Influence of three-dimensional imaging of ventricular structure and function on clinical decision-making

Abstract Both qualitative and quantitative limitations of two-dimensional (2DE) imaging have lead to the emergence of the three-dimensional echo (3DE) technique. Until recently the development of 3DE technology has been slow due to inferior image quality and longer processing and acquisition times compared with 2DE. However, in recent times 3DE technology has advanced to “real-time” (RT3D) and has become feasible as a standard clinical tool. The overall hypothesis of the studies undertaken in this thesis is that the increased accuracy of left ventricular (LV) assessment using RT3D justifies its use in clinical decision-making based on both single and sequential measurements in routine echocardiographic practice. It shows the increased accuracy and feasibility of 3DE using online and offline techniques and opacification, and evaluates the prognostic implications of LV measurements globally and regionally using 3DE compared with 2DE. Further, it assesses the accuracy of LV assessment using 3DE and demonstrates its superior accuracy to 2DE in cross-sectional studies. The thesis initially reviews the use of physiologic parameters used for clinical decision making and looks at the principles of incremental value in diagnostic testing. The assessment of ventricular structure and function is discussed as well as the current problems with clinical echocardiography, which have lead to the evolution and development of 3DE. The second chapter describes the methodologies used in this thesis, with particular emphasis on new technologies used in the studies. Chapter 3 looks at the role of experience as a barrier to incorporation of 3DE into standard practice. We tested attendees before and after an intensive interactive training course to evaluate the learning curve. An interactive teaching course with rehearsal and direct mentoring appears to overcome this limitation and may improve the uptake of the 3DE technique. Chapters 4 to 6 validate the use of 3DE by comparing it to 2DE and magnetic resonance imaging (MRI) with respect to online and offline techniques, the use of left ventricular opacification (LVO) and the serial follow-up of patients. The availability of automated on-line software may increase the feasibility of real-time 3DE for LV volume calculation in clinical practice. Chapter 4 looks at the comparison of off- and on-line approaches with magnetic resonance imaging. Online measurement of LV volumes was found to be feasible and more accurate than with 2DE, however the offline approach was more accurate - although more time-consuming. Chapter 5 involves two studies which sought to validate 3DE in a multicentre setting. They examine the variability of LV measurements across multiple sites and to validate 3DE against MRI. The first study found that the semi-automated measurement of 3DE reduces the variation of LV volume between centers. The second study provides information on the role of different potential sources of error and provides guidelines for future users on how to minimize these errors as well as how to interpret their findings. The next chapter further validates 3DE with use of LVO as a technique to improve the accuracy of LV volume measurements. We sought to examine the accuracy of non-contrast and contrast enhanced 2DE and 3DE for calculation of LV volumes and ejection fraction (EF), relative to cardiac magnetic resonance imaging. 2DE with LVO was analogous to 3DE in accurate categorization of LV function. However, 3DE with LVO was feasible and superior to other non-contrast and contrast enhanced techniques in patients with previous infarction. Chapter 7 examines the technical details of regional volume measurements. The use of a fixed external frame of reference in analyses of regional wall motion in the apical four-chamber view is prone to a systematic error and the use of a floating-axis analysis corrects for this, but this is based on landmarks by MRI (apex, annulus) and center of mass by 3DE. The axis is likely to be influenced by reverse remodeling after intervention and therefore we sought to evaluate 3DE regional volume assessment vs. cardiac MRI over follow-up. The use of a center-line based on center of mass with RT3D may contribute to problems with use of this modality to track regional volumes over time. Although global 3DE volumes compare well with MRI volumes, new developments in image quality and automated software will be needed before changes in regional volumes can be reliably followed with 3DE. 2DE ESE is limited by sampling in standard imaging planes and suboptimal acquisition may cause poor matching of pre- and post-stress images. 3DE may avoid these orientation problems and contraction front mapping (CFM) may also provide a method for defining the temporal homogeneity of contraction. Chapter 8 examines the incremental benefit of 3DE CFM to 2DE and 3DE, using coronary angiography as the reference standard. It found that analysis of the temporal distribution of contraction may be an alternate means than 3DE wall motion assessment for identification of ischemia for ESE. Another significant validation needed for 3DE is not only for the LV but for the right ventricle (RV). The non-geometric nature of the RV makes it difficult to measure. Chapter 9 shows that 3DE is superior to 2DE for follow-up of RV function by validation vs. cardiac MRI. It was found that 3DE was indeed more accurate than 2DE approaches and reduces test-retest variation of RV volumes and EF measurements in follow-up RV assessment. Chapter 10 investigates the accuracy and reproducibility of the use of 3DE for serial follow-up of LV measurements. Echocardiographic follow-up of LV volumes is difficult because of the test-retest variation of 2DE. In this follow-up study, sequential 3DE measurements of change of volumes and EF were similar to those obtained using MRI, but 2DE overestimated change in EDV. As MRI is expensive, of limited availability and cannot be performed on patients with implanted devices, 3DE appears to be an effective long-term follow-up imaging tool for when sequential measurement of LV volumes is sought to guide management decisions. Chapter 11 investigates whether 3DE is more predictive of outcome than 2DE. In this outcome study, 3DE measurements of ESV and EF showed a stronger prediction of outcomes than 2DE. 3DE now appears to be the measurement of choice for when LV volumes and EF are sought to guide management decisions. In conclusion, the work reported in this thesis demonstrates that 3DE increases the accuracy of LV assessment and justifies its use in clinical decision-making based on both single and sequential measurements in routine echocardiographic practice