Knesaurek K, Machac J
Division of Nuclear Medicine, Mount Sinai Medical Center, New York, USA.
Med Phys. 1999 Jun;26(6):917-23. doi: 10.1118/1.598483.
Collimated F-18 FDG SPECT imaging has been shown to be an acceptable alternative to F-18 FDG PET imaging for evaluating injured but viable myocardium. Ultra-energy (UHE) imaging is usually performed in simultaneous F-18 FDG/Tc-99m MIBI studies. The main limitations of this technique are degradation of the Tc-99m MIBI images due to F-18 downscatter to the Tc-99m window, and loss of resolution in Tc-99m images caused by using a UHE rather than a low-energy collimator. The quality of F-18 images has not been addressed. In our clinical and phantom studies we have found that F-18 images are inferior to simultaneously acquired Tc-99m MIBI images. This paper compares two correction methods for F-18 FDG images in a realistic cardiac phantom study. One approach is subtractive scatter correction, which employs a third 410 keV energy window image to estimate scatter. The other approach is based on restoration. The phantom acquisition was performed with 7.2 MBq of F-18 and 22.2 MBq of Tc-99m injected into the left ventricular (LV) wall. Three inserts, 3 cm, 2 cm, and 1 cm in diameter, were placed in the LV wall to simulate infarcts. Circumferential profiles were drawn from three successive short-axis slices and compared with true phantom data. The differences were calculated as root-mean-square error (rmse). Scatter correction improved rmse only 4.5 +/- 0.3%, while restoration improved rmse 16.1 +/- 0.4%, when compared with raw data. The same differences, measured as rmse, were 9.5 +/- 0.5, 6.8 +/- 0.4, and 5.1 +/- 0.5 for raw, scatter corrected, and restored F-18 data, respectively, when compared with Tc-99m window 140 keV data. The amount of noise, measured as root-mean-square % (rms%) was 5.3 +/- 0.5% for the Tc-99m image, 4.9 +/- 0.7% for the F-18 restored image, 6.2 +/- 0.6% for the raw F-18 image, and 6.5 +/- 0.9% for the scatter corrected F-18 image. The contrast measured for 2 cm and 3 cm inserts was 0.17 +/- 0.07 and 0.26 +/- 0.06 for F-18 raw data, 0.19 +/- 0.08 and 0.29 +/- 0.06 for the scatter corrected F-18 image, and 0.28 +/- 0.06 and 0.43 +/- 0.07 for the restored F-18 image. The contrast was 0.20 +/- 0.07 and 0.46 +/- 0.05 for the Tc-99m 140 keV window image. The restoration approach provided F-18 images of better contrast and detectibility than uncorrected or scatter corrected F-18 images. Restored F-18 images match better with the simultaneously acquired Tc-99m images.
准直的F - 18 FDG SPECT成像已被证明是评估受损但仍存活心肌的F - 18 FDG PET成像的一种可接受替代方法。超能量（UHE）成像通常在同时进行的F - 18 FDG/Tc - 99m MIBI研究中进行。该技术的主要局限性在于，由于F - 18向下散射到Tc - 99m能窗，导致Tc - 99m MIBI图像质量下降，以及使用超能量而非低能准直器导致Tc - 99m图像分辨率降低。F - 18图像的质量尚未得到解决。在我们的临床和体模研究中，我们发现F - 18图像不如同时采集的Tc - 99m MIBI图像。本文在一项真实心脏体模研究中比较了两种F - 18 FDG图像校正方法。一种方法是减法散射校正，它使用第三个410 keV能量窗图像来估计散射。另一种方法基于恢复。体模采集是通过向左心室（LV）壁注射7.2 MBq的F - 18和22.2 MBq的Tc - 99m进行的。在LV壁中放置三个直径分别为3 cm、2 cm和1 cm的插入物以模拟梗死灶。从三个连续的短轴切片绘制圆周轮廓，并与真实体模数据进行比较。差异以均方根误差（rmse）计算。与原始数据相比，散射校正仅使rmse改善了4.5±0.3%，而恢复使rmse改善了16.1±0.4%。与Tc - 99m 140 keV能窗数据相比时，原始、散射校正和恢复后的F - 18数据以rmse衡量的相同差异分别为9.5±0.5、6.8±0.4和5.1±0.5。以均方根百分比（rms%）衡量的噪声量，Tc - 99m图像为5.3±0.5%，F - 18恢复图像为4.9±0.7%，原始F - 18图像为6.2±0.6%，散射校正后的F - 18图像为6.5±0.9%。对于2 cm和3 cm插入物，F - 18原始数据的对比度分别为0.17±0.07和0.26±0.06，散射校正后的F - 18图像为0.19±0.08和0.29±0.06，恢复后的F - 18图像为0.28±0.06和0.43±0.07。Tc - 99m 140 keV能窗图像的对比度为0.20±0.07和0.46±0.05。恢复方法提供的F - 18图像比未校正或散射校正的F - 18图像具有更好的对比度和可检测性。恢复后的F - 18图像与同时采集的Tc - 99m图像匹配得更好。
