CZT-SPECT/CT肾动态显像中核医学断层显像对测量肾脏体积的可行性探究

Investigation of the feasibility of measuring renal volume by nuclear medicine tomography in CZT-SPECT/CT renal dynamic imaging

    摘要
  • 摘要:
    目的  以CT测量体积为参考标准，评估核医学断层显像最大临界值自动勾画法在99Tcm-二亚乙基三胺五乙酸（DTPA）肾动态显像中自动测量肾脏体积的准确性。
    方法 收集2024年3月至7月于天津医科大学总医院核医学科行碲锌镉（CZT）-SPECT/CT肾动态显像的29例患者进行前瞻性研究，其中男性20例、女性9例，年龄（43.9±2.2）岁。采用3D打印技术构建形态与体积参数符合成人肾脏解剖特征的中空肾脏体模，灌注99Tcm-DTPA溶液后行CZT-SPECT/CT核医学断层扫描，通过靶区20%最大标准摄取值（SUVmax）、30%SUVmax、40%SUVmax、50%SUVmax梯度临界值自动勾画肾脏感兴趣区（ROI），确定最佳临界值。临床验证阶段对比分析核医学断层显像法测量的肾脏体积与CT勾画法测量的肾脏体积差值的绝对值的差异。采用配对t检验进行组间比较，采用Pearson相关性分析和Bland-Altman法评估核医学断层显像法与CT勾画法得出肾脏体积的相关性与一致性，采用组内相关系数（ICC）评价2种测量方法的一致性。
    结果 肾脏体模实验中，SUVmax的临界值为靶区30%SUVmax时，测得的肾脏体模体积与实际肾脏体积差值的绝对值（3.07±5.89） cm3明显低于20%SUVmax（28.83±5.49） cm3、40%SUVmax（21.70±2.35） cm3、50%SUVmax（31.43±5.75） cm3，差异有统计学意义（t=3.961、2.937、3.444，均P<0.05）。临床验证实验中，SUVmax的临界值为靶区30%SUVmax时，29例患者核医学断层显像法与CT勾画法测得的双侧肾脏体积差异无统计学意义左肾：（166.88±9.41） cm3对（163.24±7.59） cm3，t=1.377；右肾：（166.85±8.28） cm3对（162.37±6.21） cm3，t=0.822，均P>0.05。核医学断层显像法与CT勾画法测量肾脏体积的左肾ICC为0.866，右肾ICC为0.809（均P<0.001）。核医学断层显像法与CT勾画法所测得的肾脏体积的相关系数较高（左肾：r=0.79；右肾：r=0.88，均P<0.001），相关性较好。Bland-Altman分析结果提示95%一致性界限处于临床可接受范围。
    结论 当核医学断层显像自动勾画ROI的临界值为30%SUVmax时，能够较为准确地获得肾脏体积。

     

    Abstract:
    Objective To evaluate the accuracy of automated delineation of the maximum cut-off value of nuclear medicine tomography in the automatic delineation of kidney volume in 99Tcm-diethylene-triamine pentaacetic acid (DTPA) renal dynamic scintigraphy. CT volumetric measurements were used as the reference standard.
    Methods A prospective study was conducted on 29 patients (20 males and 9 females, mean age (43.9±2.2) years) who underwent cadmium zinc telluride (CZT)-SPECT/CT renal dynamic imaging at the Department of Nuclear Medicine, Tianjin Medical University General Hospital from March 2024 to July 2024. A hollow kidney phantom, 3D printed to match adult renal anatomical morphology and volume parameters, was filled with 99Tcm-DTPA solution and scanned using CZT-SPECT/CT renal dynamic imaging. The optimal cut-off for renal region of interest (ROI) delineation was determined by applying gradient cut-off of 20% maximum standardized uptake (SUVmax), 30% SUVmax, 40% SUVmax, and 50% SUVmax of the clinical target volume. In the clinical validation phase, the absolute value of the difference between the kidney volumes measured by nuclear medicine tomography and CT-based delineation were compared. Paired t-tests were used for intergroup comparisons, and Pearson correlation analysis and Bland–Altman plots assessed correlation and agreement between the two methods. Intraclass correlation coefficient (ICC) was used to evaluate the agreement of the two measurement methods.
    Results In the phantom experiment, the absolute value of the difference between the measured and actual kidney volumes was significantly smaller at the 30% SUVmax of the clinical target volume cut-off ((3.07±5.89) cm3) compared with those at 20% SUVmax ((28.83±5.49) cm3), 40% SUVmax ((21.70±2.35) cm3), and 50% SUVmax ((31.43±5.75) cm3) (t=3.961, 2.937, 3.444, respectively; all P<0.05). In the clinical validation experiment, no significant difference was observed between the nuclear medicine tomography and CT-delineated kidney volumes at 30% SUVmax of the clinical target volume cut-off (left kidney: (166.88±9.41) cm3 vs. (163.24±7.59) cm3, t=1.377; right kidney: (166.85±8.28) cm3 vs. (162.37±6.21) cm3, t=0.822; both P>0.05). The ICC values for kidney volume measured by nuclear medicine tomography and CT contouring were 0.866 for the left kidney and 0.809 for the right kidney (both P<0.001). The correlation coefficient between the volume of the kidneys measured by nuclear medicine tomography and CT delineation was high (left kidney: r=0.79; right kidney: r=0.88; both P<0.001), and the correlation was good. Bland–Altman analysis showed 95% limits of agreement within clinically acceptable ranges.
    Conclusion Automatic delineation at the 30% SUVmax of ROI in nuclear medicine tomography provides accurate kidney volume measurements.

     

    • HTML全文
    • 参考文献(24)
    • 相关文章
    • 施引文献
  • 资源附件(0)

/

返回文章
返回