文献速递腹部泌尿生殖CT

虚拟单能量光谱衰减曲线鉴别偶发小肾脏病变

Virtual monochromatic spectral attenuation curve analysis for evaluation of incidentally detected small renal lesions using rapid kilovoltage-switching dual-energy computed tomography.
PubMed 文献 · Abdom Radiol (NY) · 2022-08-09

本研究回顾性纳入46例患者共78个肾脏病变(24个强化、54个非强化),评估快速千伏切换双能CT的光谱衰减曲线能否鉴别1-4 cm偶发小肾脏病变的强化与非强化。结果显示各定量与定性光谱参数均可显著区分两类病变(P<0.001),各斜率阈值诊断准确率达96和95,ΔHU阈值20 HU准确率达100,曲线形态目视分析准确率为94。

要点

  • DECT光谱衰减曲线可高准确度鉴别1-4 cm肾脏病变是否强化。
  • ΔHU阈值20 HU的准确率最高,斜率法与曲线形态目视分析亦佳。
  • 未做真正平扫时,DECT光谱分析价值最大。

核心结论

  • 快速千伏切换DECT的光谱衰减曲线鉴别偶发强化与非强化小肾脏病变具有极佳诊断准确率。
  • 该优势在未行真正平扫时最有帮助。

研究设计与样本

  • 回顾性研究,纳入46例患者共78个肾脏病变,其中强化24个、非强化54个。
  • 采用快速千伏切换DECT的DE模式,于门静脉期或肾实质期扫描。
  • 最终诊断由病理或按既定标准的影像随访确认。
  • 重建虚拟单能量图像,测量40-70 keV、40-100 keV、40-140 keV的斜率并记录曲线形态目视评估。
  • 诊断准确率用交叉验证Mann-Whitney U检验计算,相关性用Spearman rho及组内相关系数(ICC)评估。

关键影像表现/征象

  • 所有定量与定性光谱分析参数均能显著区分强化与非强化病变(P<0.001)。
  • 曲线形态目视分析诊断准确率为94。

诊断标准与关键数值

  • Slope HU40-70最佳阈值为读者1的3.0、读者2的3.0。
  • Slope HU40-100最佳阈值为读者1的1.8、读者2的1.9。
  • Slope HU40-140最佳阈值为读者1的1.2、读者2的1.15。
  • 使用斜率阈值时全部数据集诊断准确率为读者1的96、读者2的95。
  • ΔHU阈值20 HU时诊断准确率为100。

临床应用与阅片要点

  • 对1-4 cm偶发肾脏病变,可借助DECT虚拟单能量衰减曲线斜率或ΔHU阈值判断有无强化。
  • 斜率和曲线形态目视分析均可作为辅助判读工具,ΔHU阈值20 HU准确率最高。
  • 当未采集真正平扫时,光谱衰减曲线有助于弥补强化评估的不足。

局限性

  • 本研究为回顾性设计。
  • 扫描仅于门静脉期或肾实质期进行,未涵盖其他期相。

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

PURPOSE: To determine whether the spectral attenuation curve on a rapid kilovoltage-switching dual-energy computed tomography (DECT) scan can distinguish enhancing from nonenhancing incidental small (1-4 cm) renal lesions compared with conventional single-energy attenuation changes.

METHODS: This retrospective study enrolled 46 patients with 78 renal lesions (24 enhancing; 54 nonenhancing) who underwent DECT with DE mode performed during the portovenous or nephrographic phase. Final diagnosis of enhancing and nonenhancing masses was confirmed by pathology or imaging following the established criteria. Virtual monochromatic images (VMI) were reconstructed, and the slopes between the VMI dataset at 40-70 keV (Slope HU40-70), 40-100 keV (Slope HU40-100), and 40-140 keV (Slope HU40-140) were measured. Visual assessment of the curve pattern was recorded. Diagnostic accuracies were calculated with a cross-validated Mann-Whitney U test, and correlations of quantitative spectral parameters and intraclass correlation coefficient (ICC) were calculated using Spearman's rho correlation.

RESULTS: All quantitative and qualitative spectral analysis parameters significantly differentiated the enhancing and nonenhancing lesions (P < 0.001). The optimal slope thresholds calculated by cross-validation for Slope HU40-70, Slope HU40-100, and Slope HU40-140 were 3.0, 1.8 and 1.2, respectively for reader 1 and 3.0, 1.9 and 1.15, respectively for reader 2. Using a slope threshold at all datasets yielded a high diagnostic accuracy of 96 for reader 1 and 95 for reader 2. Using a ∆HU threshold of 20 HU yielded an accuracy of 100. Visual analysis of the curve pattern also yielded high accuracy of 94.

CONCLUSIONS: The spectral attenuation curve on rapid kilovoltage-switching DECT gives excellent diagnostic accuracy differentiating between incidental enhancing and nonenhancing renal lesions. This benefit of DECT will be most helpful when the true unenhanced phase is not performed.

来源:Abdom Radiol (NY),查看原文。版权归原作者所有。

引用

  1. Moleesaide, Maneegarn, Kaewlai, et al. Virtual monochromatic spectral attenuation curve analysis for evaluation of incidentally detected small renal lesions using rapid kilovoltage-switching dual-energy computed tomography. Abdom Radiol (NY). 2022. PMID: 35945346. doi:10.1007/s00261-022-03634-x

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。