A study on the consistency of myocardial extracellular volume quantification in the systole and diastole phases using dual-layer detector spectral CT
刘紫渲 1张钰 1李燕君 1程勇 1帅桃 1王紫薇 1李真林 1张琳琳 王春杰
扫码查看
点击上方二维码区域，可以放大扫码查看
作者信息
1. 四川大学华西医院放射科，成都 610041
折叠
摘要
目的 探究双层探测器光谱CT定量心肌细胞外容积（ECV）在心脏收缩期与舒张期的一致性。 方法 该研究为横断面研究。回顾性纳入2022年4月至12月于四川大学华西医院行心脏光谱CT检查的患者35例，扫描前3 d内采集红细胞比容，扫描时采用双层探测器光谱CT获得延迟期的心脏收缩期（45%）及舒张期（75%）全息光谱图像，并通过光谱后处理工作站分别得到收缩期和舒张期基于碘密度的ECV图像（CT-ECV）。根据美国心脏协会左心室16节段模型，重建出标准心脏短轴位图像，在左心室心底部、心中部和心尖部3个切面，将心肌标准化分为16节段。两名放射科医师采用5分法对收缩期和舒张期的3个切面以及全心整体的CT-ECV图像质量进行主观评分，并分别测量、计算收缩期和舒张期心肌16个节段和全心的心肌CT-ECV。对于两名医师的主观评价，采用Kappa检验评价评分的一致性，采用Wilcoxon符号秩和检验比较收缩期和舒张期之间的评分差异。采用配对样本t检验比较收缩期及舒张期各节段CT-ECV的差异。两名医师在CT-ECV图像上测量结果的观察者内和观察者间的一致性使用ICC检验。 结果 两名医师对收缩及舒张期CT-ECV图像质量主观评分的一致性较好（Kappa>0.80），收缩期与舒张期心底层、心中层、心尖层及全心整体之间的图像质量差异均无统计学意义（P>0.05）。35例患者通过延迟期获得的收缩期及舒张期全心CT-ECV分别为（33.29±3.46）%、（33.50±3.39）%，差异无统计学意义（t=-0.78，P=0.442）。心肌节段8收缩期、舒张期CT-ECV为（34.15±3.94）%、（35.30±3.99）%，节段9收缩期、舒张期CT-ECV为（34.03±3.76）%、（35.46±3.74）%，节段14收缩期、舒张期CT-ECV为（33.98±3.32）%、（35.05±3.98）%，节段8、9、14的平均收缩期CT-ECV低于舒张期，差异具有统计学意义（t值分别为-2.65、-3.26、-2.42，P值分别为0.012、0.003、0.022）。两名放射科医师测量的收缩期和舒张期16个节段的CT-ECV的ICC均大于0.90，一致性较好。 结论 基于双层探测器光谱CT定量心肌ECV在收缩期与舒张期之间全心CT-ECV的差异没有统计学意义，但由于部分节段的心肌CT-ECV收缩、舒张期有2%以内的差异，故心肌CT-ECV测量应在同一期相的相同节段进行，以获得稳定、准确的ECV值。 Objective To investigate the consistency of myocardial extracellular volume between systole and diastole using dual-layer detector spectral CT. Methods This was a cross-sectional study. Thirty-five patients who underwent cardiac spectral CT examination in West China Hospital of Sichuan University from April 2022 to December 2022 were retrospectively collected. Hematocrit was collected within 3 days before the CT scan. The delayed phases holographic spectral images in systole (45%) and diastole (75%) were obtained using dual-layer spectral CT. CT data were processed using a spectral post-processing workstation, and the extracellular volume (ECV) based on iodine density images, referred as CT-ECV, in systolic and diastolic phases were calculated, respectively. According to the American Heart Association′s 16-segment model of left ventricular, the standard short-axis images were constructed, and the myocardium was standardized into 16 segments at the basal, mid-cavity, and apical levels of the left ventricle. Two radiologists performed a subjective evaluation in the image quality of the CT-ECV images of the whole heart and the three sections in systole and diastole using a "five-point" scale. The ECV of the 16 segments and the whole heart in systole and diastole was calculated. The consistency of subjective evaluations between systole and diastole was assessed using Kappa statistics. Wilcoxon signed-rank tests were used to compare the differences in scores between systole and diastole. Paired sample t-test was used to compare the differences in CT-ECV scores between systole and diastole. The intraclass correlation coefficient was used to test the intra-and inter-observer consistency of CT-ECV measurements between two radiologists. P<0.05 was statistically significant. Results There was good agreement between the two radiologists on subjective scores of CT-ECV image quality between systole and diastole (Kappa>0.80), and there was no statistical difference in image quality among the basal, mid-cavity, and apical levels of the left ventricle and whole heart between systole and diastole (P>0.05). The systolic and diastolic CT-ECV for the entire heart obtained through the delay phase were (33.29±3.46)% and (33.50±3.39)%, respectively, with no statistically significant difference (t=-0.78, P=0.442). CT-ECV in systole and diastole were (34.15±3.94)% and (35.30±3.99)% for segment 8, (34.03±3.76)% and (35.46±3.74)% for segment 9, and (33.98±3.32)% and (35.05±3.98)% for segment 14, respectively. The mean values of the systolic CT-ECV of segments 8, 9 and 14 were significantly lower than those of diastolic CT-ECV (t=-2.65, -3.26, -2.42, P=0.012, 0.003, 0.022, respectively). The ICCs for CT-ECV measurements of 16 segments by the two radiologists were greater than 0.90 in both systolic and diastolic, indicating good agreement. Conclusions There is no significant difference in whole heart CT-ECV values between systolic and diastolic myocardial ECV based on dual-layer spectral CT. However, minor differences (less than 2%) are found between systolic and diastolic myocardial CT-ECV for some segments. Myocardial CT-ECV measurement should be performed on the same segment during the same phase to obtain stable and accurate ECV values.
Abstract
Objective To investigate the consistency of myocardial extracellular volume between systole and diastole using dual-layer detector spectral CT. Methods This was a cross-sectional study. Thirty-five patients who underwent cardiac spectral CT examination in West China Hospital of Sichuan University from April 2022 to December 2022 were retrospectively collected. Hematocrit was collected within 3 days before the CT scan. The delayed phases holographic spectral images in systole (45%) and diastole (75%) were obtained using dual-layer spectral CT. CT data were processed using a spectral post-processing workstation, and the extracellular volume (ECV) based on iodine density images, referred as CT-ECV, in systolic and diastolic phases were calculated, respectively. According to the American Heart Association′s 16-segment model of left ventricular, the standard short-axis images were constructed, and the myocardium was standardized into 16 segments at the basal, mid-cavity, and apical levels of the left ventricle. Two radiologists performed a subjective evaluation in the image quality of the CT-ECV images of the whole heart and the three sections in systole and diastole using a "five-point" scale. The ECV of the 16 segments and the whole heart in systole and diastole was calculated. The consistency of subjective evaluations between systole and diastole was assessed using Kappa statistics. Wilcoxon signed-rank tests were used to compare the differences in scores between systole and diastole. Paired sample t-test was used to compare the differences in CT-ECV scores between systole and diastole. The intraclass correlation coefficient was used to test the intra-and inter-observer consistency of CT-ECV measurements between two radiologists. P<0.05 was statistically significant. Results There was good agreement between the two radiologists on subjective scores of CT-ECV image quality between systole and diastole (Kappa>0.80), and there was no statistical difference in image quality among the basal, mid-cavity, and apical levels of the left ventricle and whole heart between systole and diastole (P>0.05). The systolic and diastolic CT-ECV for the entire heart obtained through the delay phase were (33.29±3.46)% and (33.50±3.39)%, respectively, with no statistically significant difference (t=-0.78, P=0.442). CT-ECV in systole and diastole were (34.15±3.94)% and (35.30±3.99)% for segment 8, (34.03±3.76)% and (35.46±3.74)% for segment 9, and (33.98±3.32)% and (35.05±3.98)% for segment 14, respectively. The mean values of the systolic CT-ECV of segments 8, 9 and 14 were significantly lower than those of diastolic CT-ECV (t=-2.65, -3.26, -2.42, P=0.012, 0.003, 0.022, respectively). The ICCs for CT-ECV measurements of 16 segments by the two radiologists were greater than 0.90 in both systolic and diastolic, indicating good agreement. Conclusions There is no significant difference in whole heart CT-ECV values between systolic and diastolic myocardial ECV based on dual-layer spectral CT. However, minor differences (less than 2%) are found between systolic and diastolic myocardial CT-ECV for some segments. Myocardial CT-ECV measurement should be performed on the same segment during the same phase to obtain stable and accurate ECV values.
万方数据