Vol 13, No 2 (2010)
Research paper
Published online: 2011-05-20
The effect of image translation table on diagnostic efficacy of myocardial perfusion SPECT studies
Nucl. Med. Rev 2010;13(2):64-69.
Abstract
BACKGROUND: The aim of this study was to determine which of
the most popular colour scales used in the Xeleris processing system
(GE) should preferably be used during a clinical interpretation
of myocardial perfusion images, and to find out whether a colour
scale saturation level affects the diagnostic efficacy of the study.
MATERIAL AND METHODS: From among 100 patients in whom a myocardial perfusion scintigraphy had been performed, a subgroup of people referred for coronary angiography, with neither prior history nor ECG signs of a myocardial infarction has been selected retrospectively. This group consisted of 41 patients (14 females) in the age group 46 to 76 years. All patients underwent two-day myocardial perfusion SPECT imaging using 99mTc-MIBI as a radiopharmaceutical. Reconstructed slices were interpreted in 3 colour scales: white-red-yellow-green-blue-black with computer-assigned thresholds (French 100%), the same French scale but without a white colour (image maximum set manually to a border value between red and white — French w.w.), and a white-yellow-violet- pink-blue-black scale (GEcolor), by consensus of two experienced nuclear medicine specialists. A semiquantitative method for evaluation of perfusion images was applied, based on myocardium segmentation. Perfusion in each segment was scored using a five-point system. Study interpretation (normal/ abnormal perfusion) was based on summed stress scores (SSS), being equal/above or below a given threshold value. The choice of optimal SSS threshold value was based on sensitivity and specificity of the study in detection of perfusion defects resulting from critical stenoses of main coronary arteries
RESULTS: SSS values differed among colour scales (p < 0.00001). The lowest values were obtained for a French 100% scale (mean value = 5.0, SD = 8.0), the highest for French w.w. (mean values = 8.1, SD = 8.7), and for GE colour scale — mean value — 5.6, SD — 7.9. A French 100% scale gave high sensitivity (88%), as well as specificity (83%), but only when a low SSS threshold value of 2, hardly acceptable for study interpreters, was used. When higher threshold values were applied, they compromised the sensitivity of the study. A French w.w. scale with SSS threshold values lower than 3 provided a slightly higher sensitivity (94%), but with a significant reduction in specificity (to values below 50%). Only a threshold value of 4 provided acceptable, but still low specificity (63%) with preserved high sensitivity (88%). At the same time, the scale GE colour provided indices of diagnostic efficacy with the SSS threshold value of 3 as high as a scale French 100% with threshold value of 2.
CONCLUSIONS: A French scale (Xeleris, GE) is not the scale of choice for the interpretation of myocardial perfusion SPECT images. It seems that a GE colour scale is better suited for this purpose. SSS threshold values accepted as diagnostic criteria for the detection of myocardial perfusion abnormalities should be suited separately for every translation table. The choice of optimal value should be verified by results of coronary angiography.
Nuclear Med Rev 2010; 13, 2: 64–69
MATERIAL AND METHODS: From among 100 patients in whom a myocardial perfusion scintigraphy had been performed, a subgroup of people referred for coronary angiography, with neither prior history nor ECG signs of a myocardial infarction has been selected retrospectively. This group consisted of 41 patients (14 females) in the age group 46 to 76 years. All patients underwent two-day myocardial perfusion SPECT imaging using 99mTc-MIBI as a radiopharmaceutical. Reconstructed slices were interpreted in 3 colour scales: white-red-yellow-green-blue-black with computer-assigned thresholds (French 100%), the same French scale but without a white colour (image maximum set manually to a border value between red and white — French w.w.), and a white-yellow-violet- pink-blue-black scale (GEcolor), by consensus of two experienced nuclear medicine specialists. A semiquantitative method for evaluation of perfusion images was applied, based on myocardium segmentation. Perfusion in each segment was scored using a five-point system. Study interpretation (normal/ abnormal perfusion) was based on summed stress scores (SSS), being equal/above or below a given threshold value. The choice of optimal SSS threshold value was based on sensitivity and specificity of the study in detection of perfusion defects resulting from critical stenoses of main coronary arteries
RESULTS: SSS values differed among colour scales (p < 0.00001). The lowest values were obtained for a French 100% scale (mean value = 5.0, SD = 8.0), the highest for French w.w. (mean values = 8.1, SD = 8.7), and for GE colour scale — mean value — 5.6, SD — 7.9. A French 100% scale gave high sensitivity (88%), as well as specificity (83%), but only when a low SSS threshold value of 2, hardly acceptable for study interpreters, was used. When higher threshold values were applied, they compromised the sensitivity of the study. A French w.w. scale with SSS threshold values lower than 3 provided a slightly higher sensitivity (94%), but with a significant reduction in specificity (to values below 50%). Only a threshold value of 4 provided acceptable, but still low specificity (63%) with preserved high sensitivity (88%). At the same time, the scale GE colour provided indices of diagnostic efficacy with the SSS threshold value of 3 as high as a scale French 100% with threshold value of 2.
CONCLUSIONS: A French scale (Xeleris, GE) is not the scale of choice for the interpretation of myocardial perfusion SPECT images. It seems that a GE colour scale is better suited for this purpose. SSS threshold values accepted as diagnostic criteria for the detection of myocardial perfusion abnormalities should be suited separately for every translation table. The choice of optimal value should be verified by results of coronary angiography.
Nuclear Med Rev 2010; 13, 2: 64–69
Keywords: myocardial perfusion SPECT studysemiquantitative analysissummed stress scoredisplay translation table
