房颤左心房几何:电解剖标测与CT对比
本研究回顾性纳入211例首次导管消融且术前行CT的房颤患者,比较电解剖标测与CT获得的左心房容积、表面积和球形度。结果显示两者相关性中等,电解剖标测系统性高估容积和表面积,低估球形度,提示其可作为补充而非替代CT的解剖评估手段。
要点
- 电解剖标测高估左心房容积13.2%,报告消融相关测量时需注意此系统偏差。
- 电解剖标测与CT对左心房整体形状的相似性尚可,但局部解剖细节差异明显。
- 消融术前CT仍是左心房解剖评估的参考标准,电解剖标测不宜单独用于精确几何定量。
核心结论
- 电解剖标测可快速提供术中左心房几何近似,但不能完全复现CT所示解剖。
- 两种模态在整体形状上有部分一致性,但存在系统性尺度和局部差异,支持电解剖标测作为补充而非替代CT的角色。
研究设计与样本
- 回顾性研究,纳入211例接受首次导管消融且术前行CT成像的房颤患者。
- 从CT和电解剖标测中分别提取左心房容积、表面积和球形度。
- 通过半轴(a、b、c)将共同椭球模型拟合至两组数据进行比较。
- 一致性采用Pearson相关与Bland-Altman分析评估。
- 几何差异通过原始网格与椭球近似上的RMS径向差异和Dice相似度量化。
关键影像表现/征象
- 电解剖标测与CT在左心房容积(r=0.58)、面积(r=0.56)和球形度(r=0.49)上均显著相关(均p<0.001)。
- Bland-Altman分析显示电解剖标测系统性高估容积11.1 cm³(13.2%)和表面积8.9 cm²(9.6%),轻微低估球形度(-0.01)。
- 原始网格的RMS径向差异高于椭球近似(0.14对0.08)。
- 椭球近似的Dice相似度高于原始几何(0.89,IQR 0.85–0.91对0.81,IQR 0.77–0.83)。
诊断标准与关键数值
- Pearson相关系数:容积r=0.58,面积r=0.56,球形度r=0.49,均p<0.001。
- Bland-Altman系统偏差:容积高估11.1 cm³(13.2%),表面积高估8.9 cm²(9.6%),球形度低估-0.01。
- RMS径向差异:原始网格0.14,椭球近似0.08。
- Dice相似度:椭球近似0.89(IQR 0.85–0.91),原始几何0.81(IQR 0.77–0.83)。
临床应用与阅片要点
- 解读电解剖标测测量的左心房容积时应考虑约13%的总体高估趋势。
- 电解剖标测对左心房整体形状和尺度的把握尚可,可用于术中大致定位,但精确解剖边界应参考术前CT。
- 当两种模态的左心房几何差异明显时,局部解剖变异和标测电极贴靠或取样密度可能是原因。
局限性
- 本研究为回顾性单中心设计。
- 未明确说明电解剖标测与CT采集之间的时间间隔及心律状态是否完全一致。
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: Atrial fibrillation (AF) involves heterogeneous remodelling of the left atrium (LA), making accurate geometric quantification clinically relevant. CT is the reference standard for anatomical assessment, whereas electroanatomical mapping (EAM) provides intraprocedural reconstruction with electrophysiological information. However, agreement between both modalities remains incompletely characterized.
METHODS: We retrospectively analyzed 211 patients undergoing first-time catheter ablation with pre-procedural CT imaging. LA volume, surface area, and sphericity were derived from CT and EAM. A common ellipsoidal model was fitted to both datasets for comparison via semi-axes (a, b, c). Agreement was assessed using Pearson correlation and Bland-Altman analysis. Geometric differences were quantified using RMS radial differences and Dice similarity on both original meshes and ellipsoidal approximations.
RESULTS: EAM correlated with CT for volume (r = 0.58), area (r = 0.56), and sphericity (r = 0.49) (all p < 0.001). Bland-Altman analysis demonstrated systematic overestimation of volume (11.1 cm3, 13.2 %) and surface area (8.9 cm2, 9.6 %), with slight underestimation of sphericity (-0.01). RMS radial differences were higher for original meshes than ellipsoidal approximations (0.14 vs 0.08), while Dice similarity was higher for ellipsoids than original geometries (0.89 IQR [0.85-0.91] vs 0.81 IQR [0.77-0.83]).
CONCLUSIONS: EAM provides a rapid intraprocedural approximation of left atrial geometry but does not fully reproduce CT-derived anatomy. While global shape is partially preserved, systematic scale differences and local discrepancies persist, supporting its role as a complementary rather than substitute modality.
引用
- Moriones, Lamata, Echebarria, et al. Left atrial geometry in atrial Fibrillation: A comparison between electroanatomical mapping and computed tomography. Eur J Radiol. 2026. PMID: 42398464. doi:10.1016/j.ejrad.2026.113042
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。