S 134 Abstracts 884 Automatic lelt ventricular volume measurements from simultaneously acquired triplane echocardiograms J. Hansegaard 1, S. Urheim 2 , E. Sleen s, H. Torp 4 , B. Olslad s, S. Maim 4 , S. Rabben 3. r University of Oslo, Department of Informatics, Oslo, Norway; 2Rikshospitalet UniversityHospital, Oslo, Norway; 3GE Vingmed Ultrasound, Horten, Norway; +NUST, Dopt of Circulation and Mod. Imaging, Trondhoim, Norway We report the results lrom a feasibility study where a new algorithm lor automatic volume measurements in the loll ventricle (LV) based on simultaneously acquired triplane echocardiograms has been evaluated. The accuracy and robustness of LV volume measurements based on automatically detected myocardial boundaries is ~0(pected to improve by making use of lriplane ultrasound (US) image sequences made available through the latest generation el 3D scanners. As a myocardial boundary detector we used an extension of the Active Appear- ance Model, which is a dynamic deformable template model thal exploits spatial coherence between the views lor handling imaging artifacts such as drop outs. This model was trained on a set el 20 manually outlined lriplano image sequences of full cardiac cycles lrom a group of 12 healthy persons and 8 persons sulloring from heart disease. The standard apical lour chamber (A4CH) view and the views rotated 60 ° and 120 ° relative to the A4CH view were recorded simultaneously us- ing a Vivid 7 scanner (GE Vingmed Ultrasound, Horlen .) Patient cases with asyn- chronies were not included. We evaluated the detection algorithm on the 20 cases using a leave one out approach. The point distance was defined as the average of the Euclidian distances between corresponding pairs of 40 unilormly distributed points along the contours. Table 1 shows the mean and 1 standard deviation el the point distance, volume error, and volume lraclional error, and the regression line for computer determined volumes (VolC) as a function of manually determined volumes (VolM.) Table 1 Mean Point distanoe (mm) 4.14-19 Error (VolM -VolO) (ml) 7.04-14 Volume fraotional error (I-VolONolM) (%) 134-12 Volume regression (y=VolC,x=VolM) (ml) y = 1.0x - 11 The detected volumes at ED, mid systole and ES showed excellent correlation with the manually determined volumes (R2=0.87). The correlation between de- tected ejection lractions (EFs) and manually determined EFs was poor (R2=0.29). However, this is probably caused by the method's tendency to idenlily an average motion pattern. Adding more samples to the training database may improve the pedormance in such cases. 885 Wave intensity analysis is an additional tool for the quantification of leR ventricular systole C. Manisly, D. Francis, A. Zambanini, S. Byrd, Z. Whinnolt, S. Them, A. Hughes, J. Mayol. ICCH, Imperial College & St. Mary's Hospital London, United Kingdom Background: Conventional markers of left ventricular systolic function such as ejection lraclion and longitudinal systolic shortening velocity (tissue Doppler veloc- ity - S') have been lound to be poor indicators of prognosis. Many epidemiological studies have demonstrated a close relationship between cardiovascular risk and pulse pressure, and this is thought to be due to decreased arterial compliance. Wave intensity analysis is a novel method el distinguishing the direction and character el the individual components of a vascular waveform and illustrates ventricular-arlerial interaction. Wave intensity is a time domain measure defined as the product of rate of change of pressure and rate el change el flow in a ves- sel. The lorward-travelling systolic wave (S) gives a measure of cardiac ejection coupled with vessel compliance. Method and Results: We used conventional ecbocardiography, tissue Doppler imaging and carotid artery wave intensity analysis to quantify left venlricular sys- tole in 280 patients with underlying hypertension, mean age 62 (SD 8) years. These subjects were a subset of patients in the ASCOT trial, randomised to either am- Iodopine and atenolol as first line antihyperlensive agents, and had good blood pressure control (mean systolic blood pressure 144 (SD 17) mmHg. • Both the total S wave energy and peak S wave intensity correlated with the av- erage S' velocity measured with tissue Doppler (r=0.22., p=0.0OOl and r=0.26, p<0.0001) but not with OlOClion fraction. • Pulse pressure is positively correlated with S wave intensity energy (r=0.374, p=<0.0001 ), however it is not related to either ejection lraction or peak seplal ve- locity. This reflects the peripheral influences in addition to venlricular contraction, on wave intensity energies - unlike tissue Doppler velocities or ejection lraclion. • Similar to ejection lraction and S' velocity - conventional measures el lell ven- lricular systole, S wave measured using wave intensity analysis did not show a relationship with age (p=0.33, r=0.O5). Conclusions: Both cardiac and peripheral elfects significantly influence the sys- tolic wave, as measured by wave intensity analysis. Conventional measurements of isolated cardiac function including ejection fraction and tissue Doppler velocity measurements, have been disappointing at predicting prognosis. Wave intensity analysis oilers additional information on vonlricular-arlorial interaction, and there- fore may oiler clinicians a new and improved prognostic marker. 886 Echocardiographic and electrophysiological definitions of regional wall function after myocardial infarction W. Wrobel, A. Hollmann, C. Czelwinski, A.M Wnuk-Wojnar, A. Rybicka-Musialik, Z. Tabor, K. Wits, M. Trusz-Gluza. Silesian Academy of Medicine,/st. Dep. of Cardiology,Katowice, Poland Background: Nonfluoroscopic 3<limensional eleclroanatomical mapping system CARTO is a novel method used to evaluate electrical activities within regions of the heart muscle. Translhoracic ecbocardiography estimates global and regional lunclion of the lell ventricle (LV). Aim of sludy: To compare in echocardiographic global and segmental function of LV with CARTO bipolar voltage segmental electrical activity in patients with LV dyslunclion and venlricular lachycardia after myocardial infarction (MI). To find cut oil point (in bipolar voltage) between normo-/hypo- and a-/dyskinetic segments Methods: A comparative analysis was performed in 32 patients (25men, 64±9 yrs) qualified to radiofrequency catheter ablation with CARTO system.Global LV lunclion and 12 segmental wall motion score index (WMSI) were assessed by echocardiography 24-48 hours before the procedure. The bipolar voltage (BP) map of LV was made during electroanalomical mapping. We analyzed contraction of 334 segments in echo and bipolar voltage of 2656 points el electroanatomical maps. The segments with less than 3 points marked during CARTO were ~0(cluded. Resulls: The LV ejection fraction (LVEF) was 33%±10%; WMSI 1,8±0,3. There was a significant difference in BP between all groups of segments (table). The receiver operator curve (ROC) was computed to find the cut off point lor BP distinguished dys- and akinetic segments lrom hypo- and normokinetic ones. Ac- curacy (ACC), positive (PPV) and negative (NPV) predictive value el this method were assessed (table). Contraction Number Mean Bip01ar V01tage Standard ANOVA of points [mV] deviation 1 - normokinesis 1216 1,82* 2,02 0,01 2 - hyp0kinesis 850 1,78* 2,32 0,01 3 - a- and dyskinesis 443 0,90* 1,09 0,01 AUC ACC Cut off point [mV] PPV NPV 0.67 0.72 0,3 0,76 0,36 AUC - Area under the ROC curve Conclusion: Electrical activity dilferentiated normo-, hypo- a- and dyskinetic seg- ments. However, the presence el the points with higher bipolar voltage in akinelic and dyskinelic regions showed preserved electrical lunction despite on advanced contraction disturbances. The BP values lower than 0,3 mV seems to give best discrimination el advanced regional dyslunction el heart muscle. 887 Impact ofenhanced platelet reactivity on lelt ventricular remodeling and functional recovery in ST-segment elevation myocardial infarction R. Pialkowski, Z. Huczok, K.J. Filipiak, J. Kochman, J. Kochanowski, G. Opolski. Warsaw Medical Universily Hospital Cardiology Department, Warsaw, Poland Purpose: Plalelel reactivity is believed to play an important role in both ischemic and reperlusion phase el myocardial infarction. We tried to determine whether measurement el plalelol reactivity on admission allows identifying patients with increased risk of impaired left venlricular lunctional recovery and remodeling in ST- segment elevation myocardial infarction (STEMI) treated with primary angioplasly (PCl). Methods: Venous blood samples were collected on admission from 110 patients presenting with STEMI and platelel reactivity (adhesion and aggregation) was esti- mated with the use of PFA-1 OO ® platelet function analyzer (Dade Behring, Newark, DE) as the lime for flowing whole blood to occlude a collagen-adenosine diphos- phale ring, with shorter closure times (CADP-CT) indicating greater reactivity A 2- D echocardiogram was performed at baseline (within 12 h alter PCI) and 1 month and 6 months lherealler. An increase of more than 20% in end diastolic volume index (EDVI) at 6-months relative to the baseline value was considered as lelt ven- lricular remodeling. Significant early (at 1 month) and late (at 6 months) functional recovery was considered to be present when the functional improvement involved at least 2 segments or t segment when only 2 were basally asynergic. Results: Study population was divided according to median CADP-CT (95 sec- onds). Loll venlricular remodeling was present in 56°/, patients of the inframe- dian (n=56) vs. 11.8% of the supramedian group (n=54), (P<0.0001). Suprame- dian group patients showed belier early and late functional recovery (73:1% and 86.3%, respectively) than inlramedian group patients (37.3%, P<0.O01 and 50%, P<0.0001, respectively). In multivariate logistic regression model, alter control- ling lot a series el possible conlounders, CADP-CT =95 sec. (enhanced reactiv- ity) remained an independent predictor of remodeling (OR 7.0, 95% CI 2.2-27, P=0.0016) and the absence el early and late functional recovery (OR 5.4, 95% CI 2.2-13.4, P=0.000G and OR 7.5, 95% CI 2.5-22.1, P=0.0003, respectively). The performance el multivariate models after incorporation el CADP-CT showed im- provement with the concordance (C) index increase from 0.02, to 0.9 for remod- eling, 0.63 to 0.8 for early lunctional recovery, and 0.73 to 0.82 lor late functional recovery. Conclusions: Enhanced platelet reactivity estimated by rapid, point-of-care plateloI lunclion analyzer (PFA-10O ~) is a strong and early predictor of impaired Eur J Echocardiography Abstracts Supplement, December 2005 by guest on March 4, 2013 http://ehjcimaging.oxfordjournals.org/ Downloaded from