Many studies have shown the feasibility of contrast two-dimensional echocardiography (2DE) for myocardial perfusion imaging. 1 The technique offers several important advantages over competing techniques including its wide availability, portability, low cost, no radiation concern, and the simultaneous visualization of wall motion and perfusion. However, contrast echocardiography has not become a routine procedure both for the diagnosis and assessment of ischaemic heart disease in daily practice. There are several reasons for this: no echo contrast agent has been specially approved for myocardial perfusion imaging, there are concerns about safety and reimbursement issues. In addition, it is well recognized that 2D images of the heart do not provide accurate information on both the extent and severity of myocardial perfusion defects.
Theoretically, 3D imaging allows a more accurate assessment of the myocardial perfusion defects but its use and advantages in patients could not be explored until the development of real-time three-dimensional echocardiography (RT3DE). This technique allows to capture the entire myocardial volume during a single contrast injection, thus obviating the need of repeated injections of boluses of a contrast agent during sequential consecutive acquisition of multiple 2DE planes needed for off-line reconstruction of a volumetric data set. 2 , 3 Early studies both in animals and humans 4 have demonstrated that RT3DE provides accurate volumetric delineation of perfusion defects in close agreement with tissue-staining anatomic reference 5 and volumetric analysis of myocardial perfusion over time. Consequently, RT3DE technology has opened the door to volumetric quantification of myocardial perfusion. 6
However, there are still several technological and methodological problems which must be solved before the technique will reach clinical practice. The low temporal ad spatial resolution, the presence of drop-out artifacts in the volumetric data set, and the limited size of the pyramidal volume, which does not allow imaging the whole ventricle in real-time, are currently the main technological limitations. From the methodological point of view, both defining the 3D region of interest (ROI) and determining the optimal contrast injection methodology are the most important challenges. These are subjects of active academic and industrial research. The technological limitations will most likely be overcome by technological advances.
Quantitative 2DE analysis of myocardial perfusion in a specific territory of the myocardium is based on manual tracing of ROIs in a single imaging plane, and subsequent frame-by-frame realignment of these ROIs to compensate for cardiac translation throughout the cardiac cycle. Myocardial videointensity in these ROIs is measured over time. Defining ROIs in 3D space is difficult and the realignment of 3D ROIs throughout the cardiac cycle to compensate for cardiac translation is even more challenging.
Veronesi et al.7 describe a method for semi-automated definition of a 3D myocardial ROI combined with 3D image registration, which potentially provides a solution for the problems of defining 3D ROIs and tracking them throughout the cardiac cycle. The method proposed by Veronesi et al.7 is original in several aspects. Rather than detecting the endocardium and then expanding the ROI outwards which is the usual approach used in 2DE, they identified the epicardium to define the myocardial ROI more accurately irrespective of variations in myocardial thickness. In addition, instead of trying to minimize cardiac translation artifacts by manually correcting the ROI size and shape on each consecutive frame, they implemented the automated frame-by-frame image registration to fit the left ventricle (LV) in every volume into its reference position. However, the results of this study should be viewed as a very initial step towards development of a technique for analysis of myocardial perfusion. The authors analysed volumetric short-axis data sets containing only the mid-portion of the LV. It is not clear whether the technique also works on a full volume LV data set. In addition, since the identification of the reference frame used to correct for translation may affect the final results, its selection from the contrast steady-state phase images needs to be clarified. Finally, myocardial perfusion was assessed only in normals and its accuracy for detecting perfusion defects needs to be explored.
Another important issue to be addressed is finding the proper contrast injection methodology. Should one use either consecutive contrast boluses or analysis of contrast replenishment after high-energy ultrasound pulses to destroy microbubbles in the myocardium (‘flash-echo’)? 8 Effects of contrast boluses on echocardiographic images are rather unpredictable and optimization of image settings is difficult to achieve in the short period of myocardial contrast enhancement. On the other end, flash-echo methodology cannot be used in RT3DE because of the excessive energy required to destroy microbubbles in the entire heart. Toledo et al.8 proposed to use a continuous contrast infusion to optimize gain settings, then interrupt the infusion to allow contrast clearance, and finally restart contrast infusion at the same infusion rate to assess contrast replenishment. In human subjects, the contrast replenishment could be completed within 45 s, which could be recorded in a single-data acquisition using RT3DE.
In conclusion, technical progress which allowed the development of RT3DE, standardization of methods for effective contrast administration, and automated myocardial segmentation could provide the basis for volumetric quantification of myocardial perfusion. However, there is still a long way to go before this technique may eventually become part of the routine clinical practice. Its use for non-invasive diagnosis of coronary artery disease in humans needs to be validated against reference techniques in large multicentre trials and in outcome studies involving patients with a spectrum of risk categories and likelihood of coronary artery disease. Finally, its cost-effectiveness is another issue that remains to be addressed. 9