从切片到表面：3D切片到体积MRI在临床实践中用于胎儿脑生物测量的可行性
三种小儿神经放射学家独立测量8项参数，SVR图像主观质量更高、测量时间更短，组间信度均优秀；胼胝体长度在SVR上系统性偏小，提示3D SVR可能纠正斜切2D采集的测量偏倚。
要点
- SVR多平面重建可用于胎儿脑标准生物测量，组间信度优秀（ICC 0.93-0.99）。
- SVR图像主观质量评分中位数为3，高于2D图像的2（p<0.001）。
- SVR测量时间显著短于2D，但胼胝体长度在SVR上系统性偏小，报告时需注意。
关键影像表现/征象
- SVR图像主观质量评分中位数为3，2D图像为2，所有评估者间差异均显著（p<0.001）。
- SVR测量时间显著短于2D图像。
- 胼胝体长度在所有评估者中均一致且显著地在SVR图像上更小。
诊断标准与关键数值
- 组间信度：2D图像ICC范围0.87-0.99，SVR图像ICC范围0.93-0.99，均属优秀。
- 主观图像质量采用0-3分评分，SVR中位3分，2D中位2分（p<0.001）。
- 测量时间比较采用配对Wilcoxon符号秩检验，SVR显著短于2D。
临床应用与阅片要点
- 日常胎儿脑MRI生物测量可优先尝试SVR多平面重建，以提高图像质量和测量效率。
- 报告胼胝体长度时需注意SVR与2D测量存在系统性差异，SVR值偏小，随访对比时应使用同一方法。
局限性
- 单中心设计，结论外推需谨慎。
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 AND PURPOSE: Fetal brain biometry on magnetic resonance imaging (MRI) is essential for assessing neurodevelopment, but the standard-of-care two-dimensional (2D) acquisitions are susceptible to motion artifact and planar obliquity, which can compromise measurement accuracy. This study aimed to evaluate the feasibility, reliability, and efficiency of performing standard fetal brain biometry on multiplanar reconstructions (MPRs) generated with slice-to-volume reconstruction (SVR) compared with the conventional 2D approach.
MATERIALS AND METHODS: This IRB-approved retrospective study included twenty consecutive fetal MRI examinations with normal-appearing brains and indications other than central nervous system abnormalities. Three fellowship-trained pediatric neuroradiologists independently measured eight key biometric parameters on both standard 2D Half-Fourier Acquisition Single-shot Turbo spin-Echo (HASTE) images and on MPRs from SVR volumes. Inter-rater reliability was assessed using the intraclass correlation coefficient (ICC). Measurement differences, subjective image quality (0-3 scale), and measurement time were compared using the paired Wilcoxon signed-rank test.
RESULTS: Excellent inter-rater reliability was observed for all biometric parameters on both 2D (ICC range, 0.87-0.99) and SVR images (ICC range, 0.93-0.99). The median image quality score was significantly higher for 3D SVR images (3) compared with 2D images (2) across all raters (p <0.001). Measurement time was significantly shorter with SVR than with 2D images. While overall 2D and SVR measurements were similar across raters, corpus callosum length was consistently and significantly smaller on SVR images for all raters.
CONCLUSIONS: Fetal brain biometry using SVR is a feasible, reliable, and efficient alternative to standard 2D imaging. The technique provides superior image quality and reduces measurement time while maintaining excellent inter-rater reliability. The systematic difference in corpus callosum measurements suggests that 3D SVR may improve accuracy by correcting for measurement bias inherent in oblique 2D acquisitions.
引用
- Steinhardt N, Brady D, Baer A, et al. From Slices to Surfaces: The Feasibility of Fetal Brain Biometry Using 3D Slice-to-Volume MRI in Clinical Practice. AJNR Am J Neuroradiol. 2026. PMID: 42097850. doi:10.3174/ajnr.A9397
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。