Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 15

CARDIOLOGY, Review Article

Gated Myocardial Perfusion SPECT Imaging

Ziada, G.* Khalil, M.** and Nasr, H. *** * Department of Nuclear , Faculty of Medicine, Kuwait University, Kuwait ** Biological Imaging Centre, Cyclotron building, MRC/CSC, Cyclotron building, Hammersmith Campus, Imperial College London, UK ***NEMROCK Center, Cairo University, Egypt

Nuclear medicine has well established role Introduction in the diagnosis of many cardiovascular Radionuclide myocardial perfusion using disorders. This is performed by several single photon emission computed diagnostic procedures that are able to tomography (SPECT) is a noninvasive identify the perfusion and functional status technique used to detect and diagnose of the . The clinical examinations patients with known or suspected coronary involved are planar and tomographic gated artery disease. The technique provides blood pool, first pass, myocardial perfusion valuable information about coronary blood single photon emission computed flow in addition to extent and severity of the tomography (SPECT) performed with or diseased myocardium. In early days of without electrocardiographic (ECG) gating. , most examinations were This range of diagnostic tools has their well performed using planar with defined use in answering many questions a number of static images acquired at raised by the referring physicians. A detailed different viewing angles. During that time, description of these examinations and their Tl-201 was the radionuclide of choice due to clinical as well as technical background are its analogy to potassium kinetics. Planar outside the focus of this review. However, imaging provides little information about the the later diagnostic test has become a 3D nature of pathological lesions and thus standard in patients undergoing myocardial unable to depict the spatial distributing of perfusion imaging since it synergist the the perfusion defects. Additionally, the interpretation process by providing valuable images acquired had poor spatial contrast. information about the heart function. Thus, However, the introduction of tomographic the main objective of this review is to give examinations using single photon the reader a comprehensive overview about tomographic imaging into clinical practice the current status of gated myocardial has resulted in an improvement of many perfusion tomographic imaging and its nuclear procedures especially those methodological aspects with some emphasis concerning cardiac examinations. placed on different computational methods Myocardial SPECT has led to improved used in clinical practice. contrast resolution and lesion detectability. ــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ Correspondence author: Prof.Dr. Gaber Ziada, Ph.D. Department of Nuclear Medicine, Faculty of Medicine, Kuwait University, Kuwait E-mail: [email protected]

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 16

Further progress has been implemented by One of the remarkable advances in nuclear using multiple detector systems with cardiology has been the combined enhancement of count sensitivity and myocardial perfusion with ECG gating in reduction of imaging time. one imaging session. This has made a The last decade has witnessed a tremendous tremendous influence on the amount of change in the practice of nuclear cardiology. information that can be obtained from Technologic advances resulted in a number cardiac . This procedure enables of dedicated gamma cameras designed with the interpreting physicians to obtain additive characteristics that specifically suit cardiac diagnostic information regarding myocardial patients. The successful results obtained perfusion in combination with myocardial from prototypes have motivated the functional parameters. The later advantage introduction of these systems into the permits one to depict regional myocardial market. Some of these systems are equipped wall motion, wall thickening in addition to with x-ray Computed Tomography (CT) that measurement of end-diastolic volume is able to provide an accurate attenuation (EDV), end-systolic volume (ESV) and correction and lesion localization. CT hence calculation of LV systems are also able to provide information (EF). about coronary and when combined with perfusion images can result Why gated SPECT is clinically useful? in substantial diagnostic information in one The rapid growth and diffusion of gated imaging session. SPECT imaging was due to a number of Another area of research being conducted advantages the technique has provided to the and have been implemented in commercial interpretation of myocardial perfusion software programs are the inclusion of images. One of these is the added diagnostic resolution recovery in the reconstruction and prognostic values that have been proved algorithms so that a significant reduction of by realizing its potential implications in image noise can be achieved. Applying these patient management [4]. Various studies have corrections to data acquired with less reported the benefits and advantages that can acquisitions time, and hence high noise be obtained from GSPECT imaging: levels, showed an adequate and comparable Gated myocardial perfusion SPECT imaging results to those acquired in longer provides an improvement for myocardial acquisition times. perfusion imaging in increasing the One more point is the development of diagnostic accuracy of detecting patients semiconductor gamma camera dedicated for with coronary artery disease [5]. It enhances imaging cardiac patients. This commercial the detection of multivessel disease and introduction of such systems came as a permits the assessment of the hemodynamic result of using a room temperature, high significance of angiographically detected density, semiconductor materials such as coronary stenotic lesions. Cardiac imaging Cadmium Telluride (CdTe) and Cadmium as a gatekeeper to cardiac catheterization is zinc telluride (CZT). Performance encouraged in management of patients with measurements yielded an improved spatial chronic stable angina in order to minimize resolution, sensitivity, count rate the rate of normal catheterizations and to performance, and energy resolution. For enrich the angiographic population with a reviews on recent advances in SPECT and greater proportion of patients with nuclear cardiac instrumentation, the reader is significant, yet treatable disease [6]. referred to [1,2,3]. Furthermore, the technique provides reliable

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 17

and reproducible measures of left ventricular patients with symptoms of stable angina, functional parameters. These quantitative with no implications as to patient outcome approaches decrease inter- and intra- [11]. observer variability and facilitate serial assessment of myocardial perfusion and How Gated SPECT is performed? function [7]. Acquisition Risk stratification based on gated SPECT The conventional way of acquiring imaging has been shown to be a reliable and myocardial perfusion study is to take a accurate tool in patient prognosis. In a study multiple 2D images around the patient. A by Sharir et al. [8], post stress EF and the camera rotates around the patient for at least extent of stress induced ischemia provide 180 degree, stopping at regular number of complementary information in the prediction angular intervals. The angular interval is of risk of cardiac death. EF was the strongest determined by the number of stops and the predictor of mortality, whereas the perfusion total acquisition arc. For each stop, the summed difference score was the strongest camera obtains a snap shot image of predictor of Myocardial infarction [8]. The myocardial activity distribution. To obtain combined variables were more effective in functional information, a 3 lead ECG-trigger risk stratification of patients than were the is attached to the patient chest to input a stress nuclear variables or post stress EF QRS signal to the acquisition computer. The alone [9]. An interesting therapeutic R-R interval is utilized to obtain a number of application of GSEPCT imaging is the gates or intervals (i.e. 8 or 16) for each detection of response to therapy, cardiac cycle, figure (1). One gate revascularizations, and treatment outcome. represents a volumetric status during the Whenever moderate to severe ischemia is periodic systolic-diastolic cycle. The study found by nuclear testing in patients with acquisition may be time, count, or trigger previous revascularization, consideration dependent. All are nearly equivalent as long should be given to repeat catheterization 10. as an adequate image statistics is achieved Cost-effectiveness analyses further indicate [12]. a role for myocardial perfusion imaging in reducing the rate of revascularization in

Figure (1). The R-R interval efficient in collecting better counts statistics While a single head gamma camera is in addition to a significant reduction of technically able to image and acquire a imaging time. This is of particular simultaneous myocardial perfusion importance in busy nuclear cardiac units and combined with ECG gating, the use of influences significantly on scanner multi-detector system would be more throughput. A short scanning time also helps to avoid patient motion and provides patient comfort. Acquisition parameters with

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 18

matrix size can be either of 64x64 or individual preferences. Factors that 128x128. The former is frequently used as it influence signal to noise ratio are imaging preserves large counts per pixel in a given time, injected dose, variation in patient patient study. Because myocardial perfusion sizes,. etc. SPECT can be performed on the same day The selection of the cut-off frequency was or on a two separate days, gated SPECT can found to influence the calculations of the left be applied on either case. However, in the ventricular volumes as well as EF. This has same day protocol, gated SEPCT data are been demonstrated by our group and others preferred to be with the study of the larger and will be discussed later. injected dose. This is to improve count Data reconstruction yield transverse cardiac sampling statistics and also important for slices for each time frame and image reliable measure of functional indices. Low reorientation is performed in the three energy high resolution collimator is orthogonal planes across the myocardial axis frequently used in the literature, however, to produce the standard horizontal, vertical general purpose collimator is a possible and short axis slices. Image reconstruction alternative since it improve the statistical for the unsummed or gated data produce a quality of the scan. set of images for each cardiac cycle that shows the ventricle kinetics in that particular Processing phase. Calculations of the LV function The raw data are received in the form of parameters are then carried out by use one of unsummed 8 or 16 gated bins for each the commercially available quantitative projection angle. To obtain data regarding methods myocardial perfusion, the cardiac frames are summed together in a composite manner to Gated SPECT Methods reveal information as if they would have Many attempts based on several and been acquired in a non-gated acquisition different theoretical assumptions were given mode. Data reconstruction can be executed to calculate the left ventricular volumes and using analytic approaches such as Filtered EF Gated SEPCT methods can generally be back projection (FBP) or iterative divided into three main categories: reconstructions such as Maximum likelihood geometric based, count-based, or hybrid expectation maximization (ML-EM) or methods. In geometric based methods, the Ordered Subset- Expectation Maximization left ventricle is modeled as ellipsoidal, (OSEM). Low pass filtering such as prolate spheroid, or spherical-cylindrical Butterworth and Hanning filter are approach. In count based methods, the count commonly used. Because the cut-off change among and within cardiac frames frequency is an important parameter that throughout the cardiac systolic functions is determine the final smoothness or sharpness utilized in the estimation process. Hybrid of the images, it is not recommended to methods combine between both approaches change the cut-off frequency on a blindly and utilized in a number of commercialized manner. The cut-off frequency can alter the software programs interpretation results of myocardial perfusion as well as myocardial function. Factors facilitated the introduction of Thus, once the values has been set, taking GSPECT into consideration factors that affect image Thallium-201 was the common radionuclide signal-to-noise ratio, then it is not advised to in cardiac imaging and has been used for change the cut-off values according to many years in assessing patients with

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 19 coronary artery diseases. However, the images. Because the activity distribution introduction of Tc-labelled compounds such along a radial profile drawn normally on the as tetrofosmin and sestamibi has resulted in myocardial wall is more likely to be a higher injection dose, better counts asymmetric in shape, an asymmetric statistics, and longer time is permitted before Gaussian function is employed to fit these isotope injection and launching image profiles using pre-defined standard acquisition. Gated SPECT imaging also deviations. These SDs were calculated by requires a fast processing speed and high phantom studies and was found to be 65% of storage capacity. Further, owing to the fact the profile width. QGS uses the count that the R-R interval is divided into frames; density change among and within frames to data sampling regime requires adequate calculate the myocardial thickening and count sensitivity. The former has been thickness utilizing the law of the tackled by using fast computer processors conservation of mass [14]. Figure 2 shows a and memory chips. The later requirement screenshot of QGS output display. has been solved by using a detection system of multiple detector heads, larger injected Emory Cardiac Toolbox (ECTb) Method dose, or relatively prolonged acquisition Emory Cardiac Toolbox was reported in time. 1999 by Faber et al [15]. The method uses two geometric systems for sampling counts These features contributed in such a way along the myocardium, Figure (3). It that made GSPECT imaging a clinically employs cylindrical geometry to sample the feasible. Manufacturers nowadays are mid and basal part of the myocardium but providing gamma cameras with two or more spherical one is used to sample the apical detectors showing better count sensitivity portion. ECTb method uses the partial and reduced statistical uncertainty. The volume effect to calculate the myocardial striking point was the development of fully thickening using Fourier transform16.By automated algorithms able to reconstruct and assuming the myocardial thickness to be 10 calculate the LV function in less than a mm at end-diastolic phase, the endocardial minute. We will present some of the and epicardial offsets are then estimated for methods proposed to evaluate the LV all the myocardial frames. Figure 4 shows function along with their theoretical an end result of myocardial perfusion study assumptions processed by ECTb

Quantitative Gated SPECT (QGS) 4D-MSPECT Method 4D-MSPECT was developed by Ficaro et al QGS was developed by Germano et al in (17) at the University of Michigan, USA. A 1995 [13]. The method is fully automated semi-automated 4D-MSPECT software and works in 3D dimensional space to processes data on the basis of a 2- sample the myocardium using ellipsoidal dimensional gradient image from which the model. The method uses fundamental image initial estimates of the ventricle are made. A processing tools such as thresholding, series of 1- and 2-dimensional weighted binarization, and data clustering to segment splines are used to refine the endocardial and and isolate the left ventricle from the nearby epicardial surface estimates. The 4D- structures. Minimum volume was set to 50 MSPECT model uses also a cylindric– ml to be as a threshold above which the spheric coordinate system. The cylindric algorithm searches the entire short axis coordinates are used to sample from the

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 20 basal wall to the distal wall and spheric a good correlation with magnetic resonance coordinates are used to sample the apex (18). imaging in the calculation of LV volumes By using Gaussian functions, wall position and EF [19]. An example of 4D MSPECT is and thickness are estimated. Like QGS and illustrated in figure 5 ECTb, 4D-MSPECT has the capability for manual processing when the automatic module fails to accurately delineate the myocardial boundaries. The method showed

Figure (2). QGS output display. It shows a 3D dimensional grid for the LV from many directions (at ED and ES) and thus allows the reader to depict regional myocardial dysfunction

Figure (3). ECTb uses two geometric systems to define the whole myocardium. The mid and basal portion of the ventricle is modeled as cylindrical in shape while the apical portion is modeled as a sphere. The valve is defined as two intersecting planes to obtain the most basal slice of the ventricle.

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 21

Figure 4. ECTb output shows the left ventricles volumes and EF for a patient stress and rest scan. Samples short, horizontal, and vertical slices are displayed in addition to polar map of the LV wall thickness.

Figure (5). 4D-MSPECT shows the three orthogonal slices of myocardial perfusion study, together with the perfusion polar maps.

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 22

Layer of Maximum Count Method cavity. This latter constant was shown to be One of the limiting factors that constrain the in a narrow physiological range [21]. By performance of quantitative gated SPECT plotting the ejection fraction calculated by a methods is data acquired for patients with reference method versus the EFmax while small left ventricles. In an attempt to solve setting the intercept to zero, the regression this problem, a relatively recent method was slope (1+constant) could be obtained, which developed to correct for such an artifact is also applicable for patients with small 20,21 This method is known as layer of . Khalil et al has recently evaluate the maximum count method (LMC). It employs performance of LMC in patients with small the prolate spheroidal transform to transform left ventricles and showed a superior the spatial activity distribution of the left performance over other commercially ventricle in the prolate spheroid system. available methods [22]. Moreover, the same Three angles are defined: A horizontal axis group has recently assessed the method that lies along the transmural direction of the versus 2D and left ventricular wall, a circumferential angle demonstrated that the method can provide a that ranges from 0 to 2π, and an azimuthal reliable estimate of the EF with comparable angle that ranges from 0 to π [21]. In each variability with other well-validated short-axis image, maximum count sampling methods23. Drawback of this method is its is performed along the horizontal direction dependency on other reference and accurate at fixed azimuthal and circumferential methods in the calibration process. angles. A threshold of 50% is used to truncate the ellipsoid in order to fit the Left Ventricular Global Thickening ventricular geometry. A tri-linear Ejection Fraction (LVGTF) interpolation is used in data transformation. This method depends on the detection of After data transformation, the left ventricle myocardial wall thickening during systolic appears as an approximated plane surface, contraction. The method relies on the partial and the layer of the maximum counts (a volume effect, a phenomenon that makes the surface area of maximum counts) is myocardial to waxes and wanes during heart determined to calculate the enclosed contraction. In these circumstances, the pixel counts in end diastolic and end systolic volume, Vmax. The Vmax of the end and end is calculated to compute the images are used to quantify the myocardial ejection fraction of the layer of maximum thickening without edge detection or geometric measurements24. However, the counts, EFmax= (EDmax–ESmax)/EDmax. EFmax is then modified by the formula: method uses the systolic and diastolic counts in addition to geometric assumptions to EF=EFmax (1+constant) to compute the left ventricular ejection fraction. The constant in derive a regional thickening fraction and the formula is a multiplication of two hence to calculate the left ventricular constants, one related to the imaging system ejection fraction using the formula: parameters like data acquisition, 3 / 2 reconstruction algorithm, smoothing kernel ⎛ 1 ⎞ LVEF = 1− ⎜ ⎟ etc., and the other is a ratio of the ⎝1+ f ⎠ myocardial volume and the ventricular thickening throughout all left ventricular regions and also an excellent correlation Where f is the measured index of myocardial with gated blood pool (r=0.91, slope=0.97). thickening. The method showed lower Furthermore, ejection fraction measurement variability in measuring the regional

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 23 by this method is not dependent on the filter number of counts collected. Because the cut-off frequency used (24, 25). Due to noise in nuclear images is stochastic in many simplifying assumptions postulated by nature, a large number of counts are LVGTF method, it showed inferior results recommended to reduce the statistical noise than other quantitative gated SPECT associated with radioactive decay. The methods 26] emission rate of Tl-201 tracer is significantly lower than that of Tc-99m such Factors affecting gated SPECT imaging that considerable differences can be seen in The addition of ECG-gating to myocardial Tl-201 based images rather than those perfusion imaging added some variables to obtained from Tc-99m based tracers. The the conventional myocardial perfusion introduction of Tc-99m based tracers imaging. Therefore, once gated SPECT (Sestamibi and Tetrofosmin) has enhanced imaging has been introduced into the clinic the quality of myocardial perfusion SPECT many researchers and practitioners were as well as the ECG-gated studies. Some interested to identify the sources of imaging investigators compared the functional errors and its consequences on the functional parameters obtained by Tl-201 and Tc-99m parameters. Several research articles were gated SPECT. EF measurements obtained by released discussing many factors that Tc-99m sestamibi have shown better influence the quality of GSPECT imaging reproducibility (5.5%) than that by Tl-201 and the resulting functional indices. On the (10.3%) (28). The low energy of Tl-201 acquisition level, proper selection of the produces a high amount of scatter in the acquisition parameters in addition to acquired images, which decreases the adequate injection dose serve to provide a certainty in edge identification. The low sufficient count sampling for image energy photons produce substantial scatter reconstruction. Patient preparation, counts and hence Tl-201 studies are blurred instructions, and comfort are also required to with less sharpening. minimize motion artifacts and thus able to The gated images are characterized by low reduce the likelihood of repeating the study count statistics which in turn may influence or generating imaging artifacts. Patients with the performance of the quantitative arrhythmia particularly those with atrial algorithms in estimation task. However, fibrillation undergoe ECG disturbances and when the acquired data are statistically hence gated acquisition may fail in such adequate, a better confidence in outlining the group of patients. Not only the functional boundaries is achieved. Count statistics are paratmers are affected but also the perfusion influenced by many factors such as injected ones. Sciagra et al found that the dose, acquisition time, and time after simultaneous acquisition of gated and injection, energy window (15% vs. 20%), nongated myocardial perfusion SPECT (a system sensitivity, patient size, collimation, feature provided by some imaging systems) number of frames set to the cardiac cycle in patients with atrial fibrillation have a The effect of background activity was technical importance in removing errors that studied and revealed a significant impact on may come from the summation process 27 volume and EF determination. Achtert et al reported that higher extracrdiac activity may Acquisition interfere with the delineation of the cardiac Tl-201 versus Tc-99m based gated borders and thereby resulted in false myocardial SPECT determination of the volumes and EF [29], An important factor that is always concern figure 6. Similarly, the injected dose and the the quality of scintigraphic images is the

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 24

imaging time are important prerequisites for along with spatial resolution. One study has controlling the count levels in the gated compared the cardiac confocal collimator study. Prolonged time after injection or low (which maintains the spatial resolution and injected dose decreases the cardiac activity provides a significant improvement in count and hence might reduce the performance of sensitivity) and high resolution collimator myocardial contouring by the mathematical for the calculation of the EF. Good algorithm. correlation was found between both collimators; however the mean EF by Image zooming Confocal collimator was significantly higher Acquisition zoom is recommended during than that by high resolution collimator and cardiac SPECT acquisition; however an Gated Blood Pool (GBP) (62%±10% vs. optimal value may be from 1.3-1.6 to 54% ±13% vs. 55.4% ±12%). Two common improve the magnification of the left types of orbits are routinely used in cardiac ventricle. Too large zooming is not SEPCT imaging. These are circular "non- recommended to avoid truncation of the elliptical" or noncircular "elliptical" orbits. heart during image acquisition. Moreover, The former provides homogenous image large zoom increases the number of pixels resolution along all projections acquired, for an entire left ventricle, the matter which given that the heart-detector distance is reduces the counts per pixel and results in approximately constant. On the other higher noise level (30). extreme, elliptical orbit minimizes the

Collimators and type of rotation orbits patient-detector distance to get the best Collimators play an important role in possible resolution. However, this may outlining the boundaries of different impact adversely as resolution may differ structures in the scintigraphic images. Also, considerably from projection to another the type of collimator determines the producing image artifacts. geometric sensitivity of the imaging system

Figure 6. Inaccurate automatic contouring. The user should review the raw data as well as the reconstructed images for any possible failure in outlining the myocardial walls. [from Hermes installation documents, QGS version 4.0]

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 25

Mean EF (%). 100. A

80. Butterworth Filter 60. QGS 40. ECTb 20.

0. 0.1 0.2 0.3 0.4 0.5 0.6 Cut-off frequency (Cycle/cm)

Mean EF (%). B 100.

80. Hamming Filte 60. QGS 40. ECTb

20.

0. 0.2 0.4 0.6 0.8 1.0 1.2 Cut-off frequency (Cycle/cm) Figure (7). The response of ECTb and QGS to wide range of filter cut-off frequencies. The relatively higher EFs showed by both methods in comparison to gated blood pool might be attributed to inclusion of small LV in the patient population.

Processing Reconstruction algorithms FBP is the conventional reconstruction reconstructed images. We have evaluated technique for SPECT images. The QGS and ECTb using two reconstruction acceptable alternative to FBP is iterative filters namely hamming and the commonly reconstruction which could incorporate used Butterworth filter [33]. Both filters many physical degrading factors. Ordered revealed similar results such that low cut-off subset expectation maximization algorithm values tend to reduce the LV volumes and was found to reveal better results than FBP overestimate the EF. However, boosting the when resolution recovery is included in filter cut-off appeared to stabilize the reconstruction. Moreover, a substantial measurements where the volume nad EF improvement was obtained when OSEM curved showed a constant response a long a reconstruction was associated with Fourier reasonable dynamic range that is used in filtering in gated myocardial SPECT using clinical setting. [34]. Low cut-off frequencies Tl-201 and Tc-99m as myocardial were found to smooth the images to a degree radiotracers [31,32] that force the quantitative algorithm to underestimate the LV volumes and Filter cut-off frequency overestimating the ejection fraction. The cut-off frequency in SEPCT However, higher cut-off frequencies serve to reconstruction determines the degree of improve the measurements but further smoothing and hence the noise level in the increase may lead to introduction of noise

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 26 and confounding the computation task. cm at end-diastolic phase [15]. In systems Therefore, the optimal cut-off was investi- with poor resolution, the reconstructed data gated to be a range of values that lie may violate this assumption leading to false between high and low frequencies, Figure results. [7] QGS method depends on count profiles drawn normally on the myocardial walls, Scatter and attenuation correction and the endocardial and epicardial offsets One of the physical factors that degrade are determined by a given percentages of the SPECT images is the contribution of the profiles (65% of the Gaussian fitting scattered photons. It falsely represents the profiles). Systems with lower resolution may actual position of the emission site. underestimate the volumes and overestimate Myocardial edges are therefore blurred and the EF. One study investigated that the left the ventricular cavity decreases in size. This ventricular volume was underestimated by might partially explain volumes up to 13% as the SPECT spatial resolution underestimation noted by gated SPECT in deteriorates [36]. Software programs that comparison to other modalities34. By depend on count density for estimating the accounting for the scattered photons in the ventricular volumes may be less affected by image acquisition, a recent report indicated system resolution; however, the count an improvement in volumes determinations statistics should be technically maintained. by QGS software in phantom and patient studies (34). Collimators A Preliminary study to assess the influence The type of collimator determines the of attenuation correction on gated geometric sensitivity of the imaging system myocardial images in comparison to data not along with spatial resolution. One study has corrected has shown that the EF is compared the cardiac confocal collimator underestimated, on average, by 2.9% due to (which maintains the spatial resolution and the use of attenuation correction. However, provides a significant improvement in count the same study suggested that the functional sensitivity) and high resolution collimator estimates are still to be determined for the calculation of the EF. Good accurately without use of attenuation correlation was found between both correction, as quantification algorithms were collimators. However the mean EF by designed on data set obtained without confocal collimator was significantly higher attenuation correction and validation studies than that by high resolution collimator and as well (35). Gated Blood Pool (62%±10% vs. 54% ±13% vs. 55.4% ±12%)37 Patients Small left ventricles Perfusion defect Gated SPECT processing methods In animal model, a trend toward volume particularly those that depend on edge overestimation was observed in presence of detection is very sensitive to the resolution perfusion defects, however, the ventricular of the myocardial walls. Poor resolution size of animals is questionable to represent enlarges the myocardial thickness and thus that by human (38). Another study included minimizes the size of ventricular cavity. patients with large myocardial infarction and This problem is more obvious in patients calculated the volumes and EF by two with small hearts. For instance, in ECTb the algorithms (QGS and MuliDim) using Tl- algorithm assumes the wall thickness to be 1

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 27

201 as a myocardial tracer revealed that 5. Sciagra R, Leoncini M. Gated single photn QGS underestimated the EDV and EF but emission computed tomography. The MultiDim overestimated the EDV and ESV. present day “one-stop-shop” for cardiac A tendency toward overestimating the imaging QJNM ventricular volumes was observed with an 6. Gibbons RJ, Abrams J, Chatterjee K, increase in infarction size [39]. Daley J, Deedwania PC, et al. ACC/AHA 2002 guideline update for the management Summary of patients with chronic stable angina-- Gated myocardial perfusion SPECT is a summary article: a report of the American useful imaging technique for patients with College of Cardiology/American Heart known or suspected coronary artery disease. It Association Task Force on practice provides additive diagnostic and prognostic guidelines (Committee on the information for the reading as well the Management of Patients With Chronic referring physicians. However, the Stable Angina). J Am Coll Cardiol. Jan 1; technicalities behind the procedure need 41(1):159-68, 2003. proper adjustments and careful selection for 7. Sharir T, Germano G, Waechter PB, et al. the different acquisition and processing A new algorithm for the quantitation of parameters. Ignoring or less recognition of myocardial perfusion SPECT, II: these parameters could lead to error validation and diagnostic yield. J Nucl propagation and unreliable quantitative Med.; 41:720-7, 2000. results. In this instance, It is not only the 8. Sharir T, Germano G, Kang X, et al. functional parameters that will be affected, Prediction of myocardial infarction versus but also the perfusion parameters can be cardiac death by gated myocardial confounded. Further studies are needed to perfusion SPECT: risk stratification by the assess the performance of hybrid imaging amount of stress-induced ischemia and the with regard to estimation of LV EF in absence post-stress ejection fraction. J Nucl Med.; and presence of CT attenuation correction. 42:831-7, 2001. 9. Johnson LL, Verdesca SA, Aude WY, et References al. Postischemic stunning can affect left 1. Madsen M. Recent Advacnes in SPECT ventricular ejection fraction and regional imaging. J Nucl Med. Apr; 48(4):661-73, wall motion on post stress gated sestamibi 2007. tomograms. J Am Coll Cardiol.; 30:1641- 2. Patton JA, Slomka PJ, Germano G, Berman 8, 1997. DS. Recent technologic advances in 10. Papaioannou G.I. and Heller G.V. Risk nuclear cardiology. J Nucl Cardiol. Jul; assessment by myocardial perfusion 14(4):501-13, 2007. imaging for coronary revascularization, 3. Slomka PJ, Patton JA, Berman DS, medical therapy, and non cardiac surgery. Germano G. Advances in technical aspects Cardiol Rev.; 11(2):60-72, 2003. of myocardial perfusion SPECT imaging. J 11. Shaw L.J., Hachamovitch R., Berman Nucl Cardiol. 2009 Mar-Apr; 16 (2):255- D.S., et al. The economic consequences of 76. Epub Feb 26. Review 2009. available diagnostic and prognostic 4. Shaw LJ, Iskandrian AE. Prognostic value strategies for the evaluation of stable of gated myocardial perfusion SPECT. J angina patients: an observational Nucl Cardiol; 11:171–85, 2004. assessment of the value of precatheterization ischemia. Economics of Noninvasive Diagnosis (END)

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 28

Multicenter Study Group. J Am Coll 19. Schaefer WM, Lipke CS, Standke D, Cardiol.; 33:661-9, 1999. Kuhl HP, Nowak B, Kaiser HJ, Koch KC, Buell U. Quantification of Left 12. Paul AK, Nabi HA. Gated Myocardial Ventricular Volumes and Ejection Perfusion SPECT: Basic Principles, Fraction from Gated 99mTc-MIBI Technical Aspects, and Clinical SPECT: MRI Validation and Comparison Applications J. Nucl Med Technol.; 32: of the Emory Cardiac Tool Box with 179-187, 2004. QGS and 4D-MSPECT J. Nucl Med.; 46: 13. Germano G, Kiat H, Kavanagh PB, 1256 - 1263, 2005. Moriel M, Mazzanti M, Su HT, Van 20. Feng B, Sitek A, Gulberg GT. The Train KF, Berman DS. Automatic prolate spheroidal transform for gated quantification of ejection fraction from SPECT. IEEE Trans Nucl Sci; 48:872– gated myocardial perfusion SPECT. J 875, 2001. Nucl Med; 36:2138-2147, 1995. 21. B, Sitek A, Gullberg GT. Calculation of 14. Germano G, Erel J, Lewin H, Kavanagh the left ventricular ejection fraction PB, Berman DS. Automatic quantitation without edge detection: application to of regional myocardial wall motion and small hearts. J Nucl Med; 43:786–794, thickening from gated technetium-99m 2002. sestamibi myocardial perfusion single- 22. Khalil MM, Elgazzar A, Khalil W, Omar photon emission computed tomography. J A, Ziada G. Assessment of left Am Coll Cardiol.; 30:1360-7,1997. ventricular ejection fraction by four 15. Faber TL, Cooke CD, Folks RD, Vansant different methods using 99mTc JP, Nichols KJ, DePuey EG, Pettigrew tetrofosmin gated SPECT in patients with RI, Garcia EV. Left ventricular function small hearts: correlation with gated blood and perfusion from gated perfusion pool. Nucl Med Commun; 26:885-93, images: an integrated method. J Nucl 2005. Med.; 40:650-659, 1999. 23. Khalil MM, Attia A, Ali M, Ziada G, 16. Cooke CD, Garcia EV, Cullom SJ, Faber Omar A, Elgazzar A. Echocardiographic TL, Pettigrew RI. Determining the validation of the layer of maximum count accuracy of calculating systolic wall method in the estimation of the left thickening using a fast Fourier transform ventricular EF using gated myocardial approximation: a simulation study based perfusion SPECT: correlation with QGS, on canine and patient data. J Nucl Med.; ECTb, and LVGTF Nucl Med Commun. 35:1185-92, 1994. Aug; 30 (8):622-8, 2009. 17. Ficaro EP, Quaife RA, Kritzman JN, 24. Smith WH, Kastner RJ, Calnon DA, Corbett JR. Accuracy and reproducibility Segalla, D, Beller GA, Watson DD. of 3D-MSPECT for estimating left Quantitative gated SPECT imaging: a ventricular ejection fraction in patients counts-based method for display and with severe perfusion abnormalities measurement of regional and global [abstract]. Circulation.; 100(suppl):I26, ventricular systolic function. J Nucl 1999. Cardiol.; 5:451-463, 1997. 18. Kritzman JN, Ficaro EP, Corbett JR. 25. Calnon DA, Kastner RJ, Smith WH, Reproducibility of 3-D MSPECT for Segalla D, Beller GA, Watson DD. quantitative gated SPECT sestamibi Validation of a new counts-based gated perfusion analysis [abstract]. J Nucl single photon emission computed Med.; 41(suppl):166P, 2000. tomography method for quantifying left

Egyptian J. Nucl. Med., Vol 2, No. 2, Dec. 2009 29

ventricular systolic function: comparison Results from Sestamibi and 201Tl 16- with equilibrium radionuclide Interval Gated SPECT. J Nucl. Med.; 46: angiography. J Nucl Cardiol; 4:464-471, 1789-1795, 2005. 1997. 33. Khalil MM, El-Motairy TN, Shehab F, 26. Khalil MM, Elgazzar A, Khalil W. Omar AM, Mohammed AM, Elgazzar Evaluation of left ventricular ejection AH. The effect of filter cut-off on the fraction by the quantitative algorithms reproducibility of QGS and ECTb in left QGS, ECTb, LMC and LVGTF using ventricular function estimation: gated myocardial perfusion SPECT: comparison with gated blood pool investigation of relative accuracy. Nucl [abstract]. J Nucl Med.; 45:225P, 2004. Med Commun. 2006 Apr; 27(4):321-32. 34. Manrique A, Hitzel A, Vera P. Impact of Erratum in: Nucl Med Commun. Oct; photon energy recovery on the 27(10):831, 2006. assessment of left ventricular volume 27. Sciagra DR, Sotgia B, Boni N, Pupi A. using myocardial perfusion SPECT. J Assessment of the Influence of Atrial Nucl Cardiol.; 11:312-7, 2004. Fibrillation on Gated SPECT Perfusion 35. Ficaro EP, Kritzman JN, Hamilton TW, Data by Comparison with Simultaneously Mitchell TA, Corbett JR. Effect on Acquired Nongated SPECT. J Nucl Med; attenuation corrected myocardial 49:1283-1287, 2008. perfusion SPECT on left ventricular 28. Hyun IY, MD, Kwan J, Park KS, Lee ejection fraction estimates [abstract]. J WH. Reproducibility of Tl-201 and Tc- Nucl Med; 41:166P, 2000. 99m sestamibi gated myocardial 36. Takayama T, Motomura N. The perfusion SPECT measurement of compensation method for the effects of myocardial function. J Nucl Cardiol.; system resolution to the left ventricular 8:182-7, 2001. volume measurements using QGS 29. Achtert AD, King MA, Dahlberg ST, program: Nuclear Science Symposium, Pretorius PH, LaCroix KJ, Tsui BM. An IEEE,; 3:1928-32,1998. investigation of the estimation of ejection 37. Groch MW, Takamiya Y, Groch PJ, fractions and cardiac volumes by a Erwin WD. Quantitative gated quantitative gated SPECT software myocardial SPECT: effect of collimation package in simulated gated SPECT on left-ventricular ejection fraction. J images. J Nucl Cardiol.; 5:144-52, 1998. Nucl Med Technol. Mar; 28 (1): 36-40, 30. Berman DS and Germano G, Editors. 2000. Clinical gated cardiac SPECT. 1st 38. Vallejo E, Dione DP, Bruni WL, Edition, Futura Publishing Company, Constable RT, Borek PP, Soares JP, Carr Inc., Armonk, N.Y; 1999. JG, Condos SG, Wackers FJ, Sinusas AJ. 31. Gremillet E, Champailler A, Soler C. Reproducibility and accuracy of gated Fourier temporal interpolation improves SPECT for determination of left electrocardiograph-gated myocardial ventricular volumes and ejection fraction: perfusion SPECT. J Nucl Med.; 46:1769- experimental validation using MRI. J 74, 2005. Nucl Med.; 41:874-82,2000. 32. Marie PV, Djaballah W, Franken PR, 39. Vera P, Koning R, Cribier A, Manrique Vanhove C, Muller MA, Boutley H, et al. A. Comparison of two three-dimensional OSEM Reconstruction, Associated with gated SPECT methods with thallium in Temporal Fourier and Depth-Dependant patients with large myocardial infarction. Resolution Recovery Filtering, Enhances J Nucl Cardiol.; 7:312-9,2000.