论文部分内容阅读
目的评价屏气CT灌注成像指标的变异性,并探讨肝细胞癌病人的测量结果是否受商业软件升级的影响。材料与方法研究得到伦理委员会批准,并获得病人的知情同意。肝细胞癌病人的CT灌注检查结果分别由3位医师独立分析。间隔4周以后,2位医师重复了之前的分析。采用组内相关系数(ICC)和Bland-Altman分析确定一致性界限(LoA)以及重复测量相关性来评估相同医师、不同医师之间以及软件之间的一致性。结果 23例肝细胞癌病人进行了93次屏气CT灌注检查。不同医师之间的ICC在不同参数中的血流量(BF)(>0.91)很高,血容量(BV)(>0.84)较高,平均通过时间(MTT)和表面通透性(PS)则较低(>0.3和>0.39)。不考虑软件版本因素的影响,在不同读片时的ICC为BF和BV较高(>0.80),MTT尚好(>0.75),而PS较低(>0.38)。应用当前版本的软件,放射科医师之间的LoA值在不同参数中分别为BF33%、BV39%、MTT55%和PS93%。即使最有经验的医师,不同读片时的LoA值在不同参数中分别为BF35%、BV43%、MTT33%、PS79%。当前版本软件的BF、BV、PS值显著高于之前版本,而MTT值显著低于之前版本(P<0.01)。结论采用当前版本的CT灌注软件对HCC病灶进行屏气灌注成像检查时,只有当BF减少超过35%、BV减少超过43%或MTT增加超过55%时才可以考虑超出了分析的变异度范围。当前版本和之前版本软件的灌注指标之间不能交换使用。本研究的结果仅限于HCC的屏气灌注CT成像,可能不适于其他的肿瘤和检查方案。
Objective To evaluate the variability of breath CT perfusion imaging and to investigate whether the measurement of hepatocellular carcinoma patients is affected by the commercial software upgrade. Research on materials and methods was approved by the ethics committee and patients’ informed consent was obtained. CT perfusion in patients with hepatocellular carcinoma was independently analyzed by three physicians. After an interval of 4 weeks, the two doctors repeated the previous analysis. ICC and Bland-Altman analyzes were used to determine the agreement limits (LoA) and the correlation of repeated measures to assess the consistency between the same physician, different physicians, and software. Results Twenty-three patients with hepatocellular carcinoma underwent breath-hold CT perfusion examination. The blood flow (BF) (> 0.91), blood volume (BV) (> 0.84), mean transit time (MTT) and surface permeability Lower (> 0.3 and> 0.39). Regardless of software version, ICC was higher for BF and BV (> 0.80), MTT was better (> 0.75), and PS was lower (> 0.38) for different reads. With the current version of the software, LoA values among radiologists were BF33%, BV39%, MTT55% and PS93%, respectively, in different parameters. Even for the most experienced physicians, LoA values for different readings were BF35%, BV43%, MTT33%, and PS79% for different parameters. BF, BV, PS values of the current version of the software were significantly higher than the previous version, while MTT values were significantly lower than the previous version (P <0.01). Conclusions With the current version of CT perfusion software for breath-hold perfusion imaging of HCC lesions, the range of variability beyond the analysis can only be considered if the reduction of BF is greater than 35%, the BV is reduced by more than 43%, or the MTT is increased by more than 55%. Between the current version and the previous version of the software perfusion index can not be used interchangeably. The results of this study are limited to breath-hold CT imaging of HCC and may not be suitable for other oncology and examination protocols.