Classification of Human Coronary Atherosclerotic Plaques Using Ex Vivo High-Resolution Multicontrast-Weighted MRI Compared With Histopathology

Transcription

1 ardiopulmonary Imaging Original Research Li et al. MR lassification of oronary therosclerotic Plaques ardiopulmonary Imaging Original Research Tao Li 1,2 Xin Li 1 Xihai Zhao 2 Weihua Zhou 1 Zulong ai 2 Li Yang 2 itao Guo 3 Shaohong Zhao 2 Li T, Li X, Zhao X, et al. Keywords: acute coronary syndrome, atherosclerosis, coronary artery disease, coronary atherosclerotic plaques, coronary vessel, histopathology, MRI DOI: /JR Received January 8, 2011; accepted after revision September 24, Department of Radiology, The General Hospital of hinese People s rmed Police Forces, eijing, hina. 2 Department of Radiology, hinese PL General Hospital, No. 28 Fuxing Rd, eijing , hina. ddress correspondence to S. Zhao (zhaoshaohong@yahoo.com.cn). 3 Department of Pathology, hinese PL General Hospital, eijing, hina. JR 2012; 198: X/12/ merican Roentgen Ray Society lassification of Human oronary therosclerotic Plaques Using Ex Vivo High-Resolution Multicontrast-Weighted MRI ompared With Histopathology OJETIVE. The objective of our study was to evaluate the feasibility of ex vivo highresolution multicontrast-weighted MRI to accurately classify human coronary atherosclerotic plaques according to the merican Heart ssociation classification. MTERILS ND METHODS. Thirteen human cadaver heart specimens were imaged using high-resolution multicontrast-weighted MR technique (T1-weighted, proton density weighted, and T2-weighted). ll MR images were matched with histopathologic sections according to the landmark of the bifurcation of the left main coronary artery. The sensitivity and specificity of MRI for the classification of plaques were determined, and ohen s kappa analysis was applied to evaluate the agreement between MRI and histopathology in the classification of atherosclerotic plaques. RESULTS. One hundred eleven MR cross-sectional images obtained perpendicular to the long axis of the proximal left anterior descending artery were successfully matched with the histopathologic sections. For the classification of plaques, the sensitivity and specificity of MRI were as follows: type I II (near normal), 60% and 100%; type III (focal lipid pool), 80% and 100%; type IV V (lipid, necrosis, fibrosis), 96.2% and 88.2%; type VI (hemorrhage), 100% and 99.0%; type VII (calcification), 93% and 100%; and type VIII (fibrosis without lipid core), 100% and 99.1%, respectively. Isointensity, which indicates lipid composition on histopathology, was detected on MRI in 48.8% of calcified plaques. greement between MRI and histopathology for plaque classification was 0.86 (p < 0.001). ONLUSION. Ex vivo high-resolution multicontrast-weighted MRI can accurately classify advanced atherosclerotic plaques in human coronary arteries. oronary artery disease is the leading cause of morbidity and mortality in the world. Every year, more than 19 million people worldwide die of an acute coronary event [1]. large percentage of these people appeared healthy and had no prior symptoms [2]. Therefore, identification of individuals who are at high risk of an acute coronary event is of paramount clinical importance. number of previous studies have shown that rupture of coronary atherosclerotic plaques with subsequent thrombosis is the major cause of acute coronary syndrome (S) [3, 4]. ecause the composition of coronary atherosclerotic plaques has shown more predictive value for plaque rupture, thrombosis, and ischemic events than the severity of lumen stenosis [5 7], it is clinically important to characterize the composition of coronary atherosclerotic plaques. Previous studies have shown that in vivo and ex vivo multicontrast-weighted MR images can be used to characterize plaque composition in the great vessels, such as the carotid artery and aorta [8 10]. ecause the classification of plaques is fundamentally based on the composition of the lesion, multicontrast-weighted MRI might play an important role in the noninvasive in vivo classification of plaque type. ai et al. [10] reported that in vivo multicontrast-weighted MRI can accurately classify human carotid atherosclerotic plaques according to a modified histologic classification scheme based on the guidelines of the merican Heart ssociation (H) [11, 12]. However, in vivo imaging of the coronary artery wall is still challenging because of respiratory and cardiac motion artifacts and the tortuous course and smaller size of the vessel. nother limitation of in vivo identification of coronary JR:198, May

2 Li et al. atherosclerotic plaque composition by MRI is its poor spatial resolution. Therefore, the characterization of coronary plaque composition using MRI has not been well studied because of the technical difficulties. This study sought, first, to determine the feasibility of using ex vivo multicontrastweighted MRI to classify human coronary atherosclerotic plaques compared with histology; and, second, to determine whether multicontrast-weighted MRI may potentially direct the development and optimization of in vivo MR vessel wall imaging techniques for the characterization of coronary atherosclerotic plaques. Materials and Methods Study Sample This study was approved by the institutional review board of the hinese PL General Hospital. The hearts from 13 consecutive autopsy subjects (12 men and one woman; age range at death, years; mean, 81.5 ± 18.2 [SD] years; cause of death, malignant tumors in five cases, infection in four, myocardial infarction in three, and trauma in one) underwent coronary MR examination. The time interval between patient death and MRI was less than 24 hours. fter being removed from the body, each heart was flushed with water to clear clots within the chambers and coronary arteries; then, a three-way tube was inserted and fixed into the left main coronary artery through the left coronary sinus and imaged immediately. oronary rtery MRI Protocol Multicontrast-weighted MRI was performed on a 1.5-T whole-body scanner (Signa Excite, GE Healthcare) with a 2.5-inch (6.4-cm) temporomandibular joint surface coil. Three-plane images were acquired for localizing heart orientation. 3D fast steady-state free precession sequence was used to image the left anterior descending artery (LD) for scouting the location of 2D axial vessel wall imaging. Three different contrast-weighted cross-sectional imaging sequences T1-weighted, proton density weighted, and T2-weighted were performed starting from the bifurcation of the left main coronary artery perpendicular to the long axis of the proximal LD. The proximal 40 mm of the LD was covered with each imaging pulse sequence and image slices per LD were generated. To suppress the signal in the vessel lumen and improve the contrast between the vessel wall and lumen, 20 ml of diluted contrast medium containing superparamagnetic iron oxide particles (ferucarbotran [Resovist, ayer Schering Pharma]) was injected into the LD via the three-way tube, and T1-weighted, proton density weighted, and T2-weighted images were acquired. The optimal proportion of MR contrast agent and saline solution was 1% and 99%, respectively. Fat suppression was used to reduce the signal from the epicardial fatty tissues. The imaging parameters were as follows for the T1-weighted sequence: TR/TE, 440/21; FOV, 12 9 cm; slice thickness, 2 mm; matrix, ; in-plane spatial resolution, mm; and number of signals acquired, 2. The imaging parameters for the proton density weighted sequence were as follows: TR/ TE, 2000/21; FOV, 12 9 cm; slice thickness, 2 mm; matrix, ; in-plane spatial resolution, mm; and number of signals acquired, 2. The parameters for the T2-weighted sequence were as follows: TR/TE, 4500/105; FOV, 12 9 cm; slice thickness, 2 mm; matrix, ; in-plane spatial resolution, mm; and number of signals acquired, 2. Histopathologic Examination fter MRI examination, the hearts were immediately fixed in a formalin solution. One week later, the left main coronary arteries and the anterior descending branches were separated from the hearts. Starting from the bifurcation of the left main coronary artery, the proximal 40 mm of the LD was sectioned into 10 tissue blocks, each of 4 mm in thickness, and the distance from the tissue block to the bifurcation of the left main coronary artery was recorded to match the MR images. fter decalcification, the tissue blocks were embedded in paraffin and prepared for histopathologic examination in 4-μm-thick sections. locks were stained with H and E. MR Evaluation of oronary therosclerotic Plaques n experienced radiologist blinded to the histopathology results evaluated the original 2D cross-sectional MR images. For each image location, three images (contrast-enhanced T1-weighted, proton density weighted, T2-weighted) were available to be reviewed for identification of the lesion type. To decrease dependence on the lesion type, MR images of the proximal 40 mm of the LD were reviewed with a 4-mm interval, which meant that odd slices (1, 3, 5 ) or even slices (2, 4, 6 ) were reviewed. This protocol generated 7 10 image locations per LD that could be compared with histopathologic sections. The thickness of the vessel wall was measured, and the presence or absence of calcification, a lipid-rich necrotic core, and intraplaque hemorrhage was determined. The plaques were classified according to the H criteria and the modified histologic classification of ai et al. and others [10, 12] (Table 1). The sensitivity and specificity of MRI for plaque classification were then determined. ccording to ai et al. [10], MRI cannot distinguish type I from type II lesions because it does not distinguish between the dispersed foam cells in type I lesions and the multiple foam cell layers of fatty streaks in type II lesions. Therefore, these lesions were combined into one category and were named type I II lesions. In addition, MRI cannot differentiate fibrous caps made of glycoprotein (type IV lesions) from those containing mainly collagen (type V lesions). Therefore, these lesions were named type IV V lesions. MRI was mainly used to identify the lipid core, fibrosis, calcification, and hemorrhage. Type I II lesions are indistinguishable from the normal. Type III lesions are characterized by a slightly diffuse or eccentric thickening of the wall containing pools of extracellular lipid. On T1-weighted and proton density weighted imaging, the lesions appear focal and show slightly high signal intensity. Type IV V lesions show high or equal signal intensity on T1-weighted and proton density weighted imaging and low or equal signal intensity on T2-weighted imaging. Type VI lesions show high signal intensity on T1-weighted, proton density weighted, and T2-weighted images because of recent intraplaque hemorrhage. Type VII lesions show signal loss on all three sequences because calcification is the predominant component in this kind of lesion. Type VIII, a highly fibrotic lesion, has equal signal intensity on T1-weighted images and high or equal signal intensity on T2- weighted images. nalysis of Histopathologic Sections The histopathologic sections were analyzed by an experienced pathologist who was blinded to the MRI results, and plaques were classified according to H criteria [11, 12]. orrelation etween MRI and Histopathology MR images and histopathologic sections were independently reviewed and classified. The left main coronary artery bifurcation was used as a landmark. The relative distance between the images and the left main coronary artery bifurcation and some morphologic features such as vessel shape and lumen size, the location of the LD branches, and calcifications were used to match the MR images with the histopathologic sections. Statistical nalysis The sensitivity, specificity, and accuracy of MRI for classifying the plaques were calculated. ohen s kappa test was used to analyze the agreement of the plaque classification between the MRI and histopathologic findings JR:198, May 2012

3 MR lassification of oronary therosclerotic Plaques TLE 1: Typical Morphologic ppearances of Different Plaque Types Using MRI ccording to the merican Heart ssociation (H) lassification and Modified H lassification of ai et al. [10] Definitions of Plaque Types Signal Intensity Plaque Type H lassification Modified H lassification for MRI T1-Weighted MRI T2-Weighted MRI I Initial lesion with foam cell II Fatty streak with multiple foam cell layers I II III IV Preatheroma with extracellular lipid pools theroma with a confluent extracellular lipid core Results The 13 hearts were all intubated successfully, and the quality of the images was satisfactory. total of 111 MR images of 13 hearts were matched with the corresponding histopathologic sections. mong the histopathologic sections, there were 15 type I II (near normal) plaque sections (13.5%) (Fig. 1), 10 type III (focal lipid pool) plaque sections (9%), 26 type IV V (lipid, necrosis, fibrosis) Near-normal wall thickness, no calcification Diffuse intimal thickening or small eccentric plaque with no calcification plaque sections (23.4%) (Fig. 2), 13 type VI (hemorrhage) plaque sections (11.7%) (Fig. 3), 43 type VII (calcification) plaque sections (38.7%) (Fig. 4), and four type VIII (fibrosis without lipid core) plaque sections (3.6%). The accuracy of MRI for the classification of plaques was 89.2% (99/111), and the distribution of plaques is shown in Table 2. The sensitivity and specificity of MRI for identifying different types of plaques were Hyperintense Iso- to hyperintense Hypointense Hypo- to isointense V Fibroatheroma Iso- to hyperintense Hypo- to isointense IV V Plaque with a lipid or necrotic core Iso- to hyperintense Hypo- to isointense surrounded by fibrous tissue with possible calcification VI omplex plaque with possible omplex plaque with possible surface Hyperintense a Hyperintense a surface defect, hemorrhage, or thrombus defect, hemorrhage, or thrombus VII alcified plaque alcified plaque Hypointense Hypointense VIII Fibrotic plaque without lipid core Fibrotic plaque without lipid core with possible small calcifications Isointense Iso- to hyperintense a ecause of subacute hemorrhage. as follows: type I II, 60% and 100%; type III, 80% and 100%; type IV V, 96.2% and 88.2%; type VI, 100% and 99.0%; type VII, 93% and 100%; and type VIII, 100% and 99.1%, respectively. In addition, isointensity in the center of dense calcification, which represented the lipid composition pathologically, was observed on T1- weighted, proton density weighted, and T2- weighted images in 21 of 43 type VII sections Fig. 1 Type I lesion in explanted heart. and, On T1-weighted () and T2-weighted () MR images, (arrow) appears normal. sterisk = lumen., Photomicrograph (H and E, 10) of pathologic section shows that intima is thickened; foam cells are detected. sterisk = lumen. JR:198, May

4 Li et al. (48.8%) (Fig. 5). ohen s kappa test showed that the agreement between MRI and histopathology findings for plaque classification was 0.86 (U = 22.16, p < 0.001), indicating that MRI had good agreement with histopathology. Discussion In this study, we performed ex vivo highresolution multicontrast-weighted MRI of the coronary artery and compared the MR images with histopathologic results. The major findings of the study were twofold: First, we found that the classification of human coronary atherosclerotic plaque types using ex vivo high-spatial-resolution multicontrastweighted MRI closely agrees with histopathologic results, with a kappa test value of 0.86; Fig. 2 Type IV V lesions in explanted heart. and, MR images show irregular thickening of left anterior descending artery wall (arrow). This thickening shows slightly high or equal signal on T1-weighted image () and slightly low signal on T2-weighted image (). sterisk = lumen., Photomicrograph (H and E, 4) of pathologic section shows extracellular lipid has accumulated into lipid pool. sterisk = lumen. Fig. 3 Type VI lesion with intraplaque hemorrhage in explanted heart. and, Plaque (arrow) presents high signal intensity on MR proton density weighted () and T2-weighted () images. sterisk in = lumen., Hemorrhage in necrotic core is evident in pathologic section (H and E, 10). sterisk = lumen. second, we found that there are isointense lipid regions within densely calcified type VII plaques on MRI. Single contrast-weighted MR images alone cannot accurately identify the compositions of coronary atherosclerotic plaque and classify the plaques. In a previous study of the classification of atherosclerotic plaques, Serfaty et al. [13] found that ex vivo high-resolution T2- weighted alone could help to accurately classify only the fibrocalcific plaque (type VII) but was of limited accuracy for classification and analysis of the other types of atherosclerotic plaques. Kawasaki et al. [14] considered that hyperintense plaques on unenhanced T1- weighted images may represent intraplaque hemorrhage or lipid-rich necrotic cores. It is difficult to differentiate intraplaque hemorrhage from the lipid component using only T1-weighted imaging. In this study, multicontrast-weighted images were crucial for determining the compositions of the plaques and for accurate classification. In an ex vivo study of the classification of coronary atherosclerotic plaques, Nikolaou et al. [15] analyzed MR images in a nonblind retrospective fashion by matching them with histopathologic sections. Their study did not include early lesions (type I II) and complicated lesions (type VI). In the current study, the types of human coronary atherosclerosis lesions were classified and differentiated in a prospective blinded fashion and all lesion types were included. High 1072 JR:198, May 2012

5 MR lassification of oronary therosclerotic Plaques Fig. 4 Type VII lesion, densely calcified plaque, in explanted heart. and, alcified plaque (arrow) presents low signal on proton density weighted () and T2-weighted () images. sterisk = lumen., alcification is evident in pathologic section (H and E, 4). sterisk = lumen. TLE 2: MRI and Histopathologic lassifications of oronary therosclerotic Plaques a Plaque Type Plaque Type ccording to Pathology Findings Total No. of ccording to MRI I II III IV V VI VII VIII Plaques I II 9 9 III 8 8 IV V VI VII VIII Total Note Dash ( ) indicates 0 lesions. a greement between MRI and histopathology: κ = 0.86, ohen s test. sensitivity and high specificity of MRI were obtained for type III through type VIII (type III, 80% and 100%; type IV V, 96.2% and 88.2%; type VI, 100% and 99.0%; type VII, 93% and 100%; and type VIII, 100% and 99.1%, respectively). However, the sensitivity for type I II lesions was low (60%). Nine type I II lesions and eight type III lesions were correctly classified on MR images, but six type I II lesions and two type III lesions were mistaken for type IV V lesions because of a diffuse thickened wall, which was caused by the distorted coronary artery and the adaptable endothelial thickening. Multicontrast-weighted MRI can accurately differentiate lipid-necrotic cores (type IV V) from recent hemorrhage (type VI). Recent hemorrhage showed as a high signal on all contrast-weighted MR images. Type IV V lesions showed a high or equal signal intensity on T1-weighted and proton density weighted images and a low or equal signal intensity on T2-weighted images. In this study, the sensitivity and specificity for type IV V and type VI lesions were high. n isointense lipid area could be detected in densely calcified plaques with ex vivo high-resolution multicontrast-weighted MRI in the current study. Type VII lesions were readily detected by MRI: They were predominant with the calcified component showing an irregular low signal on all contrast-weighted MR sequences. Interestingly, isointensity in the center of densely calcified plaques, which represents the lipid component pathologically, was detected on T1-weighted, proton density weighted, and T2-weighted images among 48.8% of the type VII lesions (21/43). To our knowledge, this article is the first report that MRI can reveal isointensity in the center of type VII lesions in the human coronary artery. In atherosclerotic plaques, cholesterol and calcium phosphate coexist, the latter in the form of hydroxyapatite. holesterol and hydroxyapatite can result in the formation of a large number of small cholesterol crystals mixed with large hydroxyapatite crystals [16]. This mixture can explain the lipid composition of densely calcified plaques on MRI. oronary calcification might indicate an attempt by the arterial wall to stabilize itself and minimize the risk of plaque rupture. However, increased stress at the interface between calcified and noncalcified areas may lead to plaque rupture [17]. Ota and colleagues [18] reported heavily calcified carotid atherosclerotic plaques with hemorrhage. It needs to be confirmed whether heavily calcified plaques with a lipid composition in coronary arteries remain unstable because of their complex composition and whether these plaques influence hemodynamics. The results of this study show that ex vivo high-resolution multicontrast-weighted MRI may enable characterization of the composition of coronary atherosclerotic plaques and the classification of plaques. However, it is still challenging to identify coronary plaque components in vivo noninvasively because of respiratory and cardiac motion artifacts, the tortuous course and smaller size of the vessel, incomplete blood suppression, and poor spatial resolution with current MRI techniques. Recently, other investigators evaluated coronary plaque burdens on MRI by measuring vessel wall thickness [19, 20]. dditionally, in recent pilot studies, T1-weighted inversion recovery MRI techniques have been used to detect hyperintensity in coronary plaques associated with S [14, 21, 22]. These investigations of coronary plaque JR:198, May

6 Li et al. features suggest that MRI may enable assessment of the vulnerability of coronary atherosclerotic plaques in vivo with improvements in fast MRI techniques. This study has some limitations. First, because MRI is time-consuming, only the proximal 40 mm of the LD from each heart was selected for comparison. Second, the number of fibrous plaques was limited because type VIII lesions are rare. The small number of type VIII lesions might have resulted in some bias. onclusions Ex vivo high-resolution multicontrastweighted MRI can accurately reveal plaque composition, and findings can be used to classify human coronary advanced atherosclerotic plaques according to the H criteria and a modified histologic classification. These results provide a theoretic foundation for directing the development and optimization of in vivo MR vessel wall imaging techniques for characterizing coronary atherosclerotic plaques. cknowledgment We thank Xin Liu (Institute of iomedical and Health Engineering, Shenzhen Institute of dvanced Technology, hinese cademy of Science) for assistance with preparing the manuscript. References 1. Yusuf S, Reddy S, Ounpuu S, nand S. Global burden of cardiovascular disease. Part I. General considerations, the epidemiologic transition, risk factors, and impact of urbanization. irculation 2001; Fig. 5 Type VII lesion, calcified plaque with cholesterol crystals, in explanted heart. and, alcified plaque (arrow) presents low signal on T1-weighted () and proton density weighted () images, but isointensity can be seen within low-signal calcification, and lumen stenosis was more than 90%. sterisk = lumen., holesterol crystals are found in calcification in pathologic section (H and E, 10) and lumen (asterisk) is crescent. 104: Virmani R, urke P, Farb. Sudden cardiac death. ardiovasc Pathol 2001; 10: Falk E, Shah PK, Fuster V. oronary plaque disruption. irculation 1995; 92: Virmani R, urke P, Farb, Kolodgie FD. Pathology of the vulnerable plaque. J m oll ardiol 2006; 47(suppl): Fuster V, adimon L, adimon JJ, hesebro JH. The pathogenesis of coronary artery disease and the acute coronary syndromes (1). N Engl J Med 1992; 326: Fuster V, adimon L, adimon JJ, hesebro JH. The pathogenesis of coronary artery disease and the acute coronary syndromes (2). N Engl J Med 1992; 326: Virmani R, Kolodgie FD, urke P, Farb, Schwartz SM. Lessons from sudden coronary death: a comprehensive morphological classification scheme for atherosclerotic lesions. rterioscler Thromb Vasc iol 2000; 20: Yuan, Mitsumori LM, Ferguson MS, et al. The in vivo accuracy of multispectral MR imaging for identifying lipid-rich necrotic cores and intraplaque hemorrhage in advanced human carotid plaques. irculation 2001; 104: Koops, Ittrich H, Petri S, et al. Multicontrastweighted magnetic resonance imaging of atherosclerotic plaques at 3.0 and 1.5 Tesla: ex vivo comparison with histopathologic correlation. Eur Radiol 2007; 17: ai JM, Hatsukami TS, Ferguson MS, Small R, Polissar NL, Yuan. lassification of human carotid atherosclerotic lesions with in vivo multicontrast magnetic resonance imaging. irculation 2002; 106: Stary H, handler, Glagov S, et al. definition of initial, fatty streak, and intermediate lesion of atherosclerosis: a report from the ommittee on Vascular Lesions of the ouncil on therosclerosis, merican Heart ssociation. irculation 1994; 89: Stary H, handler, Dinsmore RE, et al. definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis: a report from the ommittee on Vascular Lesions of the ouncil on therosclerosis, merican Heart ssociation. irculation 1995; 92: Serfaty JM, haabane L, Tabib, hevallier JM, riguet, Douek P. therosclerotic plaques: classification and characterization with T2- weighted high-spatial-resolution MR imaging-an in vitro study. Radiology 2001; 219: Kawasaki T, Koga S, Koga N, et al. haracterization of hyperintense plaque with noncontrast T1- weighted cardiac magnetic resonance coronary plaque imaging: comparison with multislice computed tomography and intravascular ultrasound. J ardiovasc Imaging 2009; 2: Nikolaou K, ecker R, Muders M, et al. Multidetector-row computed tomography and magnetic resonance imaging of atherosclerotic lesions in human ex vivo coronary arteries. therosclerosis 2004; 174: Hirsch D, zoury R, Sarig S, Kruth HS. olocalization of cholesterol and hydroxyapatite in human atherosclerotic lesions. alcif Tissue Int 1993; 52: Wexler L, rundage, rouse J, et al. oronary artery calcification: pathophysiology, epidemiology imaging methods, and clinical implications a statement for health professionals from the merican Heart ssociation Writing Group. ir JR:198, May 2012

7 MR lassification of oronary therosclerotic Plaques culation 1996; 94: Ota H, Yarnykh VL, Ferguson MS, et al. arotid intraplaque hemorrhage imaging at 3.0-T MR imaging: comparison of the diagnostic performance of three T1-weighted sequences. Radiology 2010; 254: Kim WY, strup S, Stuber M, et al. Subclinical coronary and aortic atherosclerosis detected by magnetic resonance imaging in type 1 diabetes with and without diabetic nephropathy. irculation 2007; 115: Macedo R, hen S, Lai S, et al. MRI detects increased thickness in asymptomatic individuals: the multi-ethnic study of atherosclerosis (MES). J Magn Reson Imaging 2008; 28: Tanaka, Kawasaki T, Noguchi T, et al. Hyperintense plaque with noncontrast T1-weighted magnetic resonance coronary plaque imaging leading to acute coronary syndrome. irculation 2009; 120: Oei ML, Ozqun M, Seifarth H, et al. T1-weighted MRI for the detection of coronary artery plaque haemorrhage. Eur Radiol 2010; 20: FOR YOUR INFORMTION The comprehensive book based on the RRS 2012 annual meeting categorical course on Pitfalls in linical Imaging is now available! For more information or to purchase a copy, see JR:198, May