Intravascular ultrasound plaque characterization via spectral analysis

44A 1126-20

c~

c~

c~

w

cq C~ C~ c,4 <¢

ABSTRACTS - ACCIS2002 (Angiography & Interventional Cardiology) Long-Term Outcomes After Unprotected Left Main Coronary A r t e r y S t e n t i n g

Yoshinori Yoshida. Takeshi Kimura, Takashi Tamura, Shinichi Shirai, Kei Nishiyama, Kenji Ando, Satoshi Shizuta, Kazuaki Kaitani, Masashi Iwabuchi, Hiroyoshi Yokoi, Yoshihisa Nakagawa, Naoya Hamasaki, Katsumi Inoue, Hideyuki Nosaka, Masakiyo Nobuyoshi, Kokura Memorial Hospital, Kitakyushu, Japan.

1127-2

JACC

March6, 2002

Reproducibility of Three-Dimensional Lumen, Plaque, and Outer Vessel Reconstructions and of Endothelial Shear Stress Measurements In V i v o t o Determine Progression of Atheroscleroais

Ahmet U. Coskun. Scott Kinlay, Maureen E. Clark, Milan Sonka, Andreas Wahle, Olusegun J. llegbusi, Jeffrey J. Popma, Richard E. Kuntz, Charles L. Feldman, Peter H. Stone, Brigham & Women's Hospital, Boston, Massachusetts, Northeastern Un/versi~

Background: Previous reports demonstrated that unprotected left main coronary artery Boston, Massachusetts. (LMCA) stenting has favorable short-term results, but long-term outcomes are unknown. Methods: From March 1991 to August 1998, 73 consecutive patients (pts) with unproLocal hemodynamic factors play a critical role in atherogenesis. Low endothelial shear tected LMCA stenosis underwent elective stenfing. Treated pts were divided into 2 stress tESS) (<6-12 dynes/cm2) activates the endothelium, increases intimal proliferagroups; good candidate group (GC: 33pts) for coronary artery bypass grafting (CABG), tion, and promotes atherosclerosis. We have developed a methodology to reconstruct and poor candidate group (PC: 40pts) with risk factors; old age (>75yrs) 21pts, poor left the lumen, plaque, and external elastic lamina (EEL) of coronary arteries and measure ventricular (LV) function (ejection fraction<40%) 9pfs, cerebrovascular disease 14pts, ESS in-viva. The purpose of this study was to assess the reproducibility of these 3D prior CABG 6pts, chronic obstructive pulmonary disease 2pts, poor run off 1 patient, and • reconstructions and of the ESS measurements at different regions within native and cancer 2pts. Clinical and angiographic follow-up was performed with mean interval of 48 stented coronary arteries. Methods: The 3D anatomy of an artery was determined by fusmonths. In all patients, three-year follow-up information was obtained. To identify predicing end diastolic intracoronary ultrasound (ICUS) images from a motorized pullback tive factors of death and death/myocardial infarction/CABG (MACEs), Cox regression (0.5mm/sec) with biplane angiograms. Lumen and EEL boundaries were reassembled in analysis was performed, 3D space and divided into a 3D grid. Computational fluid dynamics was used to calculate Results: The procedure success rate was 100%. A follow-up angiography was obtained local wall ESS for each area of the grid. Complete data acquisition and analysis in 98% of eligible pts. Angiographic restenosis rate was 21%, and target lesion revascusequence was then repeated during the same catherization. Coronary flow was estitarization was required in 19%. Survival rates at 1-, 3- and 5-year were 90%, 83%, and mated using TIMI frame count. Results: 7 coronary arteries without significant stenoses 76% (97%, 97%, and 91% in GC group, 85%, 72%, and 65% in PC group). LMCAwere investigated (4 native and 3 stented) in 5 patients. The grid divided the surfaces into related death, defined as all cardiac deaths excluding pts with patent LMCA, occurred 2560-10640 areas/artery (av. 5900/artery). The correlation coefficients for all arteries for infrequently. At I and 5 years, 93 percent of pts were free from LMCA-related death. Surthe 2 measurements of lumen radius, EEL radius, plaque thickness and ESS for each vival rates free from MACEs at 1-, 3-, and 5-year were 88%, 79%, and 71% (97%, 94%, area were r=0.96, 0.95, 0.91, 0.88 respectively (all p<0.0001), Reproducibility was and 84% in GC group, 80%, 67%, and 60% in PC group). In Cox regression analysis, essentially identical in native and stented arteries. Conclusions: The 3D anatomy and poor LV function (hazard ratio [HR] 5.05, p=0.04) and reference diameter before stenting ESS of coronary arteries can be reproducibly determined in viva. Areas with pathologi(pre-RD)(HR 4,96/ram, p=0,01) were significant independent predictive factors of death. cally low ESS can be identified in-viva and may indicate areas likely to progress or Poor LV function (HR 4.08, p=0.03) and pre-RD (HR 3.18/ram, p=0.04) were also signifbecome vulnerable. The contribution of arteriaJ remodeling and plaque growth to icant independent predictive factors of MACEs. changes in the lumen can therefore be reliably determined, as can be the effect of ESS Conclusion: on plaque growth and remodeling, This methodology provides a tool to examine the 1. Once eady LMCA-related death could be avoided, long-term outcome of unprotected effect of ESS on remodeling, atherogenesis, and restenosis, as well as the impact of new LMCA stenting was favorable, particularly in Iow-CABG risk group. therapies. 2. LV function and vessel diameter are most important predictive factors of adverse cardiac events in LMCA stenting. 1127-21 Diagnostic Value of Low Frequency Intrevascular

Ultrasound for Detection of Aortic Diseases

POSTER SESSION 1127 New Approaches to Imaging Monday, March 18, 2002, 3:00 p.rn.-5:00 p.m. Georgia World Congress Center, Hall G

Presentation Hour: 3:00 p.m.-4:00 p.m. 1127-1

U s e o f O p t i c a l Coherence Tomography to Determine Laminar stucturea of Porcine Arteries

Mauro Ortiz, Sharon Thomsen, Thomas E. Milner, Jeehyun Kim, Marc D. Feldman,

University of Texas Health Science Center, San Antonio, Texas, University of Texas, Austin, Texas.

Background:Optical coherence tomography (OCT) is a novel imaging technology using reflection of light to image arteries. The magnitude of reflected light depends on the optical properties of laminar structures including elastin and collagen. As laminar structures reflect light, we hypothesize that they may be identified using OCT, Methods: Porcine mesenteric arteries (n=7) were collected from swine and imaged in 1.3 mm segments using OCT at a wavelength of 1310 nm. A metal plate was pieced beneath the artery as a reference to verify light penetration. Samples were prepared for elastin staining. OCT images were compared to histology to identify laminar structures and quantify the intensity of reflected light. A-scan images quantify reflected light amplitude versus tissue depth. Results: As shown in the Figure, the internal elastic lamina (IEL) and external elastic lamina (EEL) are easily identified. A-scan images identified the IEL and EEL as regions of greater reflected light amplitude. The negative slope in the A-scan between the IEL and EEL, and the EEL and the reflective metal, implies lower concentration of elastin and collagen in the media and adventitia. Discussion: OCT has identified arterial laminar structures (IEL and EEL) for the first time. A-scan slope differences aid in the quantitation of these boundaries. Since fibrous cap thickness determination represents optical reflectance differences between elastin/ collagen and the lipid pool below, OCT is a promising tool for this application.

Ela tin stain

Holoar Eaoebrecht, Dietrich Baumgart, Clemens van Birgelen, Susanne Gans, Thomas Buck, Raimund Ethel, Department of Cardiology, Essen, Germany

Introduction : Recently, a better understanding of the clinical nature of aortic dissection and its variants has been gained, With the advent of endoluminal stent-graft placement, a new, less invasive therapeutic alternative to surgery is available. In order to detect discrete alterations or plan aortic stent-grafting, a reliable and safe high-resolution imaging modality is needed. Methods : In 32 patients with known or suspected aortic dissection, intravascular ultrasound (IVUS) imaging was performed with 10/12.5 MHz- IVUS catheter (Scimed/Boston Scientific). Results : In all patients, the IVUS procedure could be successfully performed. NO vascular access complications occurred. In 18 patients, classic Stanford type- B aortic dissection was diagnosed. The entry site could be visualized in 10/18 pts, Branch vessel involvement could be evaluated for 43/54 vessels. In all of these pts., measurements of luminar dimensions were possible. In 10/32 pts., penetrating atherosclerotic ulcers (PAU) of the thoracic and abdominal aorta (class IV aortic dissection according to Svensson) were found. In 3/32 patients, pathological aortic alterations could be excluded. In one patient, a thoracic aortic aneurysm without dissection was diagnosed. In 20 patients with indication for endoluminal stent grafting, appropriate stent-grafts could be tailored based on IVUS measurements, In these patients, IVUS was successfully used for guidance during aortic stent-greft placement (dissection n=14, PAU n=6). Conclusion : Intravascular ultrasound allowed for online assessment of the aortic wall and the dissection morphology in viva. Based on quantitative measurements, tailoring of appropriate stent- grafts was possible. Also, IVUS was useful for intraoperative guidance during aortic stent-graft placement. 1127-22

Intravascular Ultrasound Plaque Characterization via Spectral Analysis

Anuia Nair, N. Obuchowski, B. D. Kuban, P. Schoenhagen, S. E. Nissen, E. M. Tuzcu, D. G. Vince, The Cleveland Clinic Foundation, Cleveland, Ohio. Background: Coronary atherosclerosis continues to be a leading cause of death in the "developed countries. However, the current clinical setting lacks adequate identification or characterization of unstable plaques. Spectral analysis of backscatterecl intravascular ultrasound (IVUS) data can provide information on plaque composition. Spectral parameters have been shown to be indicative of scatterer size, concentration, and acoustic impedance and can be employed to classify tissue. A combination of these spectral parameters would provide a robust and accurate classification scheme for the analysis of IVUS data. Methods: We compared various spectral algorithms to obtain an accurate plaque characterization tool and used a novel quantitative method for IVUS and histology correlation. Data were collected ex viva from 32 left anterior descending coronary arteries, with 30 MHz IVUS transducers in saline at physiologic pressure, Regions of interest (ROI), selected from histology, comprised 64 fibrous (F), 24 fibro-tipidic (FL), and 60 calcifiednecrotic (CN) regions. Spectra of IVUS data, identified for each ROI, were used to calculate spectral parameters. Statistical classification trees (CT) were computed for autoregrassive tAR) models of varying order numbers and signal processing variables and

JACC

March 6, 2 0 0 2

ABSTRACTS - ACCIS2002 (Angiography & Interventional Cardiology)

classic Fourier techniques with 75% of the data. The remaining data were utilized to assess the accuracy of the CTs. Results: The classical Fourier transform method performed well, correctly identifying 87.5% of F, 83.3% of FL and 93.3% of CN in the training clara set and 43.8%, 83.3%, 73.3% in the test data, respectively. The best AR CT surpassed the classic methods and identified 97.9% of F, 72.2% of FL and 95.6% CN in the training data, and 87.5%, 66.7%, 86.7% in the test data, respectively, with high specificity (90-100%) in each case. Some overlap was observed between the F and FL regions due to presence of collagen in both, and hence this result was not unexpected. Conclusion: This study proved that AR techniques ere better suited to the analysis of IVUS data than the more widely used classical tools. The AR classification scheme can be further employed to facilitate real-time evaluation of plaques with IVUS. 1127-23

In V i t r o a n d In V i v o V a l i d a t i o n o f a N e w D u a l - S e n s o r

0,014" Guide Wire for Simultaneous Measurement o f I n t r e c o r o n e r y Pressure and Flow V e l o c i t y Maria Siebes, Dirk S. Fokkema, Jos A. Spean, Jan J. Piek, Academic Medical Center, Amsterdam, The Netherlands.

Background: Stenosis resistance during hyperemia has been shown to be more accurate in predicting reversible ischemia than flow reserve based on either pressure (FFR) or Doppler signals (CFVR). Clinical application of this concept is hampered by the lack of a suitable dual-sensor guide wire. We tested a new 0.014" guide wire (Jomed) that is equipped with both a Doppler velocity transducer and a pressure sensor. Methods: Static and dynamic tests were carried out to assess the accuracy of the dual-sensor combowire in vitro. The in vivo performance was tested in goats. A constant pressure pertusion system connected to the left circumflex coronary artery was used to vary coronary pressure and flow. The signals obtained with the combowire were compared to independently measured reference signals of pressure (Millar) and volume flow (Transonic). Results: The physical and mechanical characteristics of the combowire match those of existing single-sensor guide wires and its in vitro accuracy corresponds to their respective specifications. The response of the pressure sensor was linear over a range from 0 to 230 mm Hg (Pcombo= -0.17+0.98 Pmillar, r=0.99). The average absolute difference to the reference signals in vitro was -1.9 + 2.9% for the pressure sensor and 0.2 + 1.5% for the velocity sensor. Both sensors exhibited a delay of about 25 msee in the dynamic tests, with a slight overshoot and no oscillations in the step response of the pressure sensor. The relationship between average flow velocity (Vavg) and volume flow (Q) measured in vivo was linear (Vavg=0.11 Q + 6.1, r=0.90) over a flow range between 30 and 220 mVmin. Simultaneously obtained pressure signals were related to reference measurements by Pcombo= 1.02 Pmillar - 5.2, r=0.99, over a pressure range from 46 to 125 mm Hg. Conclusion: A new 0.014" dual-sensor guide wire for intracoronary use combines well established technology into a single guide wire and faithfully records phasic signals of both pressure and flow velocity. Simultaneous acquisition of these signals distal to a coronary stenosis facilitates the determination of both stenosis and coronary microvascular resistance in humans, in addition to the concomitant assessment of both FFR and CFVR. 11 27-24

Diagnosing Coronary Atheroeclerosls U s i n g I n t r i n s i c Fluorescence and Reflectance

Georoe O. Aneheiciu. Sweder W. E. van de Poll, Joseph T. Arendt, Markus G. Mueiler, Irene Georgakoudi, Abigail Haka, Barry Kuban, Qinggho Zhang, Jonathan Myles, Maryann Fitzmaudce, John R. Kramer, Michael S. Feld, The Cleveland Clinic Foundation, Cleveland, Ohio, Massachusetts Institute of Technology, Cambridge, Massachusetts.

Background:. We present an algorithm for accurate diagnosis of coronary atharosclerosis based on intrinsic fluorescence (IF) and diffuse reflectance spectra of in vitro arterial tissue. In contrast to fluorescence, IF spectra are free from distortions introduced by tissue scattering and absorption. Methods: We collected coronaries from heart transplant and autopsy cases. We measured within one second white light reflectance and fluorescence emission spectra generated at 11 laser excitation wavelengths (Xx) with an original instrument called the FastEEM IF spectra were extracted by combining reflectance and fluorescence spectra using a photon migration model. IF spectra were fit to a linear combination of collagen and elastin spectra at ~x=340 nm, and of collagen and component C at Zx=480 nm. The C spectrum was derived from multivariate curve resolution analysis of iF spectra and related to that of ceroid, a lipid oxidation product in atherosclerotic lesions. We calculated the contributions of collagen and elastin to IF at ~x=340 and of C at Zx=480 nm. We also calculated the contribution of beta-carotene absorption to reflectance. Two diagnostic algorithms were derived to differentiate non-atherosclerotic from atherosclerotic arteries. Specificity, sensitivity and validity were verified with a leave-one OUt cross-validation algorithm. Results: 110 coronary segments were studied: 22 normal and intimal fibrop[asia and 88 non-calcified and calcified atherosclerotic/atheromatous plaques. An algorithm using contributions of collagen, elastin and C to IF and that of beta-carotene to reflectance separated the two groups with sensitivity 95%, specificity 86% and positive predictive value (PPV) 97%. A second algorithm using percentage of collagen contribution to IF at Xx=340 nm, contribution of C to IF at Xx =480 nm and that of beta-carotene to reflectance had sensitivity 95%, specificity 91% and PPV 96%. Conclusion: We have demonstrated that fundamental parameters extracted from spectral data can accurately diagnose atherosclerotic lesions using features similar to those used by pathologists, and expect the resulting algorithms to be more robust than those previously developed.

1127-25

45A

Single Versus Multisensor intrevaecular Thermography: The Advantage of Sensing Wall Temperature In Three Dimensions

Leonidas D. Diamaotooculos, Walter Desmet, Ivan de Scheerder, Yanming Huang, Xiaoshun Liu, Frans Van De Werf, Cardiology DepL UZ Gasthuisberg, Leuven, Belgium.

Background: Single-sensor thermography methods showed that inflammed, unstable plaques are hotter than stable ones. However, the advantage of studying the temperature distdbution at 3600 on the arterial wall, has not been proven in vivo. Methods: For the purposes of the study, we catheterized 5 pigs with normal coronary arteries. In each pig, inflammation was induced at the proximal right coronary artery by traumatizing the arterial wall via high-pressure balloon inflation. After 4 days, arterial wall temperature was studied with a thermography catheter that comprises 4-therrmistor sensors attached at 360° around its dista~ tip. A single thermistor catheter was also used to study the temperature at the same area. All data were collected and analyzed in a PC console with appropriate software. Both catheters were retracted in the area of interest, using motorized pullback at speed 0.3mm/sec. Results: Both single and multi-sensor catheters detected elevated temperature at the site of injury (DT-0.15°C, p<0.01). However, the multi-sensor catheter showed additional differences in temperature around the arterial circumference, and gave the possibility of building a complete thermal map of the injured area in 3 dimensions. Conclusions: Multi-sensor thermography seems to be far better than single-sensor, not only because it gives more thsrmographic data for the same arterial length, but also because it enables visualization of temperature in 3 dimensions.

1127-26

New Invasive Imaging Technique Rotational Aortogram: Comparison With Computer Tomography for Abdominal Aortic Aneurysm Evaluation

Fernando Boccalandro. Abid Assail, Bernardo De La Guardia, Hella Achour, Barat Raval, Richard W. Smelling, University of Texas Houston Medical School, Houston, Texas.

Background: Rotational Aortogram (RA) is a new imaging modality. It consist in a fast rotation (22.5° /sec) of the C-arm around the longitudinal plane of the vessel, while a radio-opaque contrast is injected. This motion, creates a high definition picture of the aorta that aloud views from multiple angles. The accuracy of this method has not been compared with established modes of imaging. This is the first study to compare the accuracy and agreement of measurements obtained by RA and computer tomography (CT) scan for imaging of the abdominal aorta. Methods: Aortic qualitative tortuosity and quantitative angulation, upper neck, lower neck and aorta maximum luminal diameter and length were measured by RA and CT scan in 30 patients (26 males; mean age: 70±11 yr.) with abdominal aortic aneursyms (AAA). Spearman's correlation coefficients and agreement using the Bland-Attman method were analyzed for comparison and agreement between both modalities in the three aortic segments. Results: Both methods had significant correlation and agreement in diameters and lengths in all aortic segments (P
Correlation and Agreement(Bland-AirmanAnalysis)Between RA and CT Scan in AAA Evaluation Aortic Segment

Correlation 95% C.I.

P Value Agreement 95% CI (mm) (mm)

Upper neck diameter

0.7999

0,58 to 0.89

<0.0001

-0.1

Upper neck lenght

0.4876

0.14 to0.72

0.0076

7.2

4.4 to 10

Aneuryemmaxdiameter

0.9555

0.89 to 0.98

<0.0001

0.4

-0.8 to 1.6

Aneurysmlenght

0.7391

0.51 to 0.87

<0.0001

11.6

7.5 to 15.6

Lower neck diameter

0.9412

0.87 to 0.99

<0.0001

-0.2

-0.6 to 0.3

Lower neck lenght

9.9299

0.80 to 0.97

<0.0001

0.3

-0.5 to 1.1

-0.8 to 0.6

Abdominal aorta tortuosity and angulation was better assessed by RA (P<0.001) when compared with CT scan. Conclusions: Intra-luminal aortic measurements obtained by RA have a significant correlation and agreement with measurements obtained by CT scan in patients with AAA. Rotational aortogram seems superior when compared with CT scan in the evaluation of aortic tortuosity and longitudinal angulation.