Node-RADS在可切除胃癌区域淋巴结转移鉴别中优于双能CT定量参数
该回顾性研究纳入115例可切除胃癌患者的151枚区域淋巴结,比较Node-RADS、双能CT定量参数及其联合的诊断性能。结果显示Node-RADS(截断值≥3)的AUC为0.859,显著优于短径(AUC 0.780)和延迟期碘浓度(AUC 0.612),联合算法未进一步提升性能。
要点
- Node-RADS评分≥3鉴别转移淋巴结的AUC为0.859,优于短径和双能CT定量参数。
- 延迟期碘浓度是唯一与淋巴结转移相关的双能CT参数,但诊断性能有限(AUC 0.612)。
- 基于碘浓度的改良Node-RADS算法未改善诊断性能,建议直接使用初始Node-RADS评分。
关键影像表现/征象
- Node-RADS评分基于淋巴结的形态学特征(如大小、边界、强化方式等)进行标准化评估,原文未详细列出具体征象。
- 双能CT定量参数包括动脉期、静脉期和延迟期的40 keV CT衰减值、碘浓度和有效原子序数,其中仅延迟期碘浓度与淋巴结转移相关。
诊断标准与关键数值
- Node-RADS初始评分以≥3为截断值鉴别转移与非转移淋巴结,AUC为0.859(95%CI:0.793-0.910)。
- 淋巴结短径的AUC为0.780,以>8 mm为二分类截断值时AUC为0.748。
- 延迟期碘浓度的AUC为0.612。
- Node-RADS的AUC显著优于短径和延迟期碘浓度(p<0.05,DeLong检验)。
临床应用与阅片要点
- 在双能CT术前评估中,可优先采用Node-RADS评分(截断值≥3)鉴别区域淋巴结转移,其标准化和可重复性优于单纯短径测量。
- 双能CT定量参数(如延迟期碘浓度)可作为辅助参考,但不建议单独用于诊断决策。
- 注意Node-RADS评估需结合淋巴结形态学特征,避免仅依赖大小标准。
局限性
- 回顾性单中心研究,样本量有限(115例患者,151枚淋巴结)。
- 仅纳入可切除胃癌患者,结论推广至其他分期或不可切除患者需谨慎。
- Node-RADS评估可能存在观察者间差异,原文未报告一致性数据。
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
OBJECTIVE: To evaluate and compare the diagnostic performance of Node-RADS, dual-energy CT (DECT)-derived quantitative parameters, and their combination for distinguishing metastatic from non-metastatic regional lymph nodes (LNs) in patients with resectable gastric cancer (GC).
MATERIALS AND METHODS: Between Jan 2022 and Oct 2025, 115 consecutive patients with resectable GC who underwent preoperative triple-phase enhanced DECT scans and radical gastrectomy with D2 lymphadenectomy were retrospectively enrolled. A total of 151 regional LNs (70 metastatic, 81 non-metastatic) with region-level pathological confirmation were selected for node-level comparison. Two radiologists independently assessed Node-RADS scores for each LN. LN short diameter (SD) was measured at the venous phase on the maximal axial slice. DECT-derived parameters (CT attenuation at 40 keV, iodine concentration (IC), and effective atomic number (Zeff) across arterial, venous, and delayed phases) were analyzed using a freehand region of interest on the maximal axial slice of the LN. Diagnostic performances were evaluated using receiver operating characteristic curve analysis and compared using the DeLong test.
RESULTS: IC at delayed phase (ICDP) was the sole parameter associated with LN involvement. The initial Node-RADS score (cut-off ≥ 3) achieved an AUC of 0.859 (95% CI: 0.793-0.910), significantly outperforming SD (AUC: 0.780, p < 0.05) and ICDP (AUC: 0.612, p < 0.05). Two modified Node-RADS algorithms incorporating ICDP under different rule-in and rule-out criteria were established, but did not improve diagnostic performance beyond the initial Node-RADS scores.
CONCLUSION: Node-RADS demonstrates superior diagnostic performance compared with DECT-derived quantitative parameters for identifying metastatic regional LNs in resectable GC, and can be reliably applied in the DECT settings.
KEY POINTS: Question Preoperative differentiation of metastatic from non-metastatic regional lymph nodes in resectable gastric cancer remains inaccurate using conventional size criteria. Findings Node-RADS (AUC = 0.859) outperformed SD diameter (continuous: AUC = 0.780; its > 8 mm dichotomized counterpart: AUC = 0.748) and ICDP (AUC = 0.612) for metastatic LN discrimination. Clinical relevance Node-RADS is a standardized and reproducible criterion that can be applied in dual-energy CT scenarios to effectively identify regional lymph node metastasis in resectable gastric cancer.
引用
- Yao DX, Lin XX, Bei TX, et al. Node-RADS outperforms dual-energy CT-derived quantifications for distinguishing metastatic from non-metastatic regional lymph nodes in resectable gastric cancer. Eur Radiol. 2026. PMID: 42967258. doi:10.1007/s00330-026-12993-8
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。