基于骨SPECT的 segmented 衰减校正(B-SAC)用于骨转移定量分析:与CT衰减校正的比较
本研究开发了一种利用骨SPECT发射数据进行衰减校正的新方法(B-SAC),并在15例前列腺癌患者的112个骨转移灶中比较了B-SAC与CT衰减校正(CTAC)的SUV。结果显示B-SAC与CTAC的SUV相关性极好(r>0.94),但B-SACN存在约17%的恒定高估,提示B-SAC有望在缺乏CT数据时改善骨SPECT定量,但目前版本获得的SUV难以直接与CTAC比较。
要点
- B-SAC与CTAC的SUV相关性极好(r>0.94),但存在轻微恒定高估。
- B-SACN较CTAC高估骨转移瘤负荷约17%(同一代谢肿瘤体积阈值下)。
- 当前版本B-SAC获得的SUV难以直接与CTAC获得的SUV比较。
核心结论
- B-SAC有望在缺乏X线CT或透射CT数据时改善骨SPECT对骨转移的定量。
- 考虑到CTAC与B-SAC之间病灶SUV存在虽小但不可忽略的差异,当前版本B-SAC获得的SUV似乎难以直接与CTAC获得的SUV比较。
研究设计与样本
- 本研究为初步研究,纳入15例前列腺癌患者的112个骨转移瘤。
- 开发了基于骨SPECT发射数据的衰减校正方法(B-SAC),将发射数据虚拟分割为骨、软组织和空气三种组织。
- 比较CTAC联合散射校正(SC)和分辨率恢复(RR)与以下重建条件下的SUV:B-SACN(+)SC(+)RR(+)、B-SACM(+)SC(+)RR(+)、UAC(+)SC(+)RR(+)、AC(-)SC(+)RR(+)和NC。
主要发现
- 五种重建条件下的SUV均与CTAC(+)SC(+)RR(+)相关(p<0.01)。
- B-SACN或B-SACM与CTAC图像之间的相关性极好(r>0.94)。
- Bland-Altman分析显示,CTAC(+)SC(+)RR(+)与其他五种重建之间的平均SUV差异分别为:B-SACN(+)SC(+)RR(+)为0.85±2.25,B-SACM(+)SC(+)RR(+)为1.61±2.36,UAC(+)SC(+)RR(+)为1.54±3.84,AC(-)SC(+)RR(+)为-3.12±4.97,NC为-5.96±4.59。
- 与CTAC(+)SC(+)RR(+)相比,B-SACN(+)SC(+)RR(+)在采用相同代谢肿瘤体积阈值时,对骨转移瘤负荷显示出轻微但恒定的高估(约17%)。
临床怎么用
- 在无X线CT或透射CT数据时,B-SAC可用于改善骨SPECT对骨转移的定量评估。
- 当前版本B-SAC获得的SUV不宜直接与CTAC获得的SUV进行比较。
局限性
- 本研究为初步研究,样本量较小(15例患者,112个病灶)。
- B-SAC与CTAC之间存在虽小但不可忽略的病灶SUV差异。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Evidence has shown the clinical usefulness of measuring the metastatic tumor burden of bone for prognostic assessment especially in prostate cancer; quantitative evaluation by dedicated SPECT is difficult due to the lack of attenuation correction (AC) method. We developed a novel method for attenuation correction using bone SPECT emission data (bone SPECT-based segmented attenuation correction; B-SAC) where emission data were virtually segmented into three tissues (i.e., bone, soft tissue, and air). Then, the pixel values in SPECT were replaced by 50 for the virtual soft tissue, and - 1000 for the virtual air. The replaced pixel values for the virtual bone were based on the averaged CT values of the normal vertebrae (B-SACN) or the metastatic bones (B-SACM). Subsequently, the processed SPECT data (i.e., SPECT value) were supposed to realize CT data (i.e., CT value) that were used for B-SAC. The standardized uptake values (SUVs) of 112 metastatic bone tumors in 15 patients with prostate cancer were compared between CTAC with scatter correction (SC) and resolution recovery (RR) and the following reconstruction conditions: B-SACN (+)SC(+)RR(+), B-SACM (+)SC(+)RR(+), uniform AC(UAC)(+)SC(+)RR(+), AC(-)SC(+)RR(+), and no correction (NC). RESULTS: The SUVs in the five reconstruction conditions were all correlated with those in CTAC(+)SC(+)RR(+) (p < 0.01), and the correlations between B-SACN or B-SACM and CTAC images were excellent (r > 0.94). Bland-Altman analysis showed that the mean SUV differences between CTAC (+)SC(+)RR(+) and the other five reconstructions were 0.85 ± 2.25 for B-SACN (+)SC(+)RR(+), 1.61 ± 2.36 for B-SACM (+)SC(+)RR(+), 1.54 ± 3.84 for UAC(+)SC(+)RR(+), - 3.12 ± 4.97 for AC(-)SC(+)RR(+), and - 5.96 ± 4.59 for NC. Compared to CTAC(+)SC(+)RR(+), B-SACN (+)SC(+)RR(+) showed a slight but constant overestimation (approximately 17%) of the metastatic tumor burden of bone when the same threshold of metabolic tumor volume was used. CONCLUSIONS: The results of this preliminary study suggest the potential for B-SAC to improve the quantitation of bone metastases in bone SPECT when X-ray CT or transmission CT data are not available. Considering the small but unignorable differences of lesional SUVs between CTAC and B-SAC, SUVs obtained with the current version of B-SAC seem difficult to be directly compared with those obtained with CTAC.
引用
- Nakahara T, Owaki Y, Shindou T, et al. Bone SPECT-based segmented attenuation correction for quantitative analysis of bone metastasis (B-SAC): comparison with CT-based attenuation correction. EJNMMI Res. 2019. PMID: 30888521. doi:10.1186/s13550-019-0501-1
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。