It’s not all in the numbers

Abstract

Quantitative analysis is an integral part of nuclear cardiology. It is a major strength and a key advantage that separates this technique from other noninvasive imaging methods. Nuclear cardiology is ‘‘inherently quantitative’’ in that image displays depict the number of counts detected by a gamma camera, reflecting the amount of radiotracer in the entity being imaged. Perhaps the earliest use of quantitative measurements was by Parker et al who, in contrast to a previously reported geometric method, measured background corrected Tc-labeled human serum albumin ventricular blood pool counts from electrocardiographically gated end-diastolic and endsystolic digital planar left anterior oblique images to derive a counts-based left ventricular ejection fraction (LVEF). Shortly after the introduction of thallium-201 (Tl) for myocardial perfusion imaging (MPI), efforts to apply quantitation began. As the clinical potential of MPI became recognized, methods to quantitate visual regional tracer uptake were developed. While a multiple slice profile method was considered, circumferential count profile techniques became preferred. Building on these, Garcia et al developed a comprehensive computerized space/time quantitation method. After applying proximity weighted interpolative background subtraction, for each of three standard planar views, 60 6 spaced interval radii originating from the visually determined center of the ventricle were created, with maximal ventricular wall counts along each radius plotted as a function of angular coordinates aligned in reference to the apex, normalized to the maximum value. Portions of profiles from patients with suspected disease that were[2 standard deviations (SD) below ‘‘normal’’ patient (\1% likelihood of coronary disease) means were considered abnormal. It was recognized from the outset that profile curves were representations of relative rather than absolute tracer uptake, thus limited in the ability to detect balanced flow reduction. The advent of single photon emission computer tomography (SPECT) prompted efforts to quantitate three-dimensional distribution of radiotracer, with generation of two-dimensional polar coordinate maps (i.e., ‘‘bullseye’’ displays), thereby enhancing conceptualization of defect distribution and percentages (%) of abnormal myocardium. Maximal count normalized circumferential profiles were generated from short-axis (SA) slices from the most apical to most basal cuts (analogous to the planar technique), but also from the apical profiles (60 -120 ) of the vertical long-axis (VLA) slices to be placed at the center of the plot to represent the apex, with the SA profiles mapped as increasingly larger circles toward the base. Similar to planar techniques, data from low CAD likelihood cases were used such that patient counts [2.5 SD below the mean were considered abnormal. Sophistication increased after the advent of technetium-99m (Tc) tracers in that geometric methods were modified such that sampling of SA slices used cylindrical coordinates and sampling of the apical region used spherical coordinates, ensuring perpendicular myocardial radial sampling at all points to more accurately measure tracer distribution. With availability of increased computer power, Germano et al developed an alternative algorithm that replaced circumferential profiles with true three-dimensional sampling, and with the generated sampling rays assessed counts through the entire myocardial thickness from endocardial to epicardial surface (‘‘whole myocardium sampling),’’ rather than using only maximal counts. A total perfusion defect (TPD) parameter were derived, designed to combine Reprint requests: Mark I. Travin, MD, FASNC, Division of Nuclear Medicine, Department of Radiology, Montefiore Medical Center and the Albert Einstein College of Medicine, 111 E. 210th Street, Bronx, NY 10467-2490; mtravin@attglobal.net J Nucl Cardiol 2016;23:436–41. 1071-3581/$34.00 Copyright 2015 American Society of Nuclear Cardiology.