前列腺癌患者常见动态对比增强MRI参数的观察者内和观察者间变异性对T1时间校正的依赖性
该回顾性研究纳入35个MRI引导活检证实的前列腺癌病灶,由两名放射科医生分别使用个体测量T1时间(A)和固定T1时间(B)两次评估DCE参数,间隔三周。结果显示除kep外,B条件下观察者内和观察者间变异系数均小于A;个体测量T1校正使观察者内和观察者间变异性增加,kep对T1校正偏差最稳健。
要点
- 除kep外,固定T1时间(B)的观察者内和观察者间变异系数均小于个体测量T1时间(A)。
- kep对T1时间校正引起的偏差最不敏感,是所研究DCE参数中最稳健的参数。
- 个体测量T1时间校正使观察者内和观察者间变异性增加,定量DCE参数解读需考虑此影响。
核心结论
- T1时间校正对观察者内和观察者间变异性有显著影响。
- 采用个体测量T1时间校正时,观察者内和观察者间变异性均增加。
- 在所研究的DCE参数中,kep对T1时间校正引起的偏差最稳健。
研究设计与样本
- IRB批准的回顾性研究,数据集包含35个MRI引导活检证实的前列腺癌病灶。
- 两名经验丰富的放射科医生分别使用个体测量T1时间(A)和固定T1时间(B)评估DCE参数,间隔三周重复两次。
- 采用一致性限(LoA)和变异系数(CoV)评估参数的观察者内和观察者间变异性。
诊断标准与关键数值
- 观察者内平均CoV(涵盖所有四个DCE参数和两名读者)在A条件下为6.7%[0.5%, 15.1%],在B条件下为3.9%[0.2%, 11.0%]。
- 观察者间CoV在A条件下为9.0%[0.6%, 25.8%],在B条件下为7.0%[0.3%, 25.4%]。
- 除kep外,所有DCE参数在B条件下的观察者内和观察者间CoV均小于A条件。
- kep的绝对值对T1时间校正引起的偏差基本不敏感。
临床应用与阅片要点
- 定量DCE参数评估时,T1时间校正方式的选择会影响观察者内和观察者间变异性。
- kep对T1时间校正偏差较稳健,可能更适合用于需要减少校正相关变异性的定量分析。
- 采用个体测量T1时间校正时,需注意观察者内和观察者间变异性可能增加。
局限性
- 原文未报告局限性。
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)。
英文原文(摘要)· 点击展开
BACKGROUND: Dynamic contrast enhanced (DCE) MRI parameters are potential biomarkers to characterise tumour vasculature and distinguish it from the non-cancerous blood vessel system within the prostate. However, the inevitable presence of intra- and inter-observer variabilities is challenging in this context. Additionally, pre-contrast T1-time correction is a prerequisite to gain quantitative DCE parameters in the first place. The current study investigated the effect of individualized T1-time correction on intra- and inter-reader variability for quantitative DCE-parameters in prostatic lesions.
METHODS: In this IRB-approved retrospective study, two experienced radiologists assessed DCE parameters using individually measured (A) and fixed (B) T1-times twice with a time difference of three weeks. The dataset consisted of 35 MRI-guided biopsy-proven prostate cancer lesions. Limits of agreement (LoA) and coefficients of variability (CoV) were calculated to assess intra- and inter-reader variabilities of the parameters.
RESULTS: With exception of kep, for all DCE parameters both intra- and inter-reader CoV were smaller in B compared to A. Absolute kep values were largely insensitive to T1-time correction induced bias. The mean intra-reader CoVs [5%, 95% percentile] (over all four DCE parameters and both readers) were 6.7% [0.5%, 15.1%] in A and 3.9% [0.2%, 11.0%] in B. The inter-reader CoVs were 9.0% [0.6%, 25.8%] (A) and 7.0% [0.3%, 25.4%] (B).
CONCLUSIONS: T1-time correction has a significant influence on the intra- and inter-reader variability. By applying individually measured T1-time correction, both intra- and inter-observer variability were found to increase. Out of all investigated DCE parameters, kep is the most robust to this investigated bias.
引用
- Daniel M, Polanec SH, Wengert G, et al. Intra- and inter-observer variability in dependence of T1-time correction for common dynamic contrast enhanced MRI parameters in prostate cancer patients. Eur J Radiol. 2019. PMID: 31153570. doi:10.1016/j.ejrad.2019.04.015
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。