Accuracy of In Vivo Coronary Plaque Morphology Assessment A Validation Study of In Vivo Virtual Histology Compared With In Vitro Histopathology

Transcription

1 Journal of the American College of Cardiology Vol. 47, No. 12, by the American College of Cardiology Foundation ISSN /06/$32.00 Published by Elsevier Inc. doi: /j.jacc Accuracy of In Vivo Coronary Plaque Morphology Assessment A Validation Study of In Vivo Virtual Histology Compared With In Vitro Histopathology Studies With Intravascular Ultrasound Kenya Nasu, MD,* Etsuo Tsuchikane, MD, PHD,* Osamu Katoh, MD,* D. Geoffrey Vince, PHD, Renu Virmani, MD, Jean-François Surmely, MD,* Akira Murata, MD,* Yoshihiro Takeda, MD,* Tatsuya Ito, MD,* Mariko Ehara, MD,* Tetsuo Matsubara, MD,* Mitsuyasu Terashima, MD,* Takahiko Suzuki, MD, PHD* Aichi, Japan; Cleveland, Ohio; and Washington, DC OBJECTIVES BACKGROUND METHODS RESULTS CONCLUSIONS The goal of the present study was to compare the accuracy of in vivo tissue characterization obtained by intravascular ultrasound (IVUS) radiofrequency (RF) data analysis, known as Virtual Histology (VH), to the in vitro histopathology of coronary atherosclerotic plaques obtained by directional coronary atherectomy. Vulnerable plaque leading to acute coronary syndrome () has been associated with specific plaque composition, and its characterization is an important clinical focus. Virtual histology IVUS images were performed before and after a single debulking cut using directional coronary atherectomy. Debulking region of in vivo histology image was predicted by comparing pre- and post-debulking VH images. Analysis of VH images with the corresponding tissue cross section was performed. Fifteen stable angina pectoris (AP) and 15 patients were enrolled. The results of IVUS RF data analysis correlated well with histopathologic examination (predictive accuracy from all patients data: 87.1% for fibrous, 87.1% for fibro-fatty, 88.3% for necrotic core, and 96.5% for dense calcium regions, respectively). In addition, the frequency of necrotic core was significantly higher in the group than in the stable AP group (in vitro histopathology: 22.6% vs. 12.6%, p 0.02; in vivo virtual histology: 24.5% vs. 10.4%, p 0.002). Correlation of in vivo IVUS RF data analysis with histopathology shows a high accuracy. In vivo IVUS RF data analysis is a useful modality for the classification of different types of coronary components, and may play an important role in the detection of vulnerable plaque. (J Am Coll Cardiol 2006;47: ) 2006 by the American College of Cardiology Foundation Atherosclerosis and its thrombotic complications are the leading cause of morbidity and mortality in industrialized countries. Rupture of vulnerable atherosclerotic plaques is the cause of most acute coronary syndrome (). Atherosclerotic stability is related to its histologic composition, risk stratification, and treatment of high-risk lesions before rupture. Gray scale intravascular ultrasound (IVUS) is a useful modality for characterizing the extent and distribution of atherosclerotic plaques in vivo as well as for the determination of the morphology of atherosclerotic plaques and the vessel wall (1 3). However, the region of low echogenicity, which is thought to represent the composition of lipidcontaining and mixed plaque, remains relatively uncharacterized by gray scale IVUS (1,2). From the *Department of Cardiology, Toyohashi Heart Center, Toyohashi-city, Aichi, Japan; Department of Biomedical Engineering, The Cleveland Clinic Foundation, Cleveland, Ohio; and the Department of Cardiovascular Pathology, Armed Forces Institute of Pathology, Washington, DC. This study was supported by a grant from Volcano Therapeutics, Inc., Rancho Cordova, California. Manuscript received October 10, 2005; revised manuscript received January 27, 2006, accepted February 7, Spectral analysis of the radiofrequency (RF) ultrasound backscatter signals known as Virtual Histology (VH) offers an in vivo opportunity to assess plaque morphology (4 7). Indeed, the VH IVUS technology (Volcano Therapeutics, Inc., Rancho Cordova, California) has been shown to have a 80% to 92% in vitro accuracy when used to identify the four different types of atherosclerotic plaques (e.g., fibrous, fibro-fatty, dense calcium, and necrotic core) (4). However, no quantitative in vivo histologic comparisons or validations exist. Therefore, the goal of the present study was to compare the accuracy of in vivo tissue characterization obtained by VH IVUS RF data analysis to the in vitro histopathology of coronary atherosclerotic plaques obtained by directional coronary atherectomy (DCA), and also to evaluate the differences in plaque compositions between stable angina pectoris (AP) and. METHODS Study design. The present study is a prospective singlecenter registry. The inclusion criteria consisted of patients older than 18 years of age with either symptomatic stable

2 2406 Nasu et al. JACC Vol. 47, No. 12, 2006 In Vivo Virtual Histology June 20, 2006: Abbreviations and Acronyms acute coronary syndrome AP angina pectoris DCA directional coronary atherectomy IVUS intravascular ultrasound RF radiofrequency VH Virtual Histology AP or (Braunwald class IIIB, with positive serum markers) who were good candidates for percutaneous coronary revascularization. Angiographic inclusion criteria were: 1) target vessel reference diameter of 2.5 mm by visual estimation to allow the use of the FLEXI-CUT atherocatheter (Guidant Corp., Santa Clara, California; 2) one denovo culprit lesion of 50% diameter stenosis, as determined by on-line quantitative coronary angiography; 3) mild-to-moderate vessel tortuosity; and 4) left ventricular ejection fraction 30%. Exclusion criteria were: 1) contraindications to IVUS examination; 2) severe peripheral vascular diseases that precluded the use of a 8-F arterial sheath; 3) other concomitant diseases or medical conditions that could impact patient/procedural outcomes, such as history of bleeding diathesis, stroke, or transient ischemic neurological attacks within the past year or hypersensitivity to heparin, aspirin, ticlopidine or X-ray contrast media; and 4) a positive pregnancy test. The institutional review board of our institution approved the study, and all patients gave written, informed consent. Procedure and data acquisition. The schema of procedure and data acquisition is illustrated in Figure 1. After baseline angiography, a 3.2-F, 30-MHz catheter (Boston Scientific Scimed Inc., Maple Grove, Minnesota) was placed distal to the target lesion. The catheter tip position was determined by infusion of contrast media, and was subsequently pulled Figure 1. Schematic description of the procedure and data acquisition. After the initial angiogram, intravascular ultrasound (IVUS) with electrocardiogram-gated radiofrequency (RF) data was recorded. Directional coronary atherectomy was performed just one time with high pressure at the target lesion, and the tissue sample was extracted from the atherocatheter for post-processing. After debulking, an angiogram and electrocardiogram-gated RF data acquisition were performed. back to the aortic ostium using a motorized pullback system set at 0.5 cm/s. During pullback, gray scale IVUS was recorded on super VHS videotape for off-line analysis, and raw RF data was captured at the top of the R-wave for reconstruction of the color-coded map by a VH data recorder (Volcano Therapeutics, Inc.). The captured RF data were written on optical discs and sent to the Cleveland Clinic Foundation (Cleveland, Ohio) for VH IVUS analysis. Directional coronary atherectomy was performed at the region of the target lesion that corresponded to minimal luminal diameter on gray scale IVUS. Debulking was performed just one time with high pressure. The DCA procedure was performed and recorded in accordance with hospital standard of care. Angiography and IVUS with electrocardiogram-gated raw RF data acquisition were repeated after the atherectomy. The tissue sample was extracted from the atherocatheter and marked with a small clip or ink at the distal end. During DCA, the tissue sample was assumed to be cut and pushed into the nosecone. However, it was not certain that the tissue sample was only pushed into the nosecone, but could possibly be curled into it, therefore misleading the marking of the proximal and distal end of the tissue sample. To address this issue, an in vitro experiment was performed as follows: 1) red ink was injected in a piece of porcine aorta; 2) an atherocatheter was positioned with the proximal end of the cutter window placed on the ink injected region; 3) the window was pressed on the aortic wall by finger, and the debulking procedure was performed; and 4) saline was injected from the tip of the nosecone, and the direction of the extracted tissue was evaluated. Over the 20 samples obtained with this method, the end of the tissue sample was reversed only one time. Quantitative analysis. Angiography was performed in at least two projections. Pre-debulking on-line quantitative coronary angiography was conducted utilizing the view revealing the highest degree of stenosis, and severity of coronary stenosis was measured using the Cardiovascular Measurement System (CMS-MEDIS Medical Imaging System, Leiden, the Netherlands). The lesion length, reference diameter, minimal luminal diameter, and diameter stenosis was calculated off-line by an independent operator. Analysis of cine frames was performed in end-diastole. Histopathology analysis. The length of tissue sample was measured immediately after extraction from the catheter. Tissue samples were immersion-fixed in 10% neutralbuffered formalin, processed for paraffin embedding, and sliced into 4- m sections every 0.5 mm, starting proximally. After staining with hematoxylin and eosin, histology sections were forwarded to the Armed Forces Institute of Pathology (Washington, DC) and analyzed by an isolated operator who was blinded to the IVUS data acquisition. Four plaque components (fibrous tissue, fibro-fatty, necrotic core, and dense calcium) were defined (4,5) as follows: 1) fibrous tissue: areas of densely packed collagen; 2) fibro-

3 JACC Vol. 47, No. 12, 2006 June 20, 2006: Nasu et al. In Vivo Virtual Histology 2407 fatty: fibrous tissue with significant lipid interspersed in collagen; 3) necrotic core: necrotic regions consisting of cholesterol clefts, foam cell, and microcalcifications; and 4) dense calcium: calcium depositing without adjacent necrosis. After analysis, the digitized histopathologic images and a description of the data were forwarded to the Cleveland Clinic Foundation for correlative analysis. Gray scale IVUS analysis. From the pre-debulking IVUS data, the smallest lumen at the culprit lesion was identified from clockwise and longitudinal plaque distribution. Calculations were performed by an experienced operator. Vessel cross-sectional area and lumen cross-sectional area were calculated, and the difference between the two values was defined as plaque plus media cross-sectional area. Percent plaque plus media cross-sectional area was defined as plaque plus media cross-sectional area divided by vessel crosssectional area. VH IVUS analysis. Virtual histology analysis was performed at The Cleveland Clinic Foundation. Atherosclerotic coronary plaques were characterized by classification trees based on mathematical autoregressive spectral analysis of IVUS backscattered data (IVUSLab software, Volcano Therapeutics, Inc.), as described previously (8). The presence of fibrous, fibro-fatty, necrotic core, and dense calcium areas were assessed within the region of target lesion using pre- and post-debulking RF data collection scans. Fibrous areas were marked in green, fibro-fatty in yellow, dense calcium in white, and necrotic core in red. Finally, the predicted plaque composition was displayed as a colorcoded tissue map. Virtual histology analysis was performed by an experienced analyst who was blinded to the in vitro histopathologic analysis. VH IVUS-histopathology correlation analysis. Correlation analysis methodology is shown in Figure On the post-debulking gray scale IVUS, the proximal end of the debulking site was identified, and its distance from a proximal side branch (landmark) was accurately measured. In order to identify the proximal end of the Figure 2. Schematic description of the Virtual Histology (VH) intravascular ultrasound procedure and in vitro histopathology correlation analysis for a heart rate of about 60 beats/min. After the proximal end of debulking was detected, the longitudinal distance from the proximal end to the most proximal VH image (distance 0) was calculated. Knowing the R-R interval (s) and the pullback speed of the intravascular ultrasound catheter (0.5 mm/s), the distance between each VH images was calculated as follows: distance between two VH images (mm) R-R interval 0.5. The distance between sections was calculated as 0.5 mm ( each distance of cutting tissue) times the ratio of pre-fixed length of tissue sample (mm) to post-fixed length (mm). The VH image that was closest to each section was chosen as the corresponding color-coded map, and the debulked area on pre-debulking VH images was predicted by comparison of pre- with post-debulking VH images. The predicted debulking area was compared visually with the histology section to assess the presence or absence of the four different components.

4 2408 Nasu et al. JACC Vol. 47, No. 12, 2006 In Vivo Virtual Histology June 20, 2006: debulking site on the pre-debulking IVUS images, this distance was measured from the same landmark. 2. Knowing the R-R interval (s) and the pullback speed of the IVUS catheter (0.5 mm/s), the distance between each VH images was calculated as follow: distance between two VH images (mm) R-R interval Knowing the interval (s) between the proximal end of debulking site and the nearest top of R-wave, its distance could be calculated as follows: distance 0 interval The distance between histology sections was calculated as 0.5 mm ( distance of cutting tissue) times the ratio of pre-fixed length of tissue sample (mm) to post-fixed length (mm) in order to correct for a 20% to 30% tissue sample shrinking during post-processing (8). 5. The VH image that was closest to each section was chosen as the corresponding color-coded map. 6. The debulked area on pre-debulking VH images was predicted by comparison of pre- with post-debulking VH images. 7. Histology sections and correlating VH images were assessed visually for the presence or absence of the four different components. Statistical analysis. Continuous data are represented as mean values SD. Categoric data are expressed as frequencies of occurrence, and differences between groups were compared with chi-square tests. Comparison of continuous variables were performed by two-tailed unpaired Student t test for normally distributed variables and by Mann- Whitney test for variables with skewed distribution. Statview version 5.0 (Abacus Concepts Inc., Berkeley, California) was used for data analysis. A probability value of 0.05 was considered to indicate statistical significance. The results from VH images were validated with the corresponding histopathology to determine predictive accuracy, sensitivity, and specificity from widely accepted equations in biomedical literature (9). RESULTS Baseline characteristics and procedural data. Between May 2004 and July 2005, 30 patients (15 patients with stable AP and 15 patients with ) met eligibility criteria for this study and were enrolled. Baseline demographic and clinical data are summarized in Table 1. There were no significant differences between the two groups. Baseline lesion characteristics and procedural data are summarized in Table 2. All lesion characteristics and debulking pressure were well matched between the two groups. No complications, including perforation, no-reflow, need for emergency bypass surgery or death, occurred during DCA. Correlation of in vivo and in vitro histology. The predictive accuracy, sensitivity, and specificity for correlation analysis are shown in Table 3. The mean longitudinal gap between the VH image and the in vitro histology section was mm (range, 0 to 0.32 mm). Fibrous, fibro-fatty, necrotic core, and dense calcium regions were classified with high predictive accuracies of 87.1%, 87.1%, 88.3%, and 96.5%, respectively, using data from all patients. This trend was also observed in both subgroups. Representative examples of color-coded tissue maps of VH were compared, and their comparison with the corresponding hematoxylin and eosin stained sections are shown in Figure 3. Gray scale IVUS could not differentiate necrotic core and fibro-fatty plaque (labeled as low-density area; cases 1a and 2a). In contrast, color-coded maps obtained by VH IVUS analysis could distinguish fibro-fatty plaque (labeled as yellow area; cases 1b and 2b) and necrotic core (as red area; cases 1b and 2b) with dense calcium as white. The predicted debulking area (within blue circle) showed a favorable correlation with in-vitro histopathology (cases 1d and 2d). Tissue samples and in vitro histopathologic findings. The mean length of debulked tissue samples was mm, and there was no significant difference when comparing the two groups (stable AP, mm vs Table 1. Baseline Demographic and Clinical Data (n 30) (n 15) (n 15) p Value Age, yrs (mean SD) Men, % Systemic hypertension, % Diabetes mellitus, % Hyperlipidemia, % Current smoker, % Prior PCI, % Prior myocardial infarction, % Prior coronary bypass, % LVEF, % (mean SD) Values are mean SD when appropriate. A chi-square test was used for categorical data: unpaired Student t test or Mann-Whitney rank sum test was used for continuous data between stable AP and groups. Values of p 0.05 were considered statistically significant. acute coronary syndrome; AP angina pectoris; LVEF left ventricular ejection fraction; PCI percutaneous coronary intervention.

5 JACC Vol. 47, No. 12, 2006 June 20, 2006: Nasu et al. In Vivo Virtual Histology 2409 Table 2. Baseline Lesion Characteristics and Procedural Data (n 30) (n 15) (n 15) p Value Vessel treated, % LAD LCX RCA LMT AHA/ACC type, % A/B B2/C Calcified, % Ostial, % Eccentric, % Thrombus, % Lesion length, mm Reference diameter, mm Minimal luminal diameter, mm Diameter stenosis, % Vessel CSA, mm Lumen CSA, mm Percent PA, % Heart rate, beats/min (range) (45 82) (46 82) (45 80) 0.82 Debulking pressure, psi Values are mean SD when appropriate. A chi-square test was used for categorical data; unpaired Student t test or Mann-Whitney rank sum test was used for continuous data between stable AP and groups. Values of p 0.05 were considered statistically significant. ACC American College of Cardiology; acute coronary syndrome; AHA American Heart Association; AP angina pectoris; CSA cross-sectional area; LAD left anterior descending artery; LCX left circumflex artery; LMT left main trunk; PA plaque plus media CSA; RCA right coronary artery mm, p 0.12). Thrombi were observed more frequently in the group than in the stable AP group (26.7% cases of stable AP and all cases of patients). The frequency of each tissue component presence is illustrated in Table 4. A total of 307 sections were obtained (stable AP, n 144;, n 163). The presence of necrotic core tissue was observed more frequently in the group than in the stable AP group (22.6% vs. 12.5%, p 0.02). In contrast, there were no significant differences in the frequency of the other plaque components when comparing the two groups. In vivo VH images. The frequency of each plaque component at the predicted debulking area, identified from the comparison between pre- and post-debulking VH images, is illustrated in Table 4. The presence of necrotic core and dense calcium were observed more frequently in the group than in the stable AP group (24.5% vs. 10.4%, p 0.002; 11.1% vs. 4.1%, p 0.03). DISCUSSION This is the first clinical study to assess the accuracy of in vivo histology for the diagnosis of plaque composition. Our single-center experience showed that in vivo tissue characterization by IVUS-based RF analysis favorably correlated with the results of in vitro histopathologic examination of tissue samples obtained by DCA. Thus, the use of VH IVUS for differentiating components of atherosclerotic tissue was achieved with high predictive accuracy. Characterization of plaque components. Several characteristics inherent to IVUS imaging offer potential advantages in the evaluation of coronary disease. This tomographic orientation is able to visualize the full circumference of the vessel wall, examine arterial remodeling, and assess the thickness and echogenicity of atherosclerotic plaques (10 13). However, identification of atherosclerotic plaque components by densitometric category of gray scale IVUS is Table 3. Results of Virtual Histology and In Vitro Histology Correlation Analysis Sensitivity Specificity Predictive Accuracy (n 163) (n 163) (n 163) FT FF NC DC Sensitivity, specificity, and predictive accuracy are shown in %. acute coronary syndrome; AP angina pectoris; DC dense calcium; FF fibro-fatty; FT fibrous tissue; NC necrotic core.

6 2410 Nasu et al. JACC Vol. 47, No. 12, 2006 In Vivo Virtual Histology June 20, 2006: Figure 3. Pre-debulking gray scale intravascular ultrasound images, color-coded maps of pre- and post-debulking target lesions reconstructed by Virtual Histology intravascular ultrasound, and histology sections. (Case 1) Acute coronary syndrome. (1a) Gray scale intravascular ultrasound image at the target lesion. Note the heterogeneity of the plaque beside the plate of calcium with acoustic shadow. (1b) Pre-debulking color-coded map of 1a reconstructed by virtual histology intravascular ultrasound. Note the superficial necrotic core (red) beside the plate of the dense calcium (white). (1c) Post-debulking color-coded map. Predicted debulking area within the blue circle in 1b. (1d) Histologic finding with fibrous tissue, fibro-fatty plaque, and necrotic core. (Case 2) Stable angina pectoris. (2a) Gray scale intravascular ultrasound image of target lesion. Note the heterogeneity of the plaque. (2b) Pre-debulking color-coded map of 2a reconstructed by virtual histology intravascular ultrasound. Note the necrotic core (red) with thick fibrous cap consisting of fibrous tissue (green) and fibro-fatty plaque (yellow). (2c) Post-debulking color-coded map. Predicted debulking area within the blue circle in 2b. (2d) Histologic finding with fibrous tissue and fibro-fatty plaque. Bars 500 m. limited because of processing of the raw RF data (time-gain compensation, logarithmic compression, and envelope detection, and so on), and also interpretation must rely on simple visual inspection of acoustic reflections to determine plaque component. In previous in vivo or ex vivo studies, calcified and fibrous plaques were well identified by their hyperechoic appearance and homogeneous echocardiographic reflection (1 3,14 16). However, discrimination between lipid-containing and mixed (fibro-lipidic plaque), labeled as a region of low-density in gray scale IVUS remains difficult to achieve (1,2,17). Besides, analysis of tissue behind a calcification is difficult of signal attenuation. In a previous study of the correlation of gray scale IVUS image with in vitro histopathology of an atherectomy sample, gray scale IVUS could not differentiate plaque compositions (18). However, previous ex vivo studies showed that characterization of different plaque component is feasible with the analysis of IVUS RF data (4 7). Further, this modality had potential clinical applications, as colorcoding of plaque components allowed real-time evaluation during the procedure (Fig. 3). The present data from the in vivo IVUS RF analysis and the in vitro histopathologic images correlated well, indicating that VH IVUS is a useful modality for real-time characterization of clinically relevant plaque components. However, although the frequency of dense calcium in the in vitro histology was not significantly different when comparing the two groups, in vivo VH images suggested an increase in dense calcium in the group when compared with the stable AP group. Thus, although in vivo data correlated well with in vitro histology (predictive accuracy of 96.5%), the RF data analysis overestimated the frequency of calcifications. A previous report showed that the extent of calcium detected by electronbeam computed tomography was greater than would have been expected with regard to their age and gender (19). However, some post-mortem pathologic analyses of coronary arteries reported that calcium was a frequent feature of plaque rupture (20,21); others showed that ruptured plaques Table 4. Frequency of Each Plaque Phenotype Presence in In Vitro Histopathology Sections and In Vivo Virtual Histology Histopathology (n 163) p Value Virtual Histology (n 163) p Value Presence of FT, % Presence of FF, % Presence of NC, % Presence of DC, % A chi-square test was used for analysis between stable AP and groups. Values of p 0.05 were considered statistically significant. acute coronary syndrome; AP angina pectoris; DC dense calcium; FF fibro-fatty; FT fibrous tissue; NC necrotic core.

7 JACC Vol. 47, No. 12, 2006 June 20, 2006: Nasu et al. In Vivo Virtual Histology 2411 were less likely to be calcified in patients (22,23). Thus, coronary calcium is not a marker for neither unstable nor stable plaques. Possible explanations for this overestimation may be as follows: 1) the IVUS beam is approximately 300 m in longitudinal thickness, whereas the histology section is only 4- m thick. A VH image therefore includes more tissue than a histology section; 2) the presence of calcium occurred with a very low frequency in this target lesion population; and 3) artifact is colored with white because the software for VH analysis is obliged to assign one of the four colors for each pixel. Ability for detection of vulnerable plaques. The vulnerability of plaque to rupture is typically characterized by the presence of a necrotic core, which is a region of the fibroatheroma that is largely devoid of viable cells and consists of cellular debris and cholesterol clefts, a thin fibrous cap ( 65 m), and macrophage infiltration (24,25). Rupture of vulnerable plaques is defined as a necrotic core with a thin fibrous cap that is disrupted or ruptured (26 28) and their identification before rupture is an important clinical goal. In the present study, the presence of necrotic core was significantly higher in the group than in the stable AP group, and these results correlated with between VH and histopathology data. The presence of necrotic core in debulking tissue, which is the part of atherosclerotic plaque situated at the lumen border, may possibly be the sign for a thin cap fibroatheroma. The great advantage of VH IVUS is that it is based on a device that is practical for use in the clinical setting and that it generates a real-time assessment of plaque morphology. However, because this technology is based on IVUS with a maximum radial resolution of 100 m, it cannot evaluate the presence or absence of a thin fibrous cap. Various invasive and noninvasive imaging techniques have been employed to detect vulnerable plaque (29 36), and the combination of these modalities may help in overcoming their individual limitations. For example, improved results may be obtained by combining the use of anatomic methods, including IVUS, VH IVUS, elastography, and optical coherence tomography, with functional imaging methods, such as thermography. Study limitations. This study has several limitations. First, although 307 pairs of VH IVUS images and correlating histologic slices were obtained prospectively, only 30 patients from one center were involved. Study of larger patient populations from various centers is warranted to confirm these data. Second, tissue samples obtained by DCA consisted of only the superficial part of the atherosclerotic plaque, which are smaller than artery samples obtained from autopsies, and yield smaller histologic section areas. However, the cross sectional location in the atherosclerotic plaque could be identified by comparing pre- and post-procedural VH IVUS images. Third, it is possible that the extracted tissue reversed in the nosecone. Extrapolating the results from our in-vitro experiment described in the Methods section, this may have happened in one or two tissue samples of the current study. We calculated the sensitivity, specificity, and predictive accuracy that would have been obtained if a sample was reversed, and repeated it for each of the 30 samples. The results were, however, in a similar range. Fourth, for the extracted tissue sample, there is a selection bias in the kind of lesion included in the present study by the fact that DCA is not an adequate method for calcified lesions. This may explain the fact that the presence of calcium occurred with a very low frequency in the present study population. Despite this fact, correlation analysis between predicted debulking area in the in vivo histology and in the in vitro histopathology showed favorable results and high predictive accuracy. Fifth, although atherothrombi caused by plaque rupture, plaque erosion, and calcified nodules that protrude into the lumen occur in cases of (37), the present version of VH IVUS technology is unable to differentiate thrombus. This algorithm relies on the placement of two borders, namely the luminal border and the media-adventitia border, so that small thrombus within these two borders might lead incorrect tissue characterization. Sixth, the location of each VH image and histology section was identified as accurately as possible (Fig. 2). However, the pre- and post-procedural color-coded map may vary from the in-vitro histology section, because the RF data is captured at only the top of the R-wave, and the tissue sample may shrink during post-processing (8), leading to potential bias. Finally, in this study we used a mechanical IVUS catheter for the recording of the RF data. The analysis software processing these data is almost identical to the one used with the phased-array IVUS catheter, which is commercially available. In this study, an IVUS pullback was performed with both systems in 10 patients with 85 sections. In these 85 sections, the predictive accuracies for the mechanical system were 86.8% for fibrous, 83.1% for fibro-fatty, 89.6% for necrotic core, and 96.9% for dense calcium, respectively. The predictive accuracies for the phased-array system were 87.8%, 79.7%, 91.1%, and 97.8%, respectively. The results obtained by both systems are similar. Consequently, the results of this study support as well the use of VH obtained with a phased-array system. Conclusions. A color-coded mapping method using IVUS RF data analysis is useful to identify atherosclerotic plaque components of human coronary artery in vivo. This technique may play an important role in detecting vulnerable plaque. Reprint requests and correspondence: Dr. Kenya Nasu, The Department of Cardiology, Toyohashi Heart Center, 21-1 Azagobutori, Ohyama-cho, Toyohashi-city, Aichi, Japan. yuya0728@m3.kcn.ne.jp.

8 2412 Nasu et al. JACC Vol. 47, No. 12, 2006 In Vivo Virtual Histology June 20, 2006: REFERENCES 1. Potkin BN, Bartorelli AL, Gessert JM, et al. Coronary artery imaging with intravascular high-frequency ultrasound. Circulation 1990;81: Yock PG, Linker DT. Intravascular ultrasound. Looking below the surface of vascular disease. Circulation 1990;81: Nishimura RA, Edwards WD, Warnes CA, et al. Intravascular ultrasound imaging: in vitro validation and pathologic correlation. J Am Coll Cardiol 1990;16: Nair A, Kuban BD, Tuzcu EM, Schoenhagen P, Nissen SE, Vince DG. Coronary plaque classification with intravascular ultrasound radiofrequency data analysis. Circulation 2002;106: Nair A, Kuban BD, Obuchowski N, Vince DG. Assessing spectral algorithms to predict atherosclerotic plaque composition with normalized and raw intravascular ultrasound data. Ultrasound Med Biol 2001;27: Moore MP, Spencer T, Salter DM, et al. Characterisation of coronary atherosclerotic morphology by spectral analysis of radiofrequency signal: in vitro intravascular ultrasound study with histological and radiological validation. Heart 1998;79: Watson RJ, McLean CC, Moore MP, et al. Classification of arterial plaque by spectral analysis of in vitro radio frequency intravascular ultrasound data. Ultrasound Med Biol 2000;26: Siegel RJ, Swan K, Edwalds G, Fishbein MC. Limitations of postmortem assessment of human coronary artery size and luminal narrowing: differential effects of tissue fixation and processing on vessels with different degrees of atherosclerosis. J Am Coll Cardiol 1985;5: Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978;8: Erbel R, Ge J, Bockisch A, et al. Value of intracoronary ultrasound and Doppler in the differentiation of angiographically normal coronary arteries: a prospective study in patients with angina pectoris. Eur Heart J 1996;17: Mintz GS, Painter JA, Pichard AD, et al. Atherosclerosis in angiographically normal coronary artery reference segments: an intravascular ultrasound study with clinical correlations. J Am Coll Cardiol 1995;25: Lee DY, Eigler N, Luo H, et al. Effect of intracoronary ultrasound imaging on clinical decision making. Am Heart J 1995;129: Mintz GS, Pichard AD, Kovach JA, et al. Impact of preintervention intravascular ultrasound imaging on transcatheter treatment strategies in coronary artery disease. Am J Cardiol 1994;73: St. Goar FG, Pinto FJ, Alderman EL, Fitzgerald PJ, Stadius ML, Popp RL. Intravascular ultrasound imaging of angiographically normal coronary arteries: an in vivo comparison with quantitative angiography. J Am Coll Cardiol 1991;18: Gussenhoven EJ, Essed CE, Lancee CT, et al. Arterial wall characteristics determined by intravascular ultrasound imaging: an in vitro study. J Am Coll Cardiol 1989;14: Fitzgerald PJ, St. Goar FG, Connolly AJ, et al. Intravascular ultrasound imaging of coronary arteries. Is three layers the norm? Circulation 1992;86: Londero HF, Laguens R, Telayna JM, et al. Densitometric quantitative analysis of intracoronary ultrasound images: anatomopathological correlation. Int J Card Imaging 1997;13: Okimoto T, Imazu M, Hayashi Y, Fujiwara H, Ueda H, Kohno N. Atherosclerotic plaque characterization by quantitative analysis using intravascular ultrasound: correlation with histological and immunohistochemical findings. Circ J 2002;66: Raggi P, Callister TQ, Cooil B, et al. Identification of patients at increased risk of first unheralded acute myocardial infarction by electron-beam computed tomography. J Am Coll Cardiol 2000;36: Farb A, Burkle AP, Tang AL, et al. Coronary plaque erosion without rupture into a lipid core. A frequent cause of coronary thrombosis in sudden coronary death. Circulation 1996;93: Taylor AJ, Burke AP, O Malley PG, et al. A comparison of the Framingham risk index, coronary artery calcification, and culprit plaque morphology in sudden cardiac death. Circulation 2000;101: Gertz SD, Roberts WC. Hemodynamic shear force in rupture of coronary arterial atherosclerotic plaques. Am J Cardiol 1990;66: Cheng GC, Loree HM, Kamm RD, Fishbein MC, Lee RT. Distribution of circumferential stress in ruptured and stable atherosclerotic lesions. A structural analysis with histopathological correlation. Circulation 1993;87: Falk E. Stable versus unstable atherosclerosis: clinical aspects. Am Heart J 1999;138:S Virmani R, Burke AP, Farb A, Kolodgie FD. Pathology of the unstable plaque. Prog Cardiovasc Dis 2002;44: Libby P, Geng YJ, Akikawa M, et al. Macrophages and atherosclerotic plaque stability. Curr Opin Lipidol 1996;7: Zaman AG, Helft G, Worthley SG, Badimon JJ. The role of plaque rupture and thrombosis in coronary artery disease. Atherosclerosis 2000;149: Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation 1995;92: Yabushita H, Bouma BE, Houser SL, et al. Characterization of human atherosclerosis by optical coherence tomography. Circulation 2002;106: Stefanadis C, Toutouzas K, Tsiamis E, et al. Increased local temperature in human coronary atherosclerotic plaques: an independent predictor of clinical outcome in patients undergoing a percutaneous coronary intervention. J Am Coll Cardiol 2001;37: Buschman HP, Motz JT, Deinum G, et al. Diagnosis of human coronary atherosclerosis by morphology-based Raman spectroscopy. Cardiovasc Pathol 2001;10: Correia LC, Atalar E, Kelemen MD, et al. Intravascular magnetic resonance imaging of aortic atherosclerotic plaque composition. Arterioscler Thromb Vasc Biol 1997;17: Fitzgerald PJ, St. Goar FG, Connolly AJ, et al. Intravascular ultrasound imaging of coronary arteries. Is three layers the norm? Circulation 2001;104: Chen J, Tung CH, Mahmood U, et al. In vivo imaging of proteolytic activity in atherosclerosis. Circulation 2002;105: Jain KK. Nanodiagnostics: application of nanotechnology in molecular diagnostics. Expert Rev Mol Diagn 2003;3: Ropers D, Baum U, Pohle K, et al. Detection of coronary artery stenoses with thin-slice multi-detector row spiral computed tomography and multiplanar reconstruction. Circulation 2003;107: Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons from sudden coronary death: a comprehensive morphological classification scheme for atherosclerotic lesions. Arterioscler Thromb Vasc Biol 2000;20: