Abstract

Objective

To investigate if size measurements of liver observations is more variable in the arterial phase as suggested by LI-RADS and assess potential higher instability in categorization in this particular phase. Secondarily, to assess inter- and intra-reader agreement for size across phases.

Materials and methods

Patients with liver cirrhosis who underwent multi-arterial phase MRI between 2017 and 2018 were retrospectively selected. Three radiologists measured liver observations in each phase, independently, in a random order. Mean size between early and late arterial phases (AP), 2, 3 and 10 min delay and the number of observations crossing the LI-RADS size thresholds (10 and 20 mm) per phase were compared using McNemar’s test. Reader agreement was evaluated using intraclass correlation coefficient (ICC) and bootstrap-based comparisons. Bonferroni’s correction was applied to pairwise comparisons.

Results

94 observations (LR-3, LR-4, LR-5, and LR-M) were included. Mean sizes (mm) were late AP: 19.9 (95% CI 17.2, 24.2), 2 min delay: 19.8 (95% CI 17.1, 24.0), 3 min delay: 19.8 (95% CI 17.2, 24.0), 10 min delay: 20.2 (95% CI 17.5, 24.5) (p = 0.10–0.88). There was no difference between phases in number of observations that could have changed category due to variability in size (p = 0.546–1.000). Inter- and intra-reader agreement was excellent (ICC = 0.952–0.981).

Conclusion

Measurements of focal liver observations were consistent across all post-contrast imaging phases and we found no higher instability in LI-RADS category in any particular phase. Inter- and intra-reader agreement for size was excellent for each phase. Based on these findings, size measurement could be allowed on any post-contrast phase, including the arterial phase, if deemed appropriate by the radiologist.

中文翻译：

---

检查 LI-RADS 建议：是否应仅在非动脉期测量观察大小？

摘要

客观的

调查肝脏观察的大小测量是否如 LI-RADS 所建议的那样在动脉期变化更大，并评估在这个特定阶段分类中潜在的更高的不稳定性。其次，评估跨阶段大小的读者间和读者内一致性。

材料和方法

回顾性选择2017-2018年间接受多动脉期MRI检查的肝硬化患者。三位放射科医生以随机顺序独立地测量了每个阶段的肝脏观察结果。使用 McNemar 检验比较了早期和晚期动脉期 (AP)、2、3 和 10 分钟延迟之间的平均大小以及每个阶段跨越 LI-RADS 大小阈值（10 和 20 毫米）的观察次数。使用类内相关系数 (ICC) 和基于引导程序的比较来评估读者一致性。Bonferroni 的校正应用于成对比较。

结果

包括 94 个观察结果（LR-3、LR-4、LR-5 和 LR-M）。平均大小 (mm) 为晚期 AP：19.9（95% CI 17.2、24.2）、2 分钟延迟：19.8（95% CI 17.1、24.0）、3 分钟延迟：19.8（95% CI 17.2、24.0）、10 分钟延迟: 20.2 (95% CI 17.5, 24.5) ( p  = 0.10–0.88)。由于大小的可变性而可能改变类别的观察数量在阶段之间没有差异（p  = 0.546–1.000）。读者间和读者内的一致性非常好（ICC = 0.952–0.981）。

结论

在所有对比后成像阶段，局灶性肝脏观察的测量结果一致，我们发现在任何特定阶段的 LI-RADS 类别中没有更高的不稳定性。对于每个阶段，读者间和读者内的大小一致性都非常好。根据这些发现，如果放射科医生认为合适，可以在任何对比后阶段（包括动脉阶段）进行尺寸测量。