Patient Population

We studied 107 consecutive patients who were subjected to DipSE for known or suspected coronary artery disease. No patient was excluded on the basis of image quality. Patients were excluded only if they had rhythm disturbances that precluded the acquisition of electrocardiographically gated 3D full-volume data sets and/or contraindications to dipyridamole. The study protocol was approved by our institutional ethics committee, and all patients gave written informed consent.

Two-Dimensional Echocardiography

Patients were studied in the left lateral decubitus position after establishing an intravenous line for dipyridamole infusion. Conventional 2D views (parasternal long-axis; parasternal short-axis at the papillary muscle level’ and 3 apical views: 4-chamber, 2-chamber, and long-axis) were acquired by an experienced investigator using a commercially available Vivid E9 ultrasound machine (GE Vingmed Ultrasound AS, Horten, Norway) equipped with an M5S probe using the grayscale second-harmonic 2D imaging technique, with adjustment of image contrast, frequency, depth, and sector size for adequate frame rate and optimal LV border visualization. Care was taken to avoid LV foreshortening in apical views, and image acquisition was done during breath hold to minimize respiratory movements.

Electrocardiography, blood pressure and oxygen saturation were continuously assessed using a commercially available monitoring system (Vitalogik 5000; Mennen Medical Ltd, Tel Aviv, Israel).

Three-Dimensional Echocardiography

Data set acquisition of the left ventricle was performed by the same examiner at the end of the standard 2D examination, using the same echocardiographic scanner, by switching to a 3V matrix-array transducer (GE Vingmed Ultrasound AS). A full-volume scan was acquired using second-harmonic imaging from the apical approach, and care was taken to encompass the entire LV cavity in the data set by using the medium volume size to maintain adequate spatial resolution. Consecutive 4-beat electrocardiographically gated subvolumes were acquired during end-expiratory apnea to generate the full-volume data set. The quality of the acquisition was then verified in each patient by selecting 9-slice display mode available on the machine to ensure optimal imaging of the entire LV endocardium at each short-axis level, and if unsatisfactory, the data set was reacquired.

Both 2D cine loops and real-time 3D echocardiographic data sets were stored digitally in raw format and exported to a separate workstation equipped with commercially available software for offline analysis (EchoPAC PC version 108.1.4; GE Vingmed Ultrasound AS).

DipSE

Patients were asked to avoid drinks containing xanthine for ≥48 hours before the study. Dipyridamole was infused at a dose of 0.84 mg/kg over 6 minutes, unless symptoms of intolerability, positivity for ischemia, major arrhythmias, or hypotension (>30 mm Hg decrease in blood pressure) occurred. Atropine was not used to increase heart rate. Two-dimensional and 3D echocardiographic acquisitions were performed by the same investigator at baseline, during peak stress (2 minutes after the end of dipyridamole infusion), and at recovery (2 minutes after teofilline infusion) by quickly switching between the 2D and 3D echocardiographic transducers ( Figure 1 ). The sequence of image acquisition by the two techniques was similar in all patients; that is, 3D echocardiographic data sets were obtained immediately after the 2D acquisitions.

Image acquisition protocol during DipSE using 2DE and 3DE.

LV Wall Motion Analysis

The left ventricle was divided into 17 segments assigned to coronary artery supply according to the European Association of Echocardiography. Image analysis was performed by an experienced echocardiographer (L.P.B.) unaware of patients’ clinical and angiographic data, who analyzed stored 2D cine loops and 3D echocardiographic data sets in random order (the reader did not know which 2D recording corresponded to a given 3D study). Quad-screen display was used to view the 2D parasternal long-axis and short-axis views and apical 4-chamber, 2-chamber, and long-axis views at baseline ( Video 1 ). The baseline 3D echocardiographic data set was first correctly aligned to visualize the true LV apex and then automatically sliced in 3 apical and 9 equidistant 2D short-axis planes orthogonal to the z -axis of the original data sets ( Figure 2 , Video 2 ). When needed, x -axis and y -axis fine rotation was performed to ensure that these short-axis planes were perpendicular to the true LV long-axis dimension.

Multislice display of a 3D data set of the left ventricle acquired at baseline from the apical approach showing akinesia of the septal apical segment. (Left) Three apical views (from top to bottom : 4-chamber, 2-chamber, and long-axis). The 3 columns of short axes show 9 equidistant short-axis views of the left ventricle progressing from the apex (upper left) to the base (lower right) of the left ventricle. The accompanying video shows these images in real time ( Video 2 ).

Wall motion at peak stress and during recovery was assessed using a side-by-side visual assessment of the same electrocardiographically gated 2D view ( Video 3 ) and 3D echocardiographic plane from each of the 3 stages of DipSE visualized in quad-screen display ( Figure 3 , Videos 4 a and 4b). A normal response to DipSE was defined as increased wall motion or increased wall thickening during stress. The development of new or worsening resting LV wall motion abnormalities in >1 contiguous segment of the same vascular territory was considered a sign of inducible ischemia. Segmental wall motion was scored as follows: 1 = normal or hyperkinetic, 2 = hypokinetic, 3 = akinetic, and 4 = dyskinetic. A global wall motion score index was derived by dividing the sum of individual visualized segment scores by the number of visualized segments. To compare 2D echocardiography (2DE) and 3DE in the evaluation of the apex, the regional wall motion score of the 5 apical segments was calculated. Segments with resting wall motion abnormalities that did not show any change during dipyridamole infusion were defined as scar. Akinetic segments that became dyskinetic were not considered indicative of myocardial ischemia.

(Left) Side-by-side comparison of 2D parasternal short-axis views acquired at baseline, during low-dose and peak-dose infusion, and during recovery. (Right) Side-by-side comparison of 2D slices (4-chamber views) obtained from 3D data set acquired at baseline, during peak stress, and at recovery showing inducible lateral ischemia. The 2D equivalent view from the same patient is displayed for image quality comparison. The accompanying videos show these images in real time ( Video 3 and 4 ).

Coronary Angiography

Coronary angiograms were obtained in 44 patients within 1 month (10 ± 9 days) without any intervening event, on the basis of the recommendations of the referring physician. Luminal narrowing > 50% was considered significant.

Intraobserver and Interobserver Agreement

To assess the reproducibility of wall motion analysis on 3DE and 2DE, the data sets of 10 random patients were reanalyzed by the same observer (L.P.B.) 3 months after the first assessment, as well as by a second observer (F.R.) blinded to the results of the first observer. Concordance rates (normal vs abnormal) were calculated on a patient basis and a coronary perfusion territory basis.

Statistical Analysis

Data are expressed as mean ± SD. Student’s t test was used for comparisons of continuous variables. Agreement between the test results obtained using 2DE and 3DE was tested by measuring the coefficient of variation (κ). A κ value ≥ 0.45 was considered to indicate good agreement, and a κ value ≥ 0.75 was considered to indicate excellent agreement. Sensitivity, specificity, and accuracy were calculated in patients who underwent coronary angiography. A P value < .05 was considered statistically significant.