血糖水平对18F-FDG PET/CT图像质量的影响

目的 研究血糖水平对18F-脱氧葡萄糖(FDG) PET/CT图像质量的影响.方法 80例行体格检查或评价肿瘤性质的受检者,按空腹血糖水平分为9组(组间距为1mmol/L),第1组为血糖正常组,即血糖＜6.0mmol/L(10例);第2～9组血糖水平高于正常,分别为6.0～6.9mmol/L(11例),7.0～7.9mmol/L(13例),8.0～8.9mmol/L(11例),9.0～9.9mmol/L(11例),10.0～10.9mmol/L(8例),11.0～11.9mmol/L(6例),12.0～12.9mmol/L(5例),≥13.0mmol/L(5例).受检者做完PET/CT检查后,图像质量由2位有经验的医师独立判断.同时根据肝不同层面的标准摄取值(SUV)最大值(SUVmax)及SUV平均值(SUVavg)分别计算肝图像噪声.采用SPSS 12.0软件进行统计学处理.结果 (1)9组之间图像质量评分、肝噪声差异有统计学意义(P均＜0.05).将第2～9组图像质量评分、肝噪声分别与第1组进行比较,第2～7组与第1组差异无统计学意义(P均＞0.05);第8,9组与第1组比较差异有统计学意义(P均＜0.05),且评分低于第1组.(2)血糖水平与图像质量评分呈负相关(r=-0.52,P＜0.05);血糖水平与肝噪声呈正相关(SUVmax、SUVavgr值分别为0.33和0.60,P均＜0.05);SUVavg所算噪声与血糖水平的相关性优于SUVmax.结论 图像质量随血糖水平的升高而下降,血糖＜12.0mmol/L时与血糖正常者PET/CT图像质量差异无统计学意义,但当血糖水平≥12.0mmol/L时图像质量将显著下降.

Effect of blood glucose level on 18F-FDG PET/CT imaging

Objective The aim of this study was to investigate the effect of blood glucose level on the image quality of 18F-fluorodeoxyglucese (FDG) PET/CT imaging. Methods Eighty patients referred to the authors' department for routine whole-body 18F-FDG PET/CT check up were recruited into this study. The patients were classified into 9 groups according to their blood glucose level: normal group ＜6.0mmol/L (n=10),6.0～6.9mmol/L(n=11),7.0～7.9mmol/L(n=13),8.0～8.9mmol/L(n=11),9.0～9.9mmol/L(n=11),10.0～10.9mmol/L(n=8),11.0～11.9mmol/L(n=6),12.0～12.9mmol/L(n=5),≥13.0mmol/L(n=5). The image quality was reviewed independently by two experienced doctors. The noise level was calculated according to the average and maximum standardized uptake value (SUVavg and SUVmax) of liver on different slices. SPSS 12.0 was used to analyse the data. Results (1) There were significant differences among the 9 groups in image quality scores and image noises (all P＜0.05). Using the first group as normal reference, there were no significant differences in image quality scores and image noises from group 2 to group 7. However, groups 8 and 9 showed statistical significant differences contrasted to group 1(P＜0.05 ). The image quality scores of groups 8 and 9 were less but the image noises were higher than that of group 1.(2) There was negative correlation between image quality scores and blood glucose level (r=-0.52,P＜0.05). The image noise was positively correlated with blood glucose level (r of SUVavg and SUVmax:0.60 and 0.33, P＜0.05). Conclusions The higher the blood glucose level, the worse the image quality. When the blood glucose level is more than or equal to 12.0mmol/L, the image quality will significantly degrade.