correlation of Blinded Near-IR Spectroscopy
Results with Histologic Findings
Lipid Pool Thin Cap Macrophages
10. Near-IR Spectroscopy to Identify Vulnerable
Plaques
Studies 1998-2001
• Human aortic plaques in-vitro
• Human coronary plaques in-vitro
• Rabbit aortic plaques in-vivo
• Coronary pig safety study
• Human coronary plaques in-vivo
11. Coronary Composition by Near-IR Spectroscopy
147 Human Coronary Sections
Hypothesis
• Lipid pool in coronary plaques
Methods
• Spectrometer: Foss/NIRSystems
• H & E and Trichrome staining
Identification Algorithm Model
• Training Set (76 sections)
• Validation set (70 sections)
Blinded prediction
Spectrometer
Moreno PR, et al. JACC 2001;37:356A
12. 03-Coronary Tissue Scans
798 1198 1598 1998 2398
Variable
0.5
1.0
1.5
2.0
Response
Normal Artery Fibrotic Plaque
Lipid-Rich, Calcified Thick Cap Atheroma Thin Cap Atheroma
Coronary Near-Infrared Spectra
Moreno PR, et al. JACC 2001;37:356A
14. Near-IR Spectroscopy to Identify Vulnerable
Plaques
Studies 1998-2001
• Human aortic plaques in-vitro
• Human coronary plaques in-vitro
• Rabbit aortic plaques in-vivo
• Coronary pig safety study
• Human coronary plaques in-vivo
15. Identification of Lipid-rich Aortic Atherosclerotic Plaques
in Living Rabbits With a Near-IR Spectroscopy Catheter
Hypothesis
• Normal vs. atherosclerotic plaques
• Lipid-rich versus lipid-poor plaques
Model
• Atherosclerotic Rabbit Model
• Pulsed 1% cholesterol x 8 months
• Normal rabbits (controls)
Near-IR Spectroscopy
• Laser-driven catheter system
Histology
• H & E and Trichrome staining
• Computerized planimetry (Zedex software)
Moreno PR, et al. JACC 2001; 37:3A:1039-21
Normal
Atherosclerotic
Normal
16. Moreno PR, et al. JACC 2001;37:3A
In-vivo Detection of Groups of Lipid Plaques With
a Near-IR Spectroscopy Catheter
True
(+)
True (-)
False (+)
False (-)
HISTOLOGY
+ -
+
-
19 0
5 6
NEAR-INFRARED
SPECTROSCOPY
Presence of lipid : Sensitivity: 79% Specificity: 100%
Lipid area >0.75 mm2
: Sensitivity: 75% Specificity: 78%
17. Near-IR Spectroscopy to Identify Vulnerable
Plaques
Studies 1998-2001
• Human aortic plaques in-vitro
• Human coronary plaques in-vitro
• Rabbit aortic plaques in-vivo
• Coronary pig safety study
• Human coronary plaques in-vivo
18. Percutaneous Coronary Near-IR Spectroscopy
In vivo: A Safety Study
Moreno PR and Fallon JT. University of Kentucky and Mount Sinai School of Medicine, August, 2001
Six Normal Swines
• Percutaneous, over-the wire NIR coronary catheterization
performed in 2/3 coronary arteries
Results:
• Successful coronary catheterization in all cases
• Excellent angiographic and histologic results with not a
single case of dissection, thrombosis or perforation.
3 French NIR catheter
19. Near-IR Spectroscopy in Humans
3 French CatheterPhase I - Safety
• Stable angina / PTCA-Stent *
• Reference normal segment
Prospective Study
• Angioplasty/stenting
• Scan 3 major arteries
• Follow/up 12 months
• Correlation spectra/events
20. Rx of VP
Trial of Detection & Treatment of Vulnerable Plaque
Patients
with
angina
Cath Lab
PTCA/
Stenting
Near-IR
Patients
with
TCFA*
Randomize
Placebo
Patients
w/out
TCFA
1 year
follow/up
* TCFA=Thin-cap fibroatheroma
UA=unstable angina - AMI: acute myocardial infarction - SCD=sudden cardiac death
UA
AMI
SCD
21. Near-IR Spectroscopy & Vulnerable Plaques
Conclusions
• NIR spectroscopy can identify features of plaque
vulnerability in vitro and lipid-rich plaques in vivo,
through blood.
• A catheter-based system has been tested in-vivo
with excellent performance in both swine and human
coronary arteries.
• Additional clinical data are needed to definitively
apply this technique to risk stratify human
atherosclerotic lesions in the cath lab.