Summary
Background
Transthoracic echocardiography is the most commonly used tool for the detection of left ventricular wall motion (LVWM) abnormalities using “naked eye evaluation”. This subjective and operator-dependent technique requires a high level of clinical training and experience. Two-dimensional speckle-tracking echocardiography (2D-STE), which is less operator-dependent, has been proposed for this purpose. However, the role of on-line segmental longitudinal peak systolic strain (LPSS) values in the prediction of LVWM has not been fully evaluated.
Aim
To test segmental LPSS for predicting LVWM abnormalities in routine echocardiography laboratory practice.
Methods
LVWM was evaluated by an experienced cardiologist, during routine practice, in 620 patients; segmental LPSS values were then calculated.
Results
In this work, reflecting real life, 99.6% of segments were successfully tracked. Mean (95% confidence interval [CI]) segmental LPSS values for normal basal ( n = 3409), mid ( n = 3468) and apical ( n = 3466) segments were –16.7% (–16.9% to –16.5%), –18.2% (–18.3% to –18.0%) and –21.1% (–21.3% to –20.9%), respectively. Mean (95% CI) segmental LPSS values for hypokinetic basal ( n = 114), mid ( n = 116) and apical ( n = 90) segments were –7.7% (–9.0% to –6.3%), –10.1% (–11.1% to –9.0%) and –9.3% (–10.5% to –8.1%), respectively. Mean (95% CI) segmental LPSS values for akinetic basal ( n = 128), mid ( n = 95) and apical ( n = 91) segments were –6.6% (–8.0% to –5.1%), –6.1% (–7.7% to –4.6%) and –4.2% (–5.4% to –3.0%), respectively. LPSS allowed the differentiation between normal and abnormal segments at basal, mid and apical levels. An LPSS value ≥ –12% detected abnormal segmental motion with a sensitivity of 78% for basal, 70% for mid and 82% for apical segments.
Conclusions
Segmental LPSS values may help to differentiate between normal and abnormal left ventricular segments.
Résumé
Contexte
L’évaluation visuelle en échocardiographie transthoracique est l’outil le plus couramment utilisé pour analyser les anomalies de la cinétique segmentaire du ventricule gauche. Cette technique subjective requiert un opérateur expérimenté. L’imagerie de Speckle bidimensionnelle semble moins dépendante de l’opérateur et a été proposée à cet effet. Cependant, le rôle des valeurs des pics systoliques du strain longitudinal (LPSS) segmentaire obtenues en direct pendant l’examen échographique pour la prédiction des anomalies de cinétique segmentaire n’a pas été pleinement évalué.
Objectif
L’objectif de cette étude était de tester la capacité des valeurs de LPSS segmentaire à prédire les anomalies de cinétique segmentaire, dans un contexte de pratique échocardiographique de routine.
Méthodes
La cinétique segmentaire du ventricule gauche a été évaluée par un cardiologue expérimenté, chez 620 patients consécutifs. Les valeurs des LPSS segmentaires ont ensuite été calculées durant le même examen.
Résultats
Au total, 99,6 % des segments ont été traqués avec succès. Les valeurs moyennes (IC 95 %) du LPSS pour les segments basaux normaux ( n = 3409), moyens normaux ( n = 3468) et apicaux normaux ( n = 3466) étaient : –16,7 % (–16,9 % à –16,5 %), –18,2 % (–18,3 % à –18,0 %) et –21,1 % (–21,3 % à –20,9 %), respectivement. Les valeurs moyennes (IC 95 %) du LPSS pour les segments basaux hypokinétiques ( n = 114), moyens hypokinétiques ( n = 116) et apicaux hypokinétiques ( n = 90) étaient : –7,7 % (–9,0 % à –6,3 %), –10,1 % (–11,1 % à –9,0 %) et –9,3 % (–10,5 % à –8,1 %), respectivement. Les valeurs moyennes (IC 95 %) du LPSS pour les segments basaux akinétiques ( n = 128), moyens akinétiques ( n = 95) et apicaux akinétiques ( n = 91) étaient : –6,6 % (–8,0 % à –5,1 %), –6,1 % (–7,7 % à –4,6 %) et –4,2 % (–5,4 % à –3,0 %), respectivement. Les valeurs segmentaires du LPSS ont permis la différenciation entre les segments normaux et anormaux aux trois étages (basal, moyen et apical). Une valeur de LPSS ≥ –12 % détecte une cinétique segmentaire anormale avec une sensibilité de 78 % au niveau basal, 70 % au niveau moyen et 82 % au niveau apical.
Conclusion
Les valeurs segmentaires du LPSS peuvent aider à différencier les segments anormaux des segments normaux du ventricule gauche.
Background
Echocardiography is the main, but not the only tool for the evaluation of left ventricular wall motion (LVWM). The analysis is usually based on visual “naked eye” evaluation. The interpretation of LVWM abnormalities is an important component of transthoracic echocardiography (TTE), but is prone to intra- and interobserver variability , especially when the acoustic window is suboptimal and the echographer less experienced.
New echocardiographic techniques are now available for the assessment of left ventricular (LV) systolic function, one of which is two-dimensional speckle-tracking echocardiography (2D-STE). This technique allows the study of global myocardial deformation by an index called global longitudinal strain (GLS) , and segmental myocardial deformation by the measurement of segmental longitudinal peak systolic strain (LPSS) values. Deformation variables are thought to be more reproducible, and can now be obtained on-line during the ultrasound examination, therefore requiring no off-line analysis .
This study aimed to determine the usefulness of segmental LPSS values for predicting LVWM abnormalities during routine in-hospital echocardiographic activity on consecutive patients, regardless of the indication for TTE. The study is based on the hypothesis that a method that requires only limited user interaction would help to improve the robustness of echocardiographic assessments of LVWM abnormalities in routine clinical practice, as has been demonstrated to be the case for LV ejection fraction (LVEF) .
Methods
Study population
A total of 620 consecutive patients undergoing TTE assessment in our echocardiography laboratory by the same echocardiographer (N.B.) between August 2012 and March 2014 entered the analysis. The echocardiographer had 12 years of experience, carrying out 700 TTE scans per year. The indications for TTE were as follows: ischaemic stroke or transient ischaemic attack ( n = 222; 11 with ischaemic heart disease); multiple sclerosis ( n = 44); dyspnoea ( n = 24); evaluation of various treatments ( n = 19); heart murmur exploration and control of valvular disease, endocarditis and suspicion of endocarditis ( n = 17); LVEF evaluation and/or heart failure ( n = 17); preoperative assessment ( n = 15); haemorrhagic stroke ( n = 13); hypertension ( n = 12); diabetes mellitus ( n = 11); assessment of the pericardium ( n = 5); occlusion of the central retinal artery ( n = 4); ischaemic heart disease ( n = 4); dissection of supra-aortic vessels ( n = 4); subarachnoid haemorrhage ( n = 3); assessment of arterial pulmonary pressures ( n = 3); control of LV thrombus ( n = 2); chest pain ( n = 2); pulmonary embolism ( n = 2); atrial fibrillation ( n = 2); arteritis of the lower limbs ( n = 1); Lyme disease ( n = 1); non-compaction cardiomyopathy ( n = 1); Sjögren’s syndrome ( n = 1); sarcoidosis ( n = 1); intracerebral lesions ( n = 1). The other indications were classified as “miscellaneous”.
Clinical and echocardiographic characteristics of the first 507 patients have been reported elsewhere .
Echocardiographic analysis
TTE was performed using a commercial ultrasound system (Vivid 7; GE Healthcare, Horten, Norway), using a 4 MHz transducer.
Standard TTE included LV analysis, LVEF calculation (Simpson’s biplane method in patients in sinus rhythm; visual evaluation in patients in atrial fibrillation), tissue Doppler imaging, “naked eye” evaluation of LVWM and the calculation of LV GLS and segmental LPSS values. Firstly, a comprehensive analysis of cardiac anatomy was performed. Patient echogenicity was classified as 1 (good), 2 (moderate) or 3 (poor). The acoustic window had to be of sufficient quality to allow the calculation of LVEF, GLS and LPSS values and the visual evaluation of LVWM; otherwise, TTE was excluded.
Segments were classified on-line as 1 (normal), 2 (hypokinetic = 2), 3 (akinetic), 4 (dyskinetic) or 5 (paradoxical [for the septum]).
As the aim of the study was to assess the value of segmental LPSS for the detection of LVWM abnormalities in a real-life echocardiography laboratory setting, there was no rereading by a second observer and no rereading by the same operator. All measurements were performed on-line, on the Vivid 7, without off-line analysis.
Speckle-tracking strain analysis
The three apical views (four-, two- and three-chamber views) were recorded with a frame rate between 70 and 80 Hz, for the 2D-STE study. Myocardial deformation was assessed in a semiautomatic manner, based on grayscale images. The analysis was initiated through the apical three-chamber view, followed by the four- and two-chamber views. To initialize the analysis, three anatomical landmarks were set manually on each of the two points of the mitral annulus and on the apical endocardium. The software automatically placed the region of interest on the endocardial cavity. If the tracking was inadequate, manual adjustments were performed. As previously reported , each LV wall was divided into three segments: basal, mid and apical. Segmental LPSS values were hence obtained during the standard TTE.
Post-systolic shortening was not considered in this study. Fig. 1 displays an example of output.
Statistical analyses
All values are presented as means ± standard deviations, medians (interquartile ranges) or absolute numbers and frequencies.
LPSS values of normal basal, normal mid and normal apical segments were compared using the Wilcoxon signed rank sum test. The same analyses were performed for hypokinetic basal, hypokinetic mid and hypokinetic apical segments, and then akinetic basal, akinetic mid and akinetic apical segments.
To compare LPSS values at each level (basal, mid and apical), according to segmental kinetics, normal segments were labelled “1”, hypokinetic segments were labelled “2” and akinetic, dyskinetic and paradoxical segments were pooled into one sole category labelled “3”.
Analysis of variance was used for the comparisons between these three categories.
On logistic regression, the area under the receiver operating characteristic (ROC) curve (AUC) was analysed for LPSS values, in order to identify the threshold of LPSS that predicts abnormal segmental wall motion.
A P -value < 0.05 was considered to be statistically significant, except for the multiple comparisons of mean LPSS values between normal and hypokinetic segments, normal and akinetic/dyskinetic segments and hypokinetic and akinetic/dyskinetic segments at each level (basal, mid and apical). Using the Bonferroni correction, only individual tests with P < 0.0001 were considered to be significant.
Statistical analyses were performed using Stata ® software, version 13 (StataCorp LP, College Station, TX, USA).
Results
Population and echocardiographic characteristics
Among the screened patients, 52 were excluded (33 had insufficient echogenicity and 19 had atrial fibrillation with important heart rate variability, precluding strain analysis); the study therefore involved 620 TTE scans. The main clinical and echocardiographic characteristics are reported in Table 1 .
| Characteristic | |
|---|---|
| Age (years) | 59.0 ± 17.7 |
| Men | 327 (52.7) |
| Atrial fibrillation | 19 (3.1) |
| LVM indexed to body surface (g/m 2 ) | 95.9 ± 32.0 |
| LVM indexed to body surface ≥ 125 g/m 2 | 102 (16.5) |
| Wall motion abnormality | 125 (20.2) |
| LVEF (%) | |
| Overall, mean a | 64 ± 11 |
| Overall, median a | 65 (59–71) |
| Without LVWMA a | 67 ± 8 |
| With LVWMA a | 53 ± 15 |
| LVEF < 55% | 80 (12.9) |
| GLS (%) | |
| Overall, mean | –18.1 ± 4.0 |
| Overall, median | –19 (–21,–16) |
| Apical three-chamber | –18.0 ± 4.7 |
| Apical four-chamber | –17.7 ± 4.1 |
| Apical two-chamber | –18.6 ± 4.5 |
| Without LVWMA | –19.2 ± 2.9 |
| With LVWMA | –13.7 ± 4.6 |
| Echogenicity | |
| Good | 314 (50.7) |
| Moderate | 225 (36.3) |
| Poor | 81 (13.0) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
