Comparison of diagnostic performances of three different software packages in detecting coronary artery disease

Abstract

Quantitation was and is the main diagnostic goal of nuclear medicine. The ability to measure a process or to quantify a disease reflects the capability of a diagnostic tool to understand the intimal mechanism of physiological processes. Observing or measuring are completely different attitudes and they make the difference between modern, evidence-based medicine and the medicine of the nineteenth century based on subjectivity of physicians rather than objectivity of data.The measure of myocardial blood flow and the definition of coronary flow reserve [1] have defined a cultural background able to open new research perspectives to investigators interested in understanding the intrinsic mechanism of myocardial ischaemia and coronary artery disease.Measuring itself, however, is necessary but not sufficient to provide effective clinical information if investigators and physicians are not aware about the robustness of the quantitative data provided by different diagnostic tools.In nuclear cardiology, traditional quantitative measures can be categorized in hierarchical order according to the complexity of the measure: (1) relative (count/voxel), (2) absolute (MBq/cm 3 ) and (3) physiological (cm 3 /s per g). The pathophysiology of coronary blood flow and ischaemia has been widely evaluated by absolute measures of blood flow and coronary flow reserve by positron emission tomography (PET) imaging with a variety of tracers including 15 O, 13 NH 3 and 82 Rb and by standard nuclear medicine with 99m Tc-microspheres. Other non-nuclear methods including fractional flow reserve and coronary flow velocity reserve have been widely validated to measure the physiological severity of a stenosis which may or may not determine myocardial ischaemia in subtended vascular territories. All these methods, however, require sophisticated technology, including PET, cyclotrons for production of shortlived isotopes or expensive rubidium generators, complicated analysis programs which not always guarantee high reproducibility or, when non-nuclear methods are applied, invasive tests. This is the reason why the use of absolute, physiologically relevant approaches are limited to a few academic centres and are mainly utilized for research purposes. Therefore simpler, cheaper and more easily applicable methods have been developed for evaluating myocardial ischaemia. Although unable to measure the coronary blood flow, the relative methods allow the calculation of extent and severity of perfusion defects in comparison to a normal database. The degrees of extent and severity of perfusion defects are integrated in an adimensional score that is an index of myocardial scar and/or ischaemia. These scores may vary according to physiological determinants (amount of myocardial ischaemia) and methodological criteria including type of tracer administered, kind of stressors, algorithm for single photon emission computed tomography (SPECT) reconstruction, software for quantification, left ventricular segmentation model and reliability of the normal database. In t...