Cardiac MRI is well established in the clinical routinefor assessment of myocardial viability, diagnosis of car-diomyopathies and characterization of cardiac tumors.Evidence based guidelines regarding indications andimaging-protocols are available for a long time. Recentstudies have generated new evidence for perfusion MRIbeing superior to SPECT, which was previously appliedfor many patients. The reasons for the superiority of MRIare among others the high special resolution, which al-lows for assessment of subendocardial perfusion deficits.The updated guidelines already recommend MRI for as-sessment of myocardial perfusion.

T1 and T2 mapping and calculation of extracellular vol-ume fraction have recently experienced a Renaissancedue to improved sequence design and analytic tools.These quantitative measurements supplement the num-ber of tools for characterization of the myocardium andprovide a means for assessment of the degree of tissuefibrosis (T1) and edema (T2). Mapping is promising forevaluation of the severity of cardiac involvement in manydiseases in order to gain information regarding the prog-nosis and to monitor the effect of therapy. Currently, stan-dard values are generated and myocardial diseases are

systematically evaluated in order to gain evidence forthreshold values indicating pathologic changes.

MR-guided cardiovascular interventions comprisemany potential benefits, such as the inherent excellentsoft tissue contrast, direct delineation of target structureslike valves, coronary arteries and cardiac veins and my-ocardial scars and the possibility to measure flow andcardiac function. In recent years, refined animal studieshave been conducted and small series of patients havebeen treated. Transfer to the clinical arena still remainslimited to a small number of centers worldwide. The twomajor obstacles are the demanding environment in theMR suite and the lack of MR safe interventional devices.Development of MR-safe devices is expensive and will bepursued only if broad application can be expected. Thiscan be assumed only for interventions that provide para-mount advantages compared to conventional X-ray guid-ance. The development of interventions with high clinicalimpact which cannot be performed using other imagingmodalities are indispensable. The most promising indica-tions are treatment of patients with congenital heart dis-ease, electrophysiology and local delivery of drugs orcells to the myocardium.

160 KCR 2015

Cardiovascular S

ep 11, Fri

14:00-14:30 Grand Ballroom 104

MC 02 CV-01 Cutting edge in cardiac MR

Chairperson(s) : Tae-Hwan Lim University of Ulsan College of Medicine, Asan Medical Center, KoreaJung Im Jung The Catholic University of Korea, Seoul St. Mary’s Hospital, Korea

Recent update of cardiac MR

Gabriele Krombach University Hospital Giessen, Justus Liebig University, Germany.

gabriele.krombach@uniklinikum-giessen.de

4D flow MRI: three-directional velocity encoding + 3Danatomic coverage; 1 reference or magnitude image & 3velocity-encoded images (efficient data acquisition);strategies to reduce respiration motion (respiratory trig-gering, navigator gating, self-gating)

Major advantages over 2D PC MRI: comprehensiveretrospective evaluation of complex blood flow patterns;advanced hemodynamic measures (wall shear stress[WSS], pressure difference, pulse wave velocity, turbulentkinetic energy, vorticity, helicity, flow angle ); additionalevaluation of ventricle function & volumes using magni-tude cine data

Relatively long acquisition time: e.g., 8-12 min for theaorta & 10-20 min for whole heart coverage

Accelerated acquisition techniques enabling clinicallyfeasible scan duration: high performance gradients (short-er TE & TR); parallel imaging technology, radial under-sampling, kt-BLAST, kt-SENSE, kt-GRAPPA, com-pressed sensing increased clinical utilization & numberof publications

Limited spatial resolution (around 2 mm) restrictinganalyses in small vessels

Limited temporal resolution (40~60 ms) restrictinganalysis of some parameters requiring high temporal res-olution

Factors limiting a broader clinical use: demanding tech-nology & expertise required to implement 4D flow MRI;limited availability of 4D flow MRI sequences; shortage ofdedicated & user-friendly software packages; lack of stan-dardized pre- & post-processing tools; large amount ofacquired data

Preprocessing & correction: noise, aliasing, eddy cur-rents

3D PC MRA (by-product of 4D flow MRI): anatomicbackbone for flow analysis; non-contrast vs. contrast-en-hanced (blood-pool contrast agent; USPIO or gadofos-veset increase SNR & velocity-to-noise ratio; facilitatehighly accelerated acquisition)

Flow visualization:

1. (1) Velocity vector glyph(2) Streamlines: instantaneous 3D flow velocity vector

field for an individual cardiac time-frame; color-coded identification of high systolic flow velocities

(3) Pathlines from particle tracing: temporal evolutionof 3D flow over the cardiac cycle; color-codedvisualization of velocity changes or flow tracing

forward or backward

Retrospective flow quantification at any location withinthe 3D data volume: comparable or even improved accu-racy & precision relative to 2D PC MRI measurements;good scan-rescan reproducibility, & low inter- and intra-observer variability; equally accurate & precise for venous& arterial flow quantification at single arterial velocity en-coding value

Retrospective valve tracking to allow reliable & repro-ducible transvalvular flow assessment

4-component evaluation of the intracardiac blood flow:(1) direct flow; (2) retained flow; (3) delayed ejected flow;(4) residual volume

Clinical applications

1. (1) Bicuspid aortic valve: WSS, jet impingement an-gle better risk stratification for ascending aorticaneurysm formation & dissection

(2) Aortic coarctation: peak systolic pressure gradientacross coarctation & pressure fields derived from4D flow MRI as an alternative to invasivecatheterization

(3) Fontan circulation: hepatic flow distribution; opti-mal Fontan geometry leading to balanced hepaticflow distribution by using in vivo & in vitro 4D flowMRI & computational flow dynamics in order toavoid the risk for developing pulmonary arterove-nous fistulas; kinetic energy loss due to helicalflow formation in the pulmonary arteries of unfa-vorable Fontan geometry

(4) Repaired TOF: more & pathologic vertices in theRA & RV

(5) Hypertrophic cardiomyopathy: impact on aorticflow

(6) Others: pulmonary hypertension; one-and-a-halfrepair in patients with pulmonary atresia & intactventricular septum

Future perspective:Shorter acquisition time, user-friendly & standardized

software package, more clinical studies to find clinicallyrelevant hemodynamic markers using 4D flow MRI

References

1. Stankovic Z, Allen BD, Garcia J, et al. 4D flow imaging

Cardiovascular 161C

ardiovascular Sep 11, Fri

14:30-15:00 Grand Ballroom 104

MC 02 CV-02 Cutting edge in cardiac MR

Chairperson(s) : Tae-Hwan Lim University of Ulsan College of Medicine, Asan Medical Center, KoreaJung Im Jung The Catholic University of Korea, Seoul St. Mary’s Hospital, Korea

4D flow cardiac MRI

Hyun Woo Goo Asan Medical Center, University of Ulsan College of Medicine, Korea. ghw68@hanmail.net

with MRI. Cardiovasc Diagn Ther 2014;4(2):173-192.2. Calkoen EE, Roest AA, van der Geest RJ, et al.

Cardiovascular function and flow by 4-dimensional mag-netic resonance imaging techniques: new applications. JThorac Imaging 2014;29(3):185-196.

3. Riesenkampff E, Fernandes JF, Meier S, et al. Pressurefields by flow-sensitive, 4D velocity-encoded CMR in pa-tients with aortic coarctation. J Am Coll Cardiol Img2014;7:920-926.

4. Tariq U, Hsiao A, Alley M, et al. Venous and arterial flowquantification are equally accurate and precise with par-allel imaging compressed sensing 4D phase contrastMRI. J Magn Reson Imaging 2013;37:1419-1426.

5. Hsiao A, Tariq U, Alley MT, et al. Inlet and outlet valveflow and regurgitant volume may be directly and reliablyquantified with accelerated, volumetric phase-contrastMRI. J Magn Reson Imaging 2015;41:376-385.

6. Uribe S, Bachler P, Valverde I, et al. Hemodynamic as-sessment in patients with one-and-a-half ventricle repairrevealed by four-dimensional flow magnetic resonanceimaging. Pediatr Cardiol 2013;34:447-451.

7. Rolda′n-Alzate A, Garcl′a-Rodrl′guez S, Anagnostopou-los PV, et al. Hemodynamic study of TCPC using in vivoand in vitro 4D flow MRI and numerical simulation. JBiomech 2015;48(7):1325-1330.

162 KCR 2015

Cardiovascular S

ep 11, Fri

Cardiovascular14:00-18:00 Grand Ballroom 104

Cutting edge in cardiac MR

Chairperson(s)Tae-Hwan Lim University of Ulsan College of Medicine,

Asan Medical Center, KoreaJung Im Jung The Catholic University of Korea, Seoul

St. Mary’s Hospital, Korea

MC 02 CV-03 15:00 Hemodynamic influence of aortic dilatation inpatients with aortic stenosisHojin Ha1, Guk Bae Kim2, Jihhoon Kweon2, Young-Hak Kim2, Namkug Kim2, Dong Hyun Yang2

1Postech Biotech Center, 2University of Ulsan Collegeof Medicine, Asan Medical Center, Korea. donghyun.yang@gmail.com

PURPOSE: Although abnormal wall shear stress (WSS)distribution is suspected to have a significant role on thedevelopment of the aortic dilatation in patients with bicus-pid aortic valves (BAV), the relationship between hemo-dynamics characteristics and aortic dilatations is not fullyunderstood yet. The present study investigates the asso-ciation between the characteristics of WSS distributionsand aortic dilatations in aortic-stenosis (AS) patients withtricuspid aortic valves (TAV) and bicuspid aortic valves(BAV).MATERIALS AND METHODS: A total of 54 moderateand severe AS-patients (TAV = 32, BAV = 22) whounderwent multi-detector computed tomography (MDCT)and PC-MRI at the ascending aorta were retrospectivelycollected. Thoracic aortic diameter was measured at 10levels from aortic annulus to distal descending thoracicaorta (Fig. 2b). Based on PC-MRI of the ascending aorta(D of Fig. 2a), 2D velocity profiles were extracted in eachslice of the phase images (Fig. 1a), and then a total of360 velocity line-profiles were extracted from the center ofthe vessel to the wall with 1 degree angular increments.Differences of the aortic diameters and WSS distributionsbetween TAV and BAV were statistically analyzed usingSPSS Statistics 17.0 (SPSS, Chicago, IL, USA). RESULTS: The averaged WSS map shows that ASpatients with BAV have more asymmetric WSS distribu-tion compared to those with TAV. In particular, ASpatients with BAV has significantly higher systolic WSS atthe right-posterior region and lower systolic WSS at theleft and left-posterior regions of the ascending aorta (Fig.1b-d). However, the WSS differences between BAV-APand BAV-RP were not significant (Fig. 1e-f) In accor-dance with the increased WSS distribution, AS patientswith BAV are found to have larger diameters of theascending aorta compared to those with TAV (Fig. 2).Linear regression between the aortic diameter and sys-tolic WSS shows that the increment of the WSS repre-sents about 14% of the aortic dilatation.CONCLUSION: We found that the large aortic dilatationsin AS patients with BAV is associated with the increase ofhemodynamic WSS.CLINICAL RELEVANCE/APPLICATION: WSS can be

used to predict future risk of aortic dilatation in ASpatients.

Fig. 1. Hemodynamic wall shear stress (WSS) in aorticblood flows with TAV and BAV.

Fig. 2. Aortic dilatations at the ascending aorta.Comparison of aortic diameters with TAV and BAV. * Indicates p < 0.05.

MC 02 CV-04 15:10 Evaluation of pulmonary regurgitation with fourdimensional velocity encoding cine: new insightand quantificationJae Hyun Kim1, Whal Lee1, Eun-Ahun-Ah Park1, Jin Young Yoo2

1Seoul National University College of Medicine, 2SeoulNational University Hospital, Korea. whal.lee@gmail.com

PURPOSE: We have used the 4D VENC for hemody-namic evaluation of repaired tetralogy of Fallot (TOF)patients. With the partial seeding technique we can seethe spiral flow and turbulence flow in the pulmonaryartery. We will report the incidence of these findings andthe validation of proto type 4D flow measurement com-paring with established 2D flow measurement.MATERIALS AND METHODS: 20 patients (M:F = 14:6;average, 18 years) were included in this study. 4D VENCstudy was done as well as the 2D flow measurement atboth pulmonary arteries with 1.5T MR system (Avanto,Siemens). The spiral flows in the pulmonary artery aregraded by its rotation degree. The spiral flow is definedwhen the radial vector of the flow is rotated more than180 degree before branching of the right or left pulmonaryarteries. The net forward volume (NFV) and regurgitation

Cardiovascular 163C

ardiovascular Sep 11, Fri

fraction (RF) of both pulmonary arteries in 4D and 2Dmeasurement are compared with paired T-test and BlandAltman plotting were done.RESULTS: The pulmonary artery flow showed spiral pat-tern in 75% of the patent with repaired TOF and the rota-tion degrees were more than 270 degree in 50% of thepatients. The quantification measurement of 4D VENCwas failed in 6 patients in the right pulmonary artery andin 10 patients in the left pulmonary artery. The NFV of theright pulmonary artery in 4D and 2D was 37.4 + 14.6ml/stroke and 40.2 + 9.5 ml/stroke respectively (p =0.407) and RF of the right pulmonary artery in 4D and 2Dwas 0.277 + 0.22 and 0.225 + 0.19 respectively (p =0.008). In the left pulmonary artery, the NFV and RF in4D vs. 2D was 37.1 + 13.8 ml/stroke vs. 35.4 + 10.4ml/stroke (p = 0.678) and 0.32 + 0.21 vs. 0.32 + 0.19 (p =0.927) respectively. The paired t-test did not show thesignificant difference in 4D and 2D measurement exceptin RF of the right pulmonary artery The confidence inter-val of the Bland Altman plotting for the difference of NFVand RF in RPA was -22.1 to 27.7 and -0.18 to 0.07respectively and in the left pulmonary artery it was -23.6to 26.9 and -0.34 to 0.33. The difference in both NFV andRF is not clinically acceptable. CONCLUSION: The spiral pattern flow is seen in the75% of the patients with repaired TOF. The quantificationof the flow in 4D VENC is failed in 30% of the right pul-monary artery and 50% in the left pulmonary artery. Thedifference between 4D and 2D VENC measurement isnot clinically acceptable yet.

MC02 CV-05 15:20 4D flow MRI for evaluation of post-stenoticturbulent flow in phantom: comparison with 2DPC MRI and computational fluid dynamicsJihoon Kweon, Dong Hyun Yang, Guk bae Kim,Namkug Kim, Young-Hak Kim Asan Medical Center, Korea. donghyun.yang@gmail.com

PURPOSE: To assess the accuracy of 4D flow MRI inquantification and visualization of the post-stenotic flowusing flowmeter, 2D phase contrast (PC) MRI and com-putational fluid dynamics (CFD) simulation as referencestandards.MATERIALS AND METHODS: The flow phantom is con-structed to simulate turbulent stenotic flows using 75%stenosis model (22 mm internal diameter). A non-pulsatileflow of blood-equivalent fluid from a pump enters to inletof the stenosis phantom, and a flowmeter measures theflow rate (14.7 liter/minute). Image acquisition is per-formed on 3T scanner (MAGNETOM Skyra, Siemens)with varying spatial resolution, scanning direction andvelocity encoding range. In 2D PC MRI, transverse planeimages are captured at intervals of 2 cm for the compari-son. Flow information on PC-MR images is integratedwith 4D flow (Siemens) and CVI (Circle CardiovascularImaging Inc.) and the velocity information is extractedusing customized Matlab based software (MathWorks).Also, CFD analysis is conducted to investigate the flowcharacteristics in stenosis model.RESULTS: The time-averaged flow rate measured with4D flow MRI well follows the output value of the flowmeter

(-8.07 ± 8.18%) and when compared with the flow rateat the measurement positions of 2D PC MRI, 4D flow MRIexhibits comparable accuracy with smaller standard devi-ations (-1.88 ± 11.73% vs. -6.56 ± 8.12%, p = 0.428).With fine resolution (≤ 1.6×1.6×1.6 mm3), the time-averaged transverse velocity is in good agreement exceptthe most stenosed region, and flow characteristics of thepost-stenotic jet such as reverse flows (recirculation bub-ble) and flow eccentricity are well captured with the givenresolution.CONCLUSION: 4D flow MRI shows reliable performancefor the flow-rate measurement across the stenosis withcomparable accuracy to 2D PC MRI. With the spatial res-olution of 1.6 mm3 or finer, 4D flow MRI captures the prin-cipal characteristics of the post-stenotic flows.

MC 02 CV-06 15:30 Train velocity encoded phase contrast MRimaging for pulsatile velocity analysis withimproved temporal resolution and velocity-to-noise ratioPan-Ki Kim1, Whal Lee2, Hyeonjin Kim2, Eun-Ah Park2

1Seoul National University Hospital, 2Seoul NationalUniversity College of Medicine, Korea. whal.lee@gmail.com

PURPOSE: To propose new train PCMRI and to demon-strate the feasibility compared with conventional PCMRIin phantom and healthy volunteers.MATERIALS AND METHODS: The proposed trainPCMRI has composed with a velocity reference image atonly first cardiac phase and then applied with a train ofidentical velocity encoding gradients. The acquired phaseimages were slidingly subtracted between adjacent phaseimages by complex conjugate in each cardiac phase, andthen the subtracted phase images were reconstructed tothe velocity sensitized images by using cumulative sum-ming manner. Two health volunteers were measured forthrough-plane flow velocity of the abdominal aorta andfemoral arteries on 3T scanner. The imaging parameterwere: repetition time (TR) = 5.1-5.5 ms, echo time (TE) =3.1-3.6 ms, flip angle (FA) = 12 degree, number of aver-age = 1, slice thickness (TH) = 6 mm, 1 slice, readoutbandwidth 650 Hz/pixel, rectangular field of view 300 ×150 mm2, acquisition matrix 256 × 128.RESULTS AND DISCUSSION: In phantom study, themean velocities of conventional and train PCMRI haveshown a strong correlation (r = 1.00, p < 0.01), and thevelocity noise standard deviations in stationary gelatinregion are linearly dependent as function of VENC, andboth methods also strongly correlated (r = 0.994, p <0.01). The conventional PCMRI with 2 view per segment(VPS) and train PCMRI with 4 VPS have same temporalresolution to 20 ms, and the similarity of pulsatile flowcurves is statistically highly correlated to 0.998 (cl, 0.995-0.999) in Lin’s concordance-correlation coefficient (CCC).In in-vivo study, as shown in Figure 2, the CCC scored0.98 (p < 0.001) between the conventional PCMRI with 2VPS and train PCMRI with 4 VPS. The VNR of trainPCMRI (mean ± standard deviation: 19.62 ± 2.72) isabout twice as much as the conventional PCMRI (mean± standard deviation: 10.97 ± 4.82) according to VENC

164 KCR 2015

Cardiovascular S

ep 11, Fri

(p < 0.001). The mean difference between two methodsin Bland-Altman plot were 1.078 cm/s (95% CI for themeans = -10.03; 12.05) and 0.413 ml (95% CI for themeans = -2.81; 3.64) for peak velocities and flow, respec-tively.CONCLUSION: The proposed train PCMRI method wasconcurrently provided the twofold temporal resolution andthe increasable VNR using lower VENC compare withconventional PCMRI. And it can also use to reduce thetotal scan time by increasing VPS while maintaining thetemporal resolution of the conventional PCMRI.

MC 02 CV-07 15:40 Separated momentum of pulsatile stenotic flowfound at post-stenotic region and post-systolephaseGuk bae Kim1, Hojin Ha2, Dong Hyun Yang1, Sang Joon Lee2, Jihoon Kweon1, Young-Hak Kim1,Namkug Kim1

1Asan Medical Center, 2POSTECH Biotech Center,Koreanamkugkim@gmail.com

PURPOSE: To report the appearance of the separatedmomentum which can be a matter of misunderstanding indiagnosing the exact stenotic position.MATERIALS AND METHODS: Pulsatile Newtonian

flows in two stenotic phantoms having 50% and 75%reduction in area were investigated by 4D phase-contrast(PC) magnetic resonance imaging (MRI). Blood analogworking fluid was circulated via the stenotic phantom by apulsatile pump at a constant pulsating frequency of 1 Hz.The velocity profiles and fields in space and time werequantitatively analyzed in various directional points ofview.RESULTS: Pulsatile stenotic flow shows the separatedmomentum of the plug-like jet only at the specific stenoticdegree of 50% in our pulsatile waveform design. The spa-tial position where it appears is near the post-stenoticregion of 1-2 D (D: inner diameter) at the DIASTOLEphase in temporal. The observed separated momentumlooks similar to the jet-like flow of the stenotic region atthe systole phase, but it has a flow structure of the vortexring distinguishingly.CONCLUSION: The appearance of the separatedmomentum is complicatedly dependent on the stenoticdegree and the pulsatile waveform, which conditions criti-cally affect its spatial and temporal tendency as well. Inthe clinic, the separated momentum might cause a mis-understanding to diagnose the exact stenotic position. Toavoid this misdiagnosis, the local fastest region at theSYSTOLE phase has to be diagnosed as the stenoticposition and the other fastest regions found at the subse-quent phase have to be minutely ignored.

Cardiovascular 165C

ardiovascular Sep 11, Fri

166 KCR 2015

Cardiovascular S

ep 11, Fri

16:10-16:40 Grand Ballroom 104

MC 02 CV-08 Cardiac mapping: from rising star to key player

Chairperson(s) : Kyu-Ok Choe Konyang University Hospital, KoreaYeon Hyeon Choe Samsung Medical Center, Sungkyunkwan University School of Medicine, Korea

Myocardial mapping: where we are and what is new

Tim LeinerUniversity Medical Center Utrecht, Netherlands. timleiner@gmail.com

Cardiomyopathy (CM) is diseases of the heart muscle.The heart muscle may be enlarged, thick, or rigid and it issometimes replaced with scar tissue. The causes are un-known, but they can be acquired or inherited. We usuallyexclude coronary artery disease (CAD), hypertension,valvular or congenital heart diseases (VHD or CHD).Main types of cardiomyopathy are dilated CM (DCM), hy-pertrophic (HCM) and restrictive CM (RCM). Others in-clude arrhythmogenic right ventricle dysplasia (ARVD),Takotsubo CM, or myocarditis.

Normal myocardium consists of the heart muscle in75% and interstitial tissues in 25% including blood ves-sels, inflammatory cells, and collagen fiber which is usual-ly type I or III. Therefore, calculated extracellular volume(ECV) fraction on the mapping MRI is around 25% of themyocardial tissue. In this talk, I will review the mappingMRI in DCM, HCM, RCM and others including TakotsuboCM or myocarditis.

DCM is defined as a dilatation of left ventricle (LV) cavi-ty with systolic dysfunction. There is no history of CAD,hypertension, or VHD, etc. which is up to 35% of DCM.Non-familiar forms result from myocarditis, nutritional oralcoholic, ischemic or infiltrative myocardial diseases.Late Gd-enhancement (LGE) patterns are variable; noenhancement, right ventricle (RV) insertion site enhance-ment, or mid-wall enhancement of the LV. However,mean ECV was significantly higher in DCM patients (30to 31%) than those in control groups (25%) regardless ofthe LGE. Mean ECV was negatively correlated with theejection fraction (EF) either the LGE negative or positivegroups (1).

HCM in which the myocytes increase in size withoutany obvious cause results in thickening of the heart mus-cle and the alignment of muscle cell is disrupted. FamilialHCM is inherited as an autosomal dominant trait. Andgene mutation can be identified in 50-60% of the pa-tients. There is no difference in T1 values between nor-mal and remote areas. However, the reduced percent ofT1 value (RPTV) is significantly higher in peri-area as wellas halo or patchy LGE areas than in normal or remote ar-eas. Diagnostic accuracy of fibrosis detection improveswith a mapping study. Therefore, T1 mapping can be auseful method to evaluate the severity of myocardial fibro-sis in HCM patients (2). Post-contrast T1 times also shownegative correlation with LV filling pressure in HCM pa-tients. The myocardium in HCM shows shorter T1 valuesdue to diffuse myocardial fibrosis which manifests with di-

astolic dysfunction (3). In our study, we compared thepost-T1 values at the LGE area and remote area withoutenhancement. The remote area was divided into twogroups; thickened area above 15 mm and non-thickenedarea below 15 mm. ECV significantly increased in thenon-enhanced areas regardless of myocardial hypertro-phy. Therefore, we conclude that diffuse myocardial fibro-sis can be identified in HCM patients although the my-ocardium is not hypertrophied.

RCM is characterized as the rigid myocardial wallswhich are restricted from stretching and filling with bloodproperly. Primary type is endocardial fibroelastosis andsecondary one is infiltrative form from amyloidosis or sar-coidosis and interstitial form from post-radiation fibrosis.The classical form of amyloidosis shows diffuse subendo-cardial enhancement on the LGE MRI. Cardiac amyloido-sis shows amyloid deposition in the myocardium in whichinterstitial space expands and pre-T1 values markedly in-crease due to amyloid deposition (4). Pre-T1 values arealso useful to differentiate between ATTR amyloidosisand AL type amyloidosis (5). Calculated ECV markedlyincreases due to amyloid deposition in the interstitialspaces (6). In our study, mean ECV markedly increasedin amyloid patients. However, myocardial edema was notdefinite on T2 map. Sparrow P et al. (7) demonstratedthat there was no difference in mean T2 relation times be-tween the amyloid cases and normal controls (51.3 8.1vs. 52.1 3.1 msec).

Myocarditis is defined as an inflammation of the heartmuscle which is called as an inflammatory cardiomyopa-thy. The causes are from infection such as virus, bacteriaor fungus, immunologic reaction, and toxins or physicalagents. Cine MRI usually shows global hypokinesiawith/without pericardial effusion. T2WI shows diffuse my-ocardial edema in the LV wall. Mapping image showsmarkedly increased ECV fraction and T2 times due tomyocardial edema. Mean ECV and T1 or T2 values aremore sensitive in detection of myocarditis than the previ-ous LGE in the myocardium (8). Takotsubo CM is calledas apical ballooning CM or stress-induced cardiomyopa-thy which is non-ischemic cardiomyopathy. The heartmuscle is temporary weakening but recovered after prop-er treatment. Potential causes are from wraparound LAD,transient vasospasm, microvascular dysfunction or mid-ventricular obstruction due to LV outflow tract obstruction.There is no enhancement on LGE image. However,mean ECV increases and there is diffuse myocardial ede-

Cardiovascular 167C

ardiovascular Sep 11, Fri

16:40-17:10 Grand Ballroom 104

MC 02 CV-09 Cardiac mapping: from rising star to key player

Chairperson(s) : Kyu-Ok Choe Konyang University Hospital, KoreaYeon Hyeon Choe Samsung Medical Center, Sungkyunkwan University School of Medicine, Korea

The role of myocardial mapping in cardiomyopathy

Tae Hoon KimGangnam Severance Hospital, Korea. thkim1@yuhs.ac

ma.In conclusion, mean ECV fraction increases in DCM

patients although there is no enhancement in the my-ocardium. ECV fraction shows negative correlation withventricular EF. Post T1 times are shorter due to diffusemyocardial fibrosis and can be identified in non-hypertro-phied myocardium in HCM. Mean ECV expands, but my-ocardial edema is not definite in cardiac amyloidosis ascompared to myocarditis or Takotsubo cardiomyopathywhich has extensive myocardial edema.

References

1. Hong YJ, Park CH, Kim YJ, et al. Extracellular volumefraction in dilated cardiomyopathy patients without obvi-ous late gadolinium enhancement: comparison withhealthy control subjects. Int J Cardiovasc Imaging; 31Suppl 1:115-122.

2. Lu M, Zhao S, Yin G, et al. T1 mapping for detection ofleft ventricular myocardial fibrosis in hypertrophic car-diomyopathy: a preliminary study. Eur J Radiol; 82:e225-231.

3. Ellims AH, Iles LM, Ling LH, Hare JL, Kaye DM, TaylorAJ. Diffuse myocardial fibrosis in hypertrophic cardiomy-opathy can be identified by cardiovascular magnetic res-onance, and is associated with left ventricular diastolicdysfunction. J Cardiovasc Magn Reson; 14:76.

4. Karamitsos TD, Neubauer S. T1 mapping and amyloidcardiomyopathy: how much better can it get? Eur HeartJ; 36:203-205.

5. Fontana M, Banypersad SM, Treibel TA, et al. Native T1mapping in transthyretin amyloidosis. JACC CardiovascImaging; 7:157-165.

6. Robbers LF, Baars EN, Brouwer WP, et al. T1 mappingshows increased extracellular matrix size in the my-ocardium due to amyloid depositions. Circ CardiovascImaging; 5:423-426.

7. Sparrow P, Amirabadi A, Sussman MS, Paul N,Merchant N. Quantitative assessment of myocardial T2relaxation times in cardiac amyloidosis. J Magn ResonImaging 2009; 30:942-946.

8. Radunski UK, Lund GK, Stehning C, et al. CMR in pa-tients with severe myocarditis: diagnostic value of quanti-tative tissue markers including extracellular volumeimaging. JACC Cardiovasc Imaging; 7:667-675.

168 KCR 2015

Cardiovascular S

ep 11, Fri

Cardiac mapping: from rising star to key player

MC 02 CV-10 17:10 Predictive value of cardiovascular magneticresonance-derived myocardial strain for pooroutcome in patients with acute myocarditisJi Won Lee1, Ki Seok Choo2, Yeon Joo Jeong1,Geewon Lee1

1Pusan National University Hospital, 2Pusan NationalUniversity Yangsan Hospital, Korea. kschoo0618@naver.com

PURPOSE: To evaluate the value of cardiovascular mag-netic resonance (CMR)-derived myocardial strain for pre-dicting poor outcome in patients with acute myocarditis.MATERIALS AND METHODS: We retrospectivelyincluded 37 consecutive patients with acute myocarditiswho performed CMR (23 male, mean age 41.5 years).Myocardial strain parameters, left ventricular (LV) end-diastolic and end-systolic volumes, LV myocardial mass,LV ejection fraction (EF) and right ventricular EF werederived from CMR. Presence of late gadolinium enhance-ment (LGE) was also recorded. Primary outcome wasmajor adverse cardiovascular events (MACE). IncompleteLV functional recovery was used as secondary outcomein the group of patents who performed follow-up echocar-diography after 1 year.RESULTS: During an average follow-up of 41 ± 34months, 11 of 37 patients (29.7%) suffered MACE, includ-ing cardiac death (n = 2), heart transplantation (n = 1),cardiac pacemaker (n = 1), rehospitalization due to car-diac events (n = 4) or embolic stroke (n = 3). MultivariableCox proportional hazard regression analysis revealed thepresence of LGE (hazard ratio 42.88, p = 0.014) and radi-al strain obtained from the long axis views (ErrLax, hazardratio 0.77, p = 0.004) were significant predictors of MACE.Kaplan-Meier analysis showed worse outcome in patientswith LGE or ErrLax ≤ 9.48. Thirty one of 37 patients(83.7%) performed follow-up echocardiography.Multivariable backward stepwise regression analysisrevealed ErrLax was the sole significant predictor of LVfunctional recovery (hazard ratio 1.87, p = 0.042).Receiver operating characteristic curve of ErrLax wasused to find optimal cut-off values for prediction of incom-plete LV functional recovery, with corresponding areaunder the curve of 0.96. Cut-off value with the best com-bination of sensitivity and specificity for ErrLax was ≤14.86 (sensitivity 88.9%, specificity 95.5%).CONCLUSION: CMR-derived ErrLax can predict poor out-come such as MACE or incomplete LV functional recov-ery in the patients with acute myocarditis.CLINICAL SIGNIFICANCE: CMR-derived ErrLax can pre-dict poor outcome in the patients with acute myocarditis.Furthermore, presence of scar indicated by LGE is alsothe good independent predictor of MACE. This resultssupport the necessity for future large longitudinal follow-up studies to establish LGE and CMR-derived myocardialstrain as an independent predictor of MACE in acutemyocarditis.

MC 02 CV-11 17:20 Association of cardiovascular disease riskfactors with left ventricular mass, biventricularfunction, and presence of silent myocardialinfarction determined by cardiac MR inasymptomatic populationEliel Nham, Sung Mok Kim, Yeon Hyeon ChoeSamsung Medical Center, Sungkyunkwan UniversitySchool of Medicine, Korea. yhchoe@skku.edu

PURPOSE: To evaluate the relation of cardiac mass andfunction determined by cardiac magnetic resonanceimaging (cardiac MRI) with cardiovascular disease (CVD)risk factors and to investigate possible risk factors forsilent myocardial infarction (SMI).MATERIALS AND METHODS: This cross-sectionalstudy included 647 asymptomatic subjects (485 males,aged 54.8 ± 6.7 years; 162 females, aged 55.2 ± 7.6years) who underwent cardiac MRI for health checkup.Cardiac MR imaging was performed using a 1.5T scan-ner with a 32-channel cardiac coil. Both left ventricle (LV)and right ventricle (RV) were examined. Association ofventricular parameters and CVD risk factors was exam-ined by multivariable regression and analysis of variance.Logistic regression was used to model the association ofrisk factors with SMI.RESULTS: The prevalence of metabolic syndrome (n =116) and silent myocardial infarction (SMI) (n = 12) in ourstudy population was 17.9% and 1.9%, respectively.Mean LVEF and RVEF were higher in females comparedwith males (68.5 ± 7.2% vs. 66.6 ± 6.2% for LVEF, p =0.004; and 60.4 ± 8.1% vs. 56.9 ± 7.5% for RVEF, p <0.001). LV function parameters were seen associatedwith age in males (EF, p = 0.001; EDV, p = 0.006; M/V, p= 0.014) and females (EF, p = 0.004; EDV, p = 0.017;mass, 0.038; M/V, p = 0.026; EDVI, p = 0.036; MI, p =0.025) and RV function parameters were seen associatedwith aging in males (EDV, p = 0.001; EDVI, p = 0.027)and females (EDV, p = 0.002; EDVI, p = 0.007). LV masswas positively related to BMI (standardized coefficient [β]= 0.322, p < 0.001) systolic blood pressure (β= 0.251, p <0.001), more recent smoking (β= 0.129, p < 0.001). LVmass-to-volume ratio: BMI (β= 0.153, p < 0.001), systolic(β= 0.165, p = 0.001) and diastolic blood pressure (β=0.147, p = 0.002), TG (β= 0.197, p = 0.006), and CRP (β= 0.130, p < 0.001) were positively related and eGFR (β=-0.076, p = 0.025) was negatively related to LV mass-to-volume ratio. Metabolic syndrome was not associatedwith any of ventricular parameters when corrected forconfounders. Diabetes was the only risk factor with signif-icant odds ratio (OR) for SMI (estimated OR = 5.282, p =0.008).CONCLUSION: Known CVD risk factors are associatedwith altered ventricular mass, geometry and function inasymptomatic subjects. Diabetes may contribute to devel-opment of SMI.

Cardiovascular 169C

ardiovascular Sep 11, Fri

MC 02 CV-12 17:30 Quantification of left ventricular trabeculaeusing cardiac MR imaging for the diagnosis ofleft ventricular noncompaction: evaluation oftrabecular volume and new semiquantitativecriteriaYeonu Choi, Sung Mok Kim, Yeon Hyeon ChoeSamsung Medical Center, Korea. yhchoe@skku.edu

PURPOSE: Left ventricular noncompaction (LVNC) is anunclassified cardiomyopathy and there is no consensuson the diagnosis of LVNC. The aims of this study were toestablish quantitative methods to diagnose isolated LVNC(INC) using cardiac magnetic resonance (CMR) imagingand to suggest novel semiquantitative methods to diag-nose isolated LVNC.MATERIALS AND METHODS: This retrospective studyincluded 145 subjects with moderate to severe trabecula-tion of LV myocardium (24 patients with isolated LVNC,33 patients with non-isolated LVNC, 30 patients with dilat-ed cardiomyopathy [DCM] with noncompaction [DCMNC],27 patients with DCM and 31 healthy control subjectswith mild trabeculation). LVNC patients had to fulfillPetersen’s CMR criteria. Left ventricular (LV) ejectionfraction, global LV volume, trabeculated LV volume, andnumber of segments with late gadolinium enhancementwere measured. And most prominent noncompacted(NC), compacted (C), normal mid-septum, normal mid-lat-eral wall and apical trabeculation thickness on the end-diastolic frames of long-axis slices were also measured.RESULTS: In the patients with isolated LVNC, the per-centage of trabeculated LV volume (42.6 ± 13.8%) was1.4 times higher than in DCM (30.3 ± 14.3%, p <0.0001), and 1.7 times higher than in controls (24.8 ±7.1%, p < 0.0001). However, there was no significant dif-ference between INC and DCMNC (47.1 ± 17.3%, p =0.210). And a value of percentage of trabeculated LV vol-ume above 32% was predictive of isolated LVNC with aspecificity of 90.3% (CI, 74.2-98.0%) and sensitivity of79.2% (CI, 57.8-92.9%). A value of NC/septum over 1.1was considered predictive for isolated LVNC with a speci-ficity of 80.6% (CI, 62.5-92.5%) and sensitivity of 95.8%(CI, 78.9-99.9%). And a value of apex/C above 3.1 wasconsidered predictive of isolated LVNC with a specificityof 93.5% (CI, 78.6-99.2%) and sensitivity of 87.5% (CI,67.6-97.3%).CONCLUSION: As a quantitative approach, a trabeculat-ed LV volume above 32% of the average LV volume isdiagnostic for LVNC with high sensitivity and specificity.And as a semi-quantitative approach, apex/C andNC/septum ratio could be useful for supplemental diag-nostic criteria.

MC 02 CV-13 17:40 Extracellular volume fraction in hypertrophiedcardiomyopathy patients without obvious lategadolinium enhancement Chul Hwan Park, Eui-Young Choi, Tae Hoon Kim Gangnam Severance Hospital, Korea. thkim1@yuhs.ac

PURPOSE: To evaluate whether the extracellular volumefraction (ECV) measured using cardiac magnetic reso-nance (CMR) imaging can detect myocardial tissuechanges in hypertrophied cardiomyopathy (HCM) withoutlate gadolinium enhancement (LGE).MATERIALS AND METHODS: Nineteen HCM patientsand 7 healthy volunteers underwent pre- and post-T1mapping using a modified Look-Locker Inversion recov-ery sequence, LGE, and cine MRI on a 1.5T CMR sys-tem. LGE-MR findings were used to divide HCM patientsinto two groups: Group A had no apparent LGE, andGroup B had LGE apparent in at least one segment. TheECV of the left ventricle (LV) myocardium (16 segments)was calculated in the short-axis view as follows: ECV =[(ΔR1 of myocardium/ΔR1 of LV blood pool)]×(1-hematocrit), where R1=1/T1, ΔR1 = post-contrast R1-pre-contrast R1. The mean myocardial ECV in LGE (-)segments in Group A + B was compared to that of con-trols. The mean myocardial ECV in Group A was com-pared to that of LGE (-) segments in Group B.RESULTS: Among the 19 HCM patients, 9 were in GroupA, and 10 were in Group B. The mean ECV of LGE (-)segments in HCM patent (31.4±3.4%) was significantlyhigher (P<0.001) than that of the control group (24.8±2.7%).CONCLUSION: The ECV measured by MRI could be auseful parameter in evaluating diffuse myocardialchanges in HCM patients.

MC 02 CV-14 17:50 Normal range of myocardial T1 values withsegmental evaluation and relation with clinicalfactorsMoon Young Kim, Haejin Kim, Sung Mok Kim,Sang-Chol Lee, Soo Jin Cho, Yeon Hyeon Choe Samsung Medical Center, Korea. yh.choe@samsung.com

PURPOSE: To evaluate whether there is variation in pre-contrast and postcontrast myocardial T1 time (prT1 andpoT1, respectively) and extracelluar volume fraction(ECVF) according to left ventricular (LV) segments and tosearch for any correlation between them and known car-diovascular risk factors.MATERIALS AND METHODS: This study included 198asymptomatic subjects (M:F = 180:18; age, 54.4 ± 6.12years) who underwent cardiac MR imaging. PrecontrastT1 mapping and postcontrast T1 mapping 15 minutesafter 0.2 mmol gadobutrol injection were performed usingshortened modified look-locker inversion recovery[ShMOLLI] sequence at 1.5T (Magnetom Avanto,Siemens). Short-axial cine MR imaging was performedwith SSFP technique. T1 values and ECVFs were calcu-lated in 16 AHA myocardial segments. Those valueswere compared among LV segments and correlated with

170 KCR 2015

Cardiovascular S

ep 11, Fri

presence of hypertension (n = 52), diabetes mellitus (DM,n = 15), or both (n = 17). ECVF was also correlated withLV mass.RESULTS: The overall prT1 and poT1 values and ECVFwere 1006 ± 291.5 ms, 454.2 ± 38.5 ms, and 0.24 ±0.04, respectively. There was significant differencebetween apical segments and mid-basal segments inpoT1 value and ECVF (p < 0.03) and between mid-septalsegments and mid-lateral segments in T1 values andECVF (p < 0.04). ECVF showed reverse correlation withLV mass (p = 0.002). There was significantly lower poT1value (449 ± 35.6 ms) and higher ECVF (0.24 ± 0.04) insubjects with hypertension compared with those (459 ±43.3 ms and 0.23 ± 0.02) of subjects without hyperten-sion (p < 0.05). Subjects with DM showed no difference inall T1 values from subjects without DM or hypertension,except poT1 values in mid-septal segments (447 ± 23.6ms vs. 459 ± 45.6 ms, p = 0.02). Subjects with both riskfactors showed no difference in all T1 values from sub-jects without DM or hypertension, except prT1 valuebetween apical septal and lateral segments (1007 ± 126ms vs. 999 ± 156 ms, p = 0.03).CONCLUSION: The septal wall showed higher prT1value and ECVF but lower poT1 value than the lateralwall of mid- and basal levels. PoT1 value and ECVF aresignificantly affected by hypertension and LV mass.

Cardiovascular08:00-09:30 Grand Ballroom 101

Cardiovascular Imaging

Chairperson(s)Kyung-Sup Song Pohang St. Mary's Hospital, KoreaGong Yong Jin Chonbuk National University Hospital,

Korea

SS 14 CV-01 08:00 Comparison of aortic valve planimetry by 3-dimensional MR imaging and conventional cineMR imaging to assess the aortic valve stenosisHae Jin Kim, Yeon Hyeon Choe, Sung Mok Kim,Moon Young Kim, Sung-Ji Park Samsung Medical Center, Korea. yhchoe@skku.edu

PURPOSE: We intended to evaluate the novel applica-tion of high-resolution 3-dimensional MR image acquisi-tion with single-breath-hold SSFP sequence to calculatethe aortic valve area (AVA).MATERIALS AND METHODS: In 88 consecutivepatients (66.9 ± 9.59 years, 63% men) with varyingdegrees of aortic valve stenosis, high-resolution 3D SSFPimages (3D planimetry; 2.0 mm slice thickness, 20 con-tiguous slices; image matrix, 256 209) were acquiredwith single breath-hold during mid systole and mid dias-tole. SSFP cine MR imaging (2D planimetry) and velocity-encoded cine MRI (slice thickness, 4.5 mm) in three lev-els of aortic valve were also performed. AVA area wasmeasured by two experienced observers using commer-cial software (iNtuition, TeraRecon). MR imaging mea-

surements and image quality were compared withtransthoracic echocardiographic measurements of effec-tive aortic orifices (EAO) using the continuity equation (1= severe blurring of images, 2 = moderate blurring ofvalve contours; 3 = mild blurring of valve contours, 4 =excellent and no artifact). Sensitivity for accurate mea-surement and receiver operating characteristic (ROC)curve were calculated. Intra- and interobserver agree-ments were determined by using intraclass correlationcoefficient (ICC).RESULTS: Mean AVA derived by 3D planimetry, 2Dplanimetry, and echocardiography were 0.77 ± 1.04 cm2,0.72 ± 1.16 cm2, and 0.75 ± 0.32 cm2, respectively. TheICC value of 3D planimetry was higher than 2D planime-try (0.799 [CI, 0.691-0.869] vs. 0.743 [CI, 0.605-0.832])with echocardiographic EAO as the standard of refer-ence. The grade of image quality of 3D planimetry wassuperior to 2D planimetry (3.65 ± 0.65 vs. 3.17 ± 0.65).The correlation coefficients of maximum peak velocity onvelocity-encoded cine MR imaging with 3D planimetryand that with 2D planimetry were 0.42 (p < 0.05) and 0.35(p < 0.05). Intra- and interobserver agreements for 3Dplanimetry were excellent [ICC = 0.949 (CI, 881-979) and0.846 (CI, 0.636-0.935), respectively; both, p = 0.000).CONCLUSION: Novel application of high-resolution 3DSSFP breath-hold MR imaging enables planimetry ofAVA in patients with valvular aortic stenosis with betterimage quality than 2D planimetry with conventional cineMR imaging.

SS 14 CV-02 08:10 Pulmonary artery indices (measured on CTangiogram) in correlation with preoperativesurgical planning and postoperative deaths inTOF patients - a prospective studySneha Wakode, T. Mandpal, L.T.Kishore, Johann ChristopherNational Board of Examinations, Delhi, India. wakodesneha@gmail.com

INTRODUCTION: Pulmonary artery size is a key deter-minant of type of surgery and post operative mortality inTOF patients. Pulmonary artery index is a standardizedmethod of calculating the pulmonary artery size in relationto various parameters, such as body surface area usingvarious angiographic techniques such as cine-angiogra-phy, CTA and MRA.PURPOSE: 1. Measuring PAI (pulmonary artery indices)in TOF (tetralogy of Fallot) patients referred for CTangiogram. 2. Correlate PAI values with the type ofsurgery (operator specific) planned for patients and withpostoperative deaths.MATERIALS AND METHODS: CT Angiograms of 32patients were done on 64 SLICE DUAL SOURCE CTSCANNER (Somatom definition flash) and Pulmonaryartery indices (Nakata index and McGoon index) werecalculated in all patients. Follow-up was obtained in termsof type of surgery (corrective or palliative) and one monthpost operative mortality.RESULTS: In our study, PA size expressed in terms ofMcGoon index showed definite and significant correlationwith both, the type of surgery and early post deaths.Meaning, higher the McGoon index, more complex and

Cardiovascular 171C

ardiovascular Sep 11, Fri

single stage repair can be attempted in these patients.Nakata index showed significant correlation only withearly post operative deaths.

CONCLUSION: 1. Pulmonary artery indices are stan-dard, simple, practical and reproducible method of calcu-lating the pulmonary artery size on CTA.2. PAI are useful guide in planning the surgical manage-ment. Higher the pulmonary artery index, more complexand single staged cardiac can be attempted. Low pul-monary artery indeces are associated with poor surgicaloutcome and high post op mortality.3. McGoons ratio was found to be the most reliablemethod to calculate pulmonary artery indices. 4. Low Nakata index strongly correlated with death.

SS 14 CV-03 08:20 Delineation of right atrial septum using multi-phase MR angiography to prepareradiofrequency catheter ablation of atrialfibrillation: merit of additionally delayed phasescanJun Seoung Kim, Sung Ho Hwang, Yu-Whan Oh,Soo-Youn Ham Korea University Anam Hospital, Korea. gimjunsung@naver.com

PURPOSE: Additional radiofrequency catheter ablation(RFCA) of right side atrial septum (RAS) has beenapplied to improve the clinical outcome of RFCA for atrialfibrillation (AF). We aimed to assess the image quality ofmulti-phase magnetic resonance angiography (MP-MRA)to guide the RFCA of RAS.MATERIALS AND METHODS: 50 patients (38 men;mean age, 59.6 ± 9.3 years) underwent pre-proceduralMP-MRA with infusion of contrast agent to prepare theRFCA for AF. Of MP-MRA data sets, the early-phaseMRA (EP-MRA) with the strongest left atrial enhancementand the delayed phase MRA (DP-MRA) with re-enhance-ment of superior vena cava were selected. The subjectivequalitative analysis in the RAS was performed using a 3-point scale (1 = poor; 3 = excellent contour). In addition,the relative contrast (RC) ratio in the right atrium (RA) andthe longest thickness (LT) of RAS were assessed. RESULTS: The mean image quality score in the DP-MRA revealed significantly better image quality of RAS(3.44 ± 0.81 vs. 2.22 ± 0.93, p < 0.01) than that in theEP-MRA. The mean RC ratio was significantly greater inthe DP-MRA (0.72 ± 0.12 vs. 0.43 ± 0.02, p < 0.01)than in the EP-MRA. Furthermore, intra- and interobserv-er measurements of LT-RAS revealed closer correlation

and closer 95% limits of agreement using DP-MRA.CONCLUSION: The MP-MRA with infusion of contrastagent can provide a guidance of RFAC for AF. Especially,the DP-MRA with the reentry of contrast agent into theRA can allow better image quality in the delineation ofRAS.

SS 14 CV-04 08:30 Strategies for reducing radiation dose in noncontrast CT according to patients body sizeWoosup Cho1, Whal Lee1, Eun-Ah Prak1, Jin Young Yoo2

1Seoul National University College of Medicine, 2SeoulNational University Hospital, Korea. whal.lee@gmail.com

PURPOSE: The contrast noise ratio (CNR) is one of themost important parameter in CT imaging. The CNR isdetermined by the contrast and noise. Contrast is deter-mined by type and concentration of material and kVp.Noise is determined by mAs and patient body size. Thehigher mA and higher kVp will give us higher CNR butalso higher radiation dose. The purpose of this study is toinvestigate the optimal kVp for best contrast noise ratio ofiodine contrast material, bone, protein and fat in variousbody shapes in setting of same radiation dose.MATERIALS AND METHODS: A phantom with tubes ofdifferent materials including iodine contrast material,bone, protein and fat at its center and concentric cylindri-cal plastic chambers of three layers surrounding the cen-ter at its periphery was constructed for this study. Thethree layers of on centric cylinders were filled with oil orremained empty to simulate different body shape such asthin adult, normal adult and obese adult. Repeated CTscanning was performed with dual-source CT. Scan para-meters were 80, 100, 120 and 140 kVp with 11 differentmAs at each tube voltage and 0.6 mm slice thickness.Tube attenuation, image noise, contrast to noise ratio andradiation dose were obtained.RESULTS: The CNR per radiation dose of iodine con-trast material and bone are highest at 80 kVP in all threebody shape. The CNR per radiation dose of protein ishighest at 140 kVP in obese body shape, at 120 kVp atnormal body shape, and at 100 kVp at thin body shape.The CNR per radiation dose of fat is highest at 80 kVPand 100 kVp in all three body shape.CONCLUSION: In consideration of radiation dose, the 80kVp is the best for CNR of iodine contrast material, boneand fat. The best dose-CNR efficacy of the proteindepends on the body shape. In the non-contrast soft tis-sue CT, 140 kVp is the best for the obese patient, 120kVp is the best for the normal patient, 100 kVp is the bestfor the lean patient.

172 KCR 2015

Cardiovascular S

ep 11, Fri

SS 14 CV-05 08:40 Monitoring of mitral loop cerclage as a variantform of mitral cerclage annuloplasty by using CTin animal studyYeo-Jin Jeong1, Ki seok Choo1, June Hong Kim1, Min Gu Chon1, Ji Won Lee2

1Pusan National University Yangsan Hospital, 2PusanNational University Hospital, Korea. kschoo0618@naver.com

PURPOSE: To compare pre with post mitral loop cer-clage (MLC) in measuring dimension of the mitral valveand left ventricular (LV) volume by using CT in animalstudy.MATERIALS AND METHODS: Nine healthy farm swineunderwent MLC and CT (128-slice dual source CT,Definition Flash, Siemens Medical Solution, Forchheim,Germany) was performed to these swines before andafter MLC. CT protocol was as follows: 140 kVp, 320effective mAs using a continuous helical scan, a rotationtime of 280 msec, slice collimation of 2 × 128 × 0.6 mm,and 120 mL of iopromide (Ultravist 370, Bayer, Germany)administered at a rate of 5 ml/sec followed by 20 mLsaline at the same rate. LV volume and dimension of themitral valve were measured by one radiologist in all swinebefore and after MLC, respectively.RESULTS: The MLC procedural success rate was 100%.CT showed significant difference between pre MLC andpost MLC of septal lateral dimension of the mitral annulus(24.58 ± 2.16. vs. 21.26 ± 1.43 mm, p = 0.04) and LVvolume in diastole (75.9 ± 3.9 vs. 70.6 ± 5.0 ml, p =0.04).CONCLUSION: CT can be good monitoring tool in MLCas a novel approach for catheter-based mitral valverepair.

SS 14 CV-06 08:50 Thoracic aorta and arch vessel anomalies inTurner syndrome on chest CT: compared withnormal control groupWon Jin Choi1, Eun-Ju Kang1, Ki-Nam Lee1, Jongmin Lee2

1Dong-A University Hospital, 2Kyungpook NationalUniversity Hospital, Korea. medcarrot@naver.com

PURPOSE: To investigate the spectrum and frequency ofthoracic aorta and arch vessel anomalies using contrastenhanced chest CT in patients with Turner syndrome(TS) compared with normal control females.MATERIALS AND METHODS: A total number of 20 con-secutive TS patients (15.8 ± 6.61 years) who underwentcontrast enhanced chest CT from May 2012 to May 2015for clinical evaluation were retrospectively enrolled.Twenty females (15.75 ± 4.64 years) who underwentcontrast enhanced chest CT from November 2014 to May2015 with various clinical needs were retrospectivelyenrolled as control group. Elongation of the transverseaortic arch (ETA) was defined by the presence of both (1)posterior origin of the left subclavian artery (LSCA)behind the trachea on axial images and (2) inward inden-tation or convex kinking of the inferior aortic contour alongthe lesser curvature. Average cross sectional diameters

and areas of right brachiocephalic artery (RBCA), LSCAwere measured within 5 mm distal to their origin oncurved multiplanar images using commercial software.The ratio of the area of LSCA by RBCA (‘LSCA ratio’)was calculated. The other findings including coarctation ofthe aorta (CoA), bovine arch, aberrant right subclavianartery, and anomalous origin of left vertebral artery (VA)were also recorded. We compared the results of eachdata of TS with that of control group.RESULTS: There were no statistically significant differ-ence between TS and control group in age, weight, BSA(p > 0.05) but significant difference in height (140.88 ±10.22 cm vs. 151.15 ± 12.97 cm, p = 0.001). ETA wasonly detected in 11 patients of TS (55%) and showed sta-tistically significant difference with control group (p <0.001). The cross sectional area of both RBCA and LSCAwere relatively larger in TS patients than control group,especially LSCA (91.67 ± 45.90 mm2 vs. 50.72 ± 15.30mm2, p = 0.001). LSCA ratio was significantly higher in TSpatients than that of control (0.86 ± 0.38 vs. 0.57 ±0.16, p = 0.004). CoA (n = 2), aberrant RSCA (n = 1)were only seen in TS patients and anomalous origin ofleft VA were more common in TS patients (5 in TS vs. 1in control). Bovine aortic arches were equally seen in bothTS (n = 2) and control (n = 2).CONCLUSION: The ETA was characteristic finding in TSpatient with 55% incidence, and the arch vessel size wasrelatively large on TS patients, especially LSCA. Theother vascular anomaly or variant such as CoA, aberrantRSCA were more common in TS than control.

SS 14 CV-07 09:00 Comparison of objective and subjective imagequality between 80 kVp and 120 kVp CTvenography for the lower extremity with model-based iterative reconstruction at same noiseindex: preliminary reportHyun-Jung Baek1, Ji Won Lee2, Ki Seok Choo1

1Pusan National University Yangsan Hospital, 2PusanNational University Hospital, Korea. kschoo0618@naver.com

PURPOSE: To compare objective and subjective imagequality of 80 kVp CT venography (CTV) with 120 kVpCTV with model-based iterative reconstruction (MBIR) atsame noise index.MATERIALS AND METHODS: This retrospective studywas approved by our Institutional Review Board. Total 66patients (mean age, 56.1 ± 18.1) who were underwentCTV using 80 kVp (33 patients, group A) or 120 kVp (33patients, group B) at same noise index (21) were enrolledand these images were reconstructed with MBIR for eval-uation of deep vein thrombosis (DVT) or varicose vein.Objective image quality (vascular enhancement, noise,contrast to noise ratio [CNR]) were measured in the inferi-or vena cava (IVC), femoral vein (FV) and popliteal vein,respectively by independent two radiologists. In addition,subjective image parameters (image quality, image noise,confidence to find DVT) were assessed using a 5 pointscale system by same two radiologists independently.Data were analyzed using paired t-test and Mann-Whitney U test and effective dose was estimated usingthe dose-length product (DLP).

Cardiovascular 173C

ardiovascular Sep 11, Fri

RESULTS: All images were acceptable for diagnosis.The mean vascular enhancement and CNR of Group Awas significantly higher than those of Group B in the IVC,FV and popliteal vein (p < 0.05). However, images inGroup A had significantly higher objective image noisethan group B (p < 0.05). In addition, the subjective imagequality and confidence of find DVT of Group A was signifi-cantly higher than those of Group B (p < 0.01). The meanDLP was not statistically different (Group A vs. Group B,354.31 ± 51.20 mGy cm vs. 348.68 ± 56. 40).CONCLUSION: CTV using 80 kVp with MBIR providedhigher vascular enhancement and CNR as well as bettersubjective image quality than CTV using 120 kVp withMBIR at same noise index level.

SS 14 CV-08 09:10 Ventricular noncompaction in adult: imaging andclinical findings from 63 patientsSejin Cho, Donghyun Yang, Jun-won Kang, Tae-Hwan Lim Asan Medical Center, Korea. donghyun.yang@gmail.com

PURPOSE: To describe imaging and clinical findings ofleft ventricular noncompaction (LVNC) in adult.MATERIALS AND METHODS: From 2000 to 2014, 63patients were diagnosed to LVNC by echocardiography,computed tomography (CT), and magnetic resonanceimaging (MRI) in a single institution. Baseline characteris-tics, clinical manifestation, combined cardiac or systemicanomalies, and imaging findings were reviewed and com-pared between isolated group (LVNC without cardiacanomaly) and combined group (LVNC with cardiac anom-aly). Institutional Review Board approved the study andinformed consent was waived.RESULTS: Among 63 patients with LVNC, 32 (51%)patients did not have combined cardiac anomaly, isolatedgroup of LVNC. Mean age at initial diagnosis of isolatedLVNC group is higher than the combined group (54.2years vs. 40.2 years; p < 0.001). The combined grouppresented symptoms more frequently at initial diagnosisthan isolated group (93.6% vs. 75%). Heart failure symp-toms were most common (60.3% in all patients).Thromboembolic events were shown in 20 (31.7%)patents in all patients, which were much higher in com-bined group (66.7%) than isolated group (33.3%). Themost common combined cardiac abnormality was dilatedcardiomyopathy (n = 15, 24%) followed by other congeni-tal heart disease (n = 12, 19%) (Imaging findings of CTand MRI are under analysis. Final presentation willinclude the imaging findings of LVNC in CT and MRI).CONCLUSION: The isolated and combined groups ofLVNC show differences in age at diagnosis and clinicalmanifestations. The most common combined cardiacabnormality was dilated cardiomyopathy.CLINICAL RELEVANCE: We provided clinical character-istics of LVNC from 14-year imaging database of a singleinstitution. Clinical and imaging findings of LVNC present-ed in the study may be helpful to understand LVNC todetermine management plan of the disease.

SS 14 CV-09 09:20 Assessment of mitral annuloplasty ring bycardiac CT: correlation with echocardiographicparameters and comparison between twodifferent ring typesYoung Joo Suh, Young Jin Kim, Yoo Jin Hong, Hye-Jeong Lee, Jin Hur, Byoung Wook Choi Severance Hospital, Korea. dryj@yuhs.ac

PURPOSE: To investigate CT appearance after mitralring annuloplasty, especially to compare CT findingsbetween patients with normal pressure gradient (PG) andpatients with functional mitral stenosis (MS) and betweentwo commonly used types of annuloplasty ring.MATERIALS AND METHODS: A total of 45 cardiac CTscans in patients who underwent mitral ring annuloplasty(Carpentier-Edwards [CE] ring [n = 27], Duran ring [n =18]) were retrospectively reviewed. On CT scan, pres-ence of significant pannus around the annuloplasty ring,presence of leaflet thickening, and the maximal mitralopening area were analyzed. CT findings were comparedbetween patients with normal PG and patients with func-tional MS (mean-diastolic PG ≥ 5 mmHg). Incidence offunctional MS and CT findings were compared betweenthe two ring types.RESULTS: Significant pannus was present in 10 casesand leaflet thickening in 31 cases, and the maximal open-ing area was 2.34 ± 0.717 cm2. The valve opening areaon CT was positively correlated with mitral valve area(MVA) on TTE, and negatively correlated with mean-dias-tolic PG. The mean-diastolic PG was significantly elevat-ed with increasing pannus severity. Patients with function-al MS had more significant pannus formation thanpatients with normal range PG. Patients with the Duranring had a higher mean-diastolic PG, smaller MVA andhigher incidence of functional MS than patients with theCE ring (p < 0.05). The proportion of pannus and signifi-cant pannus was significantly higher in patients with theDuran ring (p < 0.05).

Fig. 1. Significant pannus formation causing mitral steno-sis in a 52-year-old male patient with the Duran annulo-plasty ring.

174 KCR 2015

Cardiovascular S

ep 11, Fri

CONCLUSION: Significant pannus around the annulo-plasty ring on CT may cause functional MS after mitralring annuloplasty and this may occur more frequently forthe Duran ring.

Cardiovascular09:50-11:20 Grand Ballroom 101

Cardiac CT

Chairperson(s)Joon-Won KANG University of Ulsan College of

Medicine, Asan Medical Center,Korea

Jeong A Kim Inje University Ilsan Paik Hospital, Korea

SS 17 CV-01 09:50 Coronary artery disease in patients withhypertrophic cardiomyopathy- coronary CT angiography studyYoon Joo Shin, Yeo-Koon Kim, Sang Il Choi, Eun Ju Chun Seoul National University Bundang Hospital, Korea. drejchun@hanmail.net

PURPOSE: To evaluate the prevalence and clinicalimpact of coexistent coronary artery disease (CAD) inpatients with hypertrophic cardiomyopathy (HCM) usingcoronary CT angiography (CCTA).MATERIALS AND METHODS: A total 248 patients withHCM diagnosed by clinical findings, electrocardiography,and echocardiography were retrospectively enrolled. Weevaluated the prevalence of the obstructive CAD (> 50%luminal reduction) and type of the plaque (calcified, non-calcified, high-risk plaque) according to 16-segmentmodel. High-risk plaque (HP) was defined as a plaqueshowing density with < 30 HU, positive remodeling withnapkin ring sign and spotty calcification. Clinical risk fac-tors were also evaluated from all patients. The primaryendpoint was defined as cardiac death, non-fatal myocar-dial infarction, unstable angina requiring hospitalization,revascularization after 90 days from index CCTA, orimplantable cardioverter defibrillator insertion.RESULTS: In patients with HCM, the prevalence of coex-istent obstructive and non-obstructive CAD was 16.5%and 42.7%, respectively. During the median of 37-monthsobservation period (range, 3-108 months), total cardiacevents were occurred in 11.7% of patients with HCM. Theprevalence of cardiac events were significantly increasedin HCM patients with obstructive CAD (39.0%) as com-pared to non-obstructive CAD (6.6%) and no CAD (0%)(p < 0.05). Using univariate and multivariate Cox regres-sion analysis model, the presence of obstructive CAD isindependent factor for cardiac events in HCM (Hazardratio 7.3, 95% confidence interval 2.5-21.5, p < 0.001).CONCLUSION: The prevalence of obstructive CAD inpatients with HCM was approximately one-fifth of theHCM population on this study. The presence of obstruc-tive CAD itself is associated with higher risk of majoradverse cardiac events in HCM patients compared withthose with no CAD or non-obstructive CAD. CCTA can be

helpful to provide information regarding myocardial hyper-trophy and CAD at the same time.

SS 17 CV-02 10:00Comparison between the quantitativeparameters of myocardial perfusion CT andinvasive fractional flow reserveHyun Jung Koo, Dong Hyun Yang, Joon-Won Kang,Soo-Jin Kang, Young-Hak Kim, Tae-Hwan Lim Asan Medical Center, Korea. donghyun.yang@gmail.com

PURPOSE: We attempted to compare the quantitativeparameters which are represent the presence, location,and extent of myocardial ischemia in coronary artery dis-ease on myocardial perfusion CT using invasive fractionalflow reserve (FFR).MATERIALS AND METHODS: We enrolled 68 patients(mean age, 62.8 years; 56 males) undergoing myocardialperfusion CT who underwent cardiac catheterization withinvasive FFR from a prospective CT perfusion (CTP) reg-istry. CTP was performed using second generation dualsource CT with adenosine-induced, stress-first, and staticCTP protocol. Myocardial attenuation and area in each 16myocardial segment and perfusion defect were evaluatedin short-axis stress-CTP image covering whole left ventri-cle (LV). For quantitative parameters of CTP, the perfu-sion defect-to-normal (D/N) ratio of myocardial attenua-tion and the ischemic burden (%) defined as the ratio ofperfusion defect volume to LV myocardial volume wereassessed in both systolic and diastolic images. An associ-ation between D/N ratio or ischemic burden and invasiveFFR were analyzed.RESULTS: Among the 119 coronary arteries with mea-sured invasive FFR, 39 (33.1%) arteries showed hemody-namically significant stenosis (FFR ≤ 0.80). The D/Nratio of myocardial attenuation showed significant correla-tion with FFR (r = 0.58, p < 0.001). However, ischemicburden did not correlated with FFR (r = -0.18, p = 0.37).The mean ischemic burdens of left anterior descendingartery, right coronary artery, and left circumflex arterystenosis were 29.7 ± 15.5%, 23.7 ± 5.3%, 17.2 ±9.8%, respectively. The D/N ratio in systolic phaseshowed better correlation with invasive FFR than that indiastolic phase (r = 0.58 vs. 0.44).

CONCLUSION: Myocardial attenuation ratio of perfusiondefect-to-normal area in CTP showed significant correla-tion with invasive FFR. However, ischemic burden in CTP

Cardiovascular 175C

ardiovascular Sep 11, Fri

did not correlated with FFR. The quantitative parametersof CTP in the study may be helpful to predict invasiveFFR using automatic analysis software of CTP.

SS 17 CV-03 10:10Influence of contrast concentration on imagequality for coronary CT angiography:comparative assessment of iobitridol350,iopromide370 and iomeprol400 in a multicentre,randomized double-blind trialFrantz HebertGuerbet Company, France. frantz.hebert@guerbet-group.com

PURPOSE: This study aimed to demonstrate non-inferi-ority of iobitridol 350 for coronary CT angiography (CTA)when compared to contrast media with higher iodine con-tent in terms of rate of patients with CT scans fully evalu-able for the presence or absence of coronary arterystenoses.MATERIALS AND METHODS: In this European multi-centre double-blind trial, patients suspected for coronaryartery disease and scheduled for clinically indicated coro-nary CTA using CT systems with 64-detector rows ormore were randomized to receive either iobitridol 350,iopromide 370 or iomeprol 400 in a blinded fashion, atstandardized volume and injection rate based on weight.Per patient, 18 coronary segments were graded regard-ing image quality and interpretability (score 0 = non diag-nostic to 4 = excellent quality) by two core lab readers.For the primary endpoint each patient’s coronary CT scanwas considered as ‘evaluable’ if no segment had a scoreof 0 and the rates of patients with evaluable CT scanswere compared between contrast agents. Mean vascularsignal attenuation, signal to noise ratio (SNR), contrast tonoise ratio (CNR) and the safety profile were assessed assecondary endpoints.RESULTS: A total of 452 patients were analyzed. Theiodine load (in g) injected was significantly differentbetween the 3 groups: 27.8 ± 3.4 (iobitridol), 29.3 ± 3.8(iopromide) and 31.7 ± 3.8 (iomeprol). The rate ofpatients with evaluable CT scans was 92.1%, 95.4% and94.6% for iobitridol, iopromide and iomeprol, respectively,and the non-inferiority of iobitridol compared to the high-est rated comparator was demonstrated with a 95% confi-dence interval of the difference of [8.8%; 2.1%]. Althoughaverage arterial signal attenuation increased with higheriodine concentrations in contrast agent, SNR, CNR andthe safety profiles did not differ between groups.CONCLUSION: In coronary CTA, iobitridol 350 is notinferior in terms of evaluability for the presence orabsence of coronary artery stenosis when compared toiopromide 370 and iomeprol 400. This is evidence that incoronary CTA, iodine concentration can be decreased to350 mg/ml while maintaining the necessary image quality.

SS 17 CV-04 10:20Common causes and clinical significance ofdiscrepancy between coronary CT angiographyand invasive coronary angiographyJunghoon Kim, Eun Ju Chun, Sang Il Choi Seoul National University Bundang Hospital, Korea. humandr.eunju@gmail.com

PURPOSE: To evaluate matched and mismatchedresults between invasive coronary angiography (ICA) andcoronary computed tomography angiography (CCTA)findings, and analyze the causes and clinical significanceof discrepancy between ICA and CCTA.MATERIALS AND METHODS: We evaluated 123 sub-jects with suspicious CAD who underwent CCTA andthen were referred for ICA from December 2009 until July2011. Patient- and segment-based diagnostic perfor-mances were calculated to detect the significant CAD (≥50% in luminal stenosis). We also analyzed the causes offalse positive (FP) and false negative (FN) with CCTAfindings in comparison to ICA as reference standard.During the follow-up (median 40 months, 6-54M), weassessed the predictive value of the true positive (TP),true negative (TN), FP and FN for the major adverse car-diac events (MACE: death, myocardial infarction, unsta-ble and stable angina with revascularization).RESULTS: A total 97 of 123 patients (78.9%) had signifi-cant CAD according to ICA. The diagnostic accuracy ofCCTA on patient- and segment-based analysis was93.5% and 95.0%, respectively. Among 1968 segmentanalysis, 58 segments (2.9%) revealed FP with followingcauses: blooming artifact with calcification (n = 35), posi-tive arterial remodeling (n = 10), noise (n = 7) and motionartifact (n = 6). 41 segments (2.1%) revealed FN with fol-lowing causes; small vessel caliber (n = 18), noise (n =11), motion artifact (n = 9) and calcification (n = 3). Duringthe follow up, cardiovascular event has occurred in 82 of95 patients with TP (86.3%), 4 of 6 FP (66.7%) and all FN(n = 2, 100%), however, no event in TN. Three re-eventsoccurred in other target lesions of TP group. The causesof 91 total events were significant stenosis (n = 85) andpositive arterial remodeling (n = 6), one of the main caus-es of FP.CONCLUSION: CCTA shows high diagnostic accuracyfor CAD. Moreover, among the causes of FP of CCTA,positive remodeling should be considered as significantCAD, because it reflects vulnerable plaque, commoncause of cardiovascular events.

SS 17 CV-05 10:30Modified protocol of triple rule-out CTangiography with restricting field view fordetection of pulmonary thromboembolism andaortic dissection in emergency departmentpatientsHyun Su Kim, Sung Mok Kim, Hae Jin Kim, Moon Young Kim, Jin-Ho Choi, Yeon Hyeon Choe Samsung Medical Center, Korea. holymosquito@hanmail.net

BACKGROUND: Triple rule-out (TRO) CT protocolrequiring whole thorax scan coverage results in higherradiation does compared with coronary CT angiography

176 KCR 2015

Cardiovascular S

ep 11, Fri

(CCTA) using cardiac scan coverage. However, the inci-dences of pulmonary thromboembolism (PTE) and aorticdissection (AD) are low in patients presenting with acutechest pain. There is no study comparing the detectabilityof PTE and AD in each field of view (FOV) of TRO CTand CCTA in large number of patients who underwentconventional TRO CT.PURPOSE: To evaluate potential diagnostic performanceof TRO CT with restricted field of view for detection ofPTE and AD.MATERIALS AND METHODS: This study included1,242 consecutive patients with acute chest pain who vis-ited the emergency department and underwent TRO CTusing a 128-slice dual-source CT. Image data was recon-structed according to the FOV of CCTA and TRO CT pro-tocols in each patient. Two radiologists performed con-sensus analysis for the presence of PTE and AD in eachFOV. The radiation dose was measured using the phan-tom model to evaluate the potential benefits by restrictingcardiac scan coverage instead of the whole thorax.RESULTS: Among all patients, fifteen cases with PTE(1.2%) and nine cases with aortic dissection (0.7%) werefound. Except for one PTE case, all cases were detectedon both FOV of TRO CT and CCTA. In the phantomstudy, the effective radiation doses of whole-thorax andcardiac scans were 4.5 mSv and 2.5 mSv, respectively.CONCLUSION: Isolated PTE and AD outside the cardiacFOV rarely occur. Therefore, modified TRO CT protocolusing cardiac scan coverage can be adopted to detectPTE and AD with reduced radiation dose.

SS 17 CV-06 10:40Evaluation of effectiveness of intelligentboundary registration to correct mis-registrationartifact in coronary artery CTSeung-Hyeon Lim, Joon-Won KANG, Hyun Jung Koo, Dong Hyun Yang, Tae-Hwan Lim Asan Medical Center, Korea. jwonkang@amc.seoul.kr

PURPOSE: To evaluate the effectiveness of IntelligentBoundary Registration (IBR) on the vessel images withdifferent signal-to-noise ratio (SNR), contrast-to-noiseratio (CNR), and different ECG-gating mode in a phantommodel.MATERIALS AND METHODS: Coronary artery phantommodel has 6 vessels which have four different diameters(0.5, 1, 2, 4 mm) of vessel and three different levels ofiodine concentration in (0.5, 1, 2, 4)-mm diameter vessel.We selected 2 mm (vessel 1, 3, 5) and 4 mm (vessel 2, 4,6) diameter of vessel and assigned numbers (vessel 1, 2,3, 4, 5, 6) as seen in the phantom in order. First group ofvessel 1, 2 which included iodine concentration of 10mg/cm3 and second group of vessel 3, 4 which includediodine of 2.5 mg/cm3 and third group of vessel 5, 6 whichincluded iodine concentration of 5 mg/cm3. We analyzedwhether misregistration occurred or not for evaluation ofeffectiveness of IBR. And we measured diameter of eachvessel, vessel HU, background HU, Standard Deviationfor CNR and SNR, respectively. Pearson’s correlationanalysis was performed for evaluation of association ofeffectiveness of IBR effect with scan parameters.RESULTS: Misregistration was observed in all cases

both prospective and retrospective ECG-gating mode. Inretrospective ECG-gating, when vessels have sameiodine concentration level, IBR was more effective inlarge-diameter vessel more than small-diameter vessel.IBR was more effective at high contrast more than poorcontrast. Pearson coefficients for correlation between theIBR effect and the Vessel HU were R = 0.638 (p < 0.01),between IBR effect and kV were R = -0.78 (p = 0.425),between IBR effect and mA were R = 0.21 (p < 0.05),between IBR effect and CNR were R = 0.554 (p < 0.01),between IBR effect and SNR were R = 0.487 (p < 0.01),respectively. On the other hand, IBR was not effectiveregardless of all of scan parameter and conditions ofdiameter of the vessel, even high CNR and high SNR inprospective ECG-gating. It seems to be Z-gap in the ves-sel images.CONCLUSION: Even if lower heart rate is maintainedduring examination, IBR doesn’t work in the conditionswith lower SNR, CNR, small-diameter vessel andprospective ECG-gating. We recommend to proper usescan parameter and high-iodine concentration of contrastmedia for good effect of IBR.

SS 17 CV-07 10:50Influence of Scan technique on intracoronarytransluminal attenuation gradient in coronary CTangiography using 128-slice dual source CT:multi-beat versus one-beat scanSung Mok Kim, Hae Jin Kim, Jin-Ho Choi, Moon Young Kim, Yeon Hyeon Choe Samsung Medical Center, Korea. holymosquito@hanmail.net

PURPOSE: The purpose of this study was to compareintracoronary TAG on CCTA acquired from multi-beat andone-beat scan techniques in the same patients.MATERIALS AND METHODS: We retrospectivelyreviewed CT images of patients with aortic disease whounderwent coronary and aorta CT scan at the same timeusing 128-slice dual source CT. Coronary and aorta CTscans were performed with standard-pitch helical mode(multi-beat imaging) and high-pitch helical mode (one-beat imaging), consequently. Both two scan data wereacquired using ECG-gating. We divided into two groupsdepending on the presence of significant coronary steno-sis with greater than 50% diameter stenosis. 150 signifi-cant stenoses (50 lesions in each 3 vessels) from 121patients (68 ± 9.2 years, 79% men) were included. 150normal coronary arteries from 50 patients (49 ± 14years, 50% men) were included for control group. TAGwas determined from the change in HU per 10 mm lengthof the coronary artery, and defined as the linear regres-sion coefficient between intraluminal radiological attenua-tion (HU) and length from the ostium (mm). TAGs werecompared between two scan techniques using paired t-test or Mann Whitney test.RESULTS: TAGs were calculated on each major coro-nary artery as follows (multi-beat/one beat): RCA, -0.57± 0.70 / -0.70 ± 0.79; LAD, -0.83 ± 0.65 / -0.91 ±0.75; LCX, -0.79 ± 0.69 / -1.08 ± 1.02 in normal group.There were no significant differences on TAGs betweentwo scan techniques (p = NS, all). In stenosis group,TAGs were as follows: RCA, -1.36 ± 0.91 / -1.30 ±

Cardiovascular 177C

ardiovascular Sep 11, Fri

0.70; LAD, -1.54 ± 0.62 / -1.25 ± 0.53; LCX, -1.98 ±0.86 / -1.73 ± 0.94. There were no significant differencesexcept LAD (p = 0.014). In comparison of normal andstenosis groups, there were significant differences ineach vessel and each technique (p < 0.001).CONCLUSION: TAGs were no significant differencesbetween multi-beat scan and one-beat scan except signif-icant LAD stenosis. This finding suggests that TAGs werenot affected from CT acquisition time in significant steno-sis of RCA and LCX as well as normal arteries. Althoughthe TAG of stenotic LAD did not show temporal uniformitybetween two scan techniques, there were significant dif-ferences between normal and stenotic LAD in each scantechnique.

SS 17 CV-08 11:00 Diagnostic performance of TAG of adenosinestress single-shot coronary CTA in patients withsignificant coronary artery stenosis: comparisonwith stress perfusion cardiac MRIHee Yeong Kim1, Hwanseok Yong2, Eun-Young Kang2

1Kangnam Sacred Heart Hospital, 2Korea UniversityGuro Hospital, Korea. yhwanseok@naver.com

OBJECTIVES: To evaluate the value of transluminalattenuation gradient (TAG) of adenosine stress single-shot coronary computed tomography angiography (CTA)using wide area detector CT comparing to cardiac mag-netic resonance (MR) with adenosine stress perfusion.BACKGROUND: Coronary CTA has diagnostic limitationfor predicting functionally significant coronary arterystenosis. TAG of adenosine stress coronary CTA mayimprove the diagnostic accuracy of coronary CT to detectstenosis inducing myocardial ischemia. Stress perfusioncardiac MR is a reliable, noninvasive study for evaluatingmyocardial ischemia.MATERIALS AND METHODS: Twenty one patients withsignificant coronary stenosis on invasive coronary angiog-raphy were prospectively included. All patients underwentadenosine stress single-shot coronary CTA using widearea detector CT and adenosine stress cardiac MRI usinga 3.0-T scanner as a gold standard. TAG of coronaryCTA in all vessels was calculated and stress perfusioncardiac MR images were analyzed by two cardiac radiolo-gists. A TAG cut off value of -15.1 HU/10 mm as previ-ously described was defined as significant.RESULTS: On a vessel-based analysis, TAG of coronaryCTA in all coronary arteries had sensitivity, specificity,positive and negative predictive values of 90.5% (19/21),90% (27/30), 86.4% (19/22) and 93.1% (27/29), respec-tively. In coronary arteries with moderate stenosis, TAGof coronary CTA had sensitivity, specificity, positive andnegative predictive values of 93.3% (14/15), 100%(12/12), 100% (14/14) and 92.31% (12/13), respectively.Mean effective radiation doses for adenosine stress sin-gle-shot coronary CTA and rest coronary CTA were 8.09mSv and 2.13 mSv, respectively.CONCLUSION: TAG of adenosine stress single-shotcoronary CTA using wide area detector CT provides highdiagnostic accuracy for the detection of hemodynamicallysignificant coronary artery stenosis.

SS 17 CV-09 11:10 A vulnerability marker for coronary artery plaqueusing ECG-gated CT histographic analysis:comparison with intravascular ultrasound (IVUS)Jongmin Park1, Jongmin Lee1, Eun-Ju Kang2, Jihoon Hong1

1Kyeongpook National University Hospital, 2Dong-AUniversity Hospital, Korea. jonglee@knu.ac.kr

PURPOSE: To develop and validate a new vulnerabilitymarker (V-marker) for coronary artery plaque, based onhistographic distribution and stratified HU acquired duringCT coronary angiography (CTCA).MATERIALS AND METHODS: Forty-one plaques from30 patients examined by CTCA and IVUS were includedin this study. Using a freeware (ImageJ), 2-dimensionalplaque segmentations were manually performed toinclude whole plaque volume. Per plaque level, a pixelcount-HU scatter plot was acquired. HU axis was strati-fied to three zones: -40 HU-40 HU, 41 HU-120 HU, and121 HU-1500 HU. Pixel count axis was divided to threesame-sized zones. To enhance fat-equivalent density, 9zones by pixel count-HU matrix were multiplied byweighting factors from 1 to 9. The V-marker was validatedusing ANOVA test based on IVUS plaque classifications:lipid-rich, fibrous, and calcified plaques.RESULTS: From IVUS data, 5 lipid-rich, 7 fibrous, and 29calcified plaques were depicted. The V-markers showedsignificantly different mean values among the threegroups (5.68 ± 0.67, 4.11 ± 0.75, and 2.27 ± 0.36,respectively, p < 0.0001). Although mean HUs showedsignificant differences (67 ± 37 HU, 131 ± 59 HU, and691 ± 267 HU, respectively, p < 0.0001), less overlaps of95% confidence interval were noted in the V-marker thanmean HU method.CONCLUSION: A new vulnerability marker for coronaryartery plaque was developed and validated to be feasible.The V-marker could reliably predict IVUS-diagnosed lipid-rich plaques during CTCA.

178 KCR 2015

Cardiovascular S

ep 11, Fri