文献速递冠心病

中度冠状动脉狭窄中静息全周期比值与尼可地尔FFR的诊断一致性及阈值对应性

Diagnostic Agreement and Threshold Correspondence Between Resting Full‐Cycle Ratio and Nicorandil‐Based Fractional Flow Reserve in Intermediate Coronary Stenosis: A Prospective Single‐Center Registry
RSS 订阅 · Catheterization and Cardiovascular Interventions · 2026-09-30

在一项前瞻性单中心注册研究中,300例患者(399处病变)序贯接受静息全周期比值(RFR)和基于尼可地尔的血流储备分数(FFR)测量。总体阈值一致性为85.0%(95%CI 81.1%–88.5%),RFR平均比FFR高+0.067。不一致主要集中在FFR 0.80决策边界附近及RFR 0.87–0.91区间,该区间与既往混合策略的灰区重叠。

要点

  • RFR与尼可地尔FFR总体一致性较高,但决策边界附近一致性明显下降。
  • RFR 0.89约对应FFR 0.823,RFR 0.87–0.91区间内一致性仅63.2%。
  • 前降支病变是不一致的探索性相关因素,可能反映其FFR分布特征。

核心结论

  • RFR与基于尼可地尔的FFR在中等程度冠状动脉狭窄中具有较高一致性,与既往队列结果一致。
  • 不一致主要集中于FFR 0.80决策边界附近,符合两个相关连续指标在固定阈值处二分类化的预期。
  • 不一致在RFR 0.87–0.91区间最集中,该探索性区间与既往混合策略灰区重叠,需结局研究验证。

研究设计与样本

  • 前瞻性单中心注册研究,纳入300例患者共399处中等程度冠状动脉狭窄病变。
  • 每处病变序贯进行静息全周期比值(RFR)和基于尼可地尔的FFR测量。
  • 诊断准确性采用患者水平聚类自助法计算置信区间,阈值对应性采用普通最小二乘法和Deming回归分析。
  • 探索性采用单变量logistic回归评估不一致的相关因素。

主要发现

  • 总体阈值一致性为85.0%(95%CI 81.1%–88.5%),κ值为0.684,AUC为0.932(95%CI 0.906–0.954)。
  • RFR平均比FFR高+0.067,患者均值的一致性界限为−0.040至+0.173。
  • 远离FFR 0.80决策边界时一致性为91.6%–92.8%,但在FFR决策边界两侧相邻分层中一致性为72.0%–73.5%。
  • RFR 0.89对应FFR约为0.823,Deming回归结果相似。
  • 在RFR 0.87–0.91区间外一致性超过92%,但该探索性区间内一致性为63.2%,既往混合策略0.86–0.92区间内为68.8%。

关键数值与推荐

  • 总体阈值一致性85.0%(95%CI 81.1%–88.5%),κ=0.684,AUC 0.932(95%CI 0.906–0.954)。
  • RFR与FFR的患者均值差为+0.067,一致性界限−0.040至+0.173。
  • FFR 0.80决策边界两侧相邻分层的一致性为72.0%–73.5%。
  • RFR 0.89对应FFR约0.823。
  • RFR 0.87–0.91探索性区间内一致性为63.2%,0.86–0.92区间内为68.8%。

临床怎么用

  • RFR可用于中等程度冠状动脉狭窄的功能学评估,但决策边界附近结果需谨慎解读。
  • RFR 0.87–0.91区间与既往混合策略灰区重叠,该范围内可考虑加测FFR确认。
  • 前降支病变出现不一致的可能性需在临床决策中予以关注。

局限性

  • RFR 0.87–0.91为事后探索性区间,尚需前瞻性结局研究验证。
  • 前降支相关分析为探索性单变量结果,未进行多因素校正。

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)。

原文

ABSTRACT Background RFR has shown good aggregate agreement with FFR, including in prior Asian cohorts using intracoronary nicorandil. Where RFR-FFR disagreement is concentrated relative to the FFR 0.80 decision boundary, and how the two binary thresholds correspond, remain incompletely characterized. Methods In a prospective single-center registry, 300 patients (399 lesions) underwent sequential RFR and nicorandil-based FFR measurement. Diagnostic accuracy was assessed with patient-level cluster bootstrap CIs across post hoc FFR and RFR strata. Threshold correspondence used ordinary least squares and Deming regression. Univariable logistic regression evaluated correlates of discordance (exploratory). Results Overall threshold-based agreement was 85.0% (95% CI 81.1%−88.5%; κ = 0.684) with AUC 0.932 (95% CI 0.906−0.954). RFR exceeded FFR by +0.067 (patient-mean limits of agreement −0.040 to +0.173). Agreement was 91.6−92.8% away from the FFR 0.80 decision boundary but 72.0−73.5% in the two flanking strata. RFR 0.89 corresponded to an FFR of approximately 0.823, similar with Deming regression. Agreement exceeded 92% outside RFR 0.87−0.91 but was 63.2% within this exploratory band and 68.8% within the 0.86−0.92 interval of previous hybrid strategies. LAD territory was an exploratory univariable correlate of discordance, which likely reflected the FFR distribution of LAD lesions. Conclusions RFR showed high agreement with nicorandil-based FFR, consistent with prior cohorts. Disagreement was concentrated near the FFR decision boundary, as expected for two correlated continuous indices dichotomized at fixed thresholds, and this study quantifies its frequency. Agreement was lowest near the RFR cut-off (0.87-0.91), an exploratory interval that overlaps with previous hybrid gray zones and requires validation in outcome studies.

来源:Catheterization and Cardiovascular Interventions,查看原文。版权归原作者所有。

引用

  1. Catheterization and Cardiovascular Interventions:Diagnostic Agreement and Threshold Correspondence Between Resting Full‐Cycle Ratio and Nicorandil‐Based Fractional Flow Reserve in Intermediate Coronary Stenosis: A Prospective Single‐Center Registry

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。