快速时间分辨率动态对比增强MRI:直方图分析与视觉分析鉴别乳腺良恶性病变
本研究纳入2015年1月至2016年12月90例连续注册患者的105个乳腺病变(40个良性,65个恶性),行1.5-T动态对比增强MRI快速时间分辨率(10秒)全乳采集,比较直方图分析与视觉识别峰值强化分析基于经验数学模型得出的动力学参数鉴别良恶性的价值。两种分析基于α、A×α和AUC30均能成功鉴别良恶性病变,且两者ROC AUC无显著差异(p=0.21)。
要点
- 视觉识别峰值强化分析基于α、A×α和AUC30可鉴别乳腺良恶性病变。
- 直方图分析(第90、75百分位)与视觉分析基于经验数学模型的参数鉴别效能无显著差异(p=0.21)。
- 快速时间分辨率(10秒)动态对比增强MRI全乳采集可用于乳腺病变动力学评估。
核心结论
- 基于视觉识别峰值强化的动力学评估可用于鉴别乳腺良恶性病变。
研究设计与样本
- 2015年1月至2016年12月连续注册的90例患者共105个乳腺病变(40个良性,65个恶性),行1.5-T动态对比增强MRI,包括1次平扫和8次对比增强快速时间分辨率(10秒)全乳采集;采用经验数学模型ΔS(t)=A×(1-e-αt)拟合动力学曲线,计算初始斜率(A×α)和初始AUC(AUC30)。
诊断标准与关键数值
- 基于经验数学模型的α、A×α和AUC30,视觉识别峰值强化分析与直方图分析(第90和第75百分位)均能成功鉴别良恶性病变;两种分析的ROC AUC无显著差异(p=0.21)。
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: The purpose of this study was to validate a kinetic assessment based on visually identified peak enhancement, which is routinely used in clinical practice, for differentiating benign from malignant lesions during fast dynamic contrast-enhanced MRI.
MATERIALS AND METHODS: Between January 2015 and December 2016, 90 consecutively registered patients with 105 breast lesions (40 benign, 65 malignant) underwent dynamic contrast-enhanced 1.5-T MRI that included one unenhanced and eight contrast-enhanced fast temporal resolution (10 seconds) whole-breast acquisitions. Histogram analysis was performed to measure the voxel-based enhancement of the entire lesion to obtain 90th, 75th, and 50th percentile values at each time point and to generate kinetic curves. Two observers selected visually identified peak enhancement within the lesions to generate the kinetic curves. The kinetic curves from histogram and visually identified peak enhancement analyses were fitted by means of an empiric mathematic model (EMM): ΔS(t) = A × (1 - e-αt), where A is the upper limit of signal intensity, e indicates the exponential function, and α (min-1) is the rate of increase in signal intensity. The initial slope of the kinetic curve (A × α) and the initial AUC (AUC30) were calculated. These parameters were compared between benign and malignant lesions, and results from visually identified peak enhancement analysis were compared with those from histogram analysis.
RESULTS: Benign lesions were successfully differentiated from malignant lesions in both visually identified peak enhancement and histogram analyses (90th and 75th percentile values) on the basis of α, A × α, and AUC30 from the EMM. There was no significant difference in ROC AUC in these EMM parameters between visually identified peak enhancement and histogram analyses (p = 0.21).
CONCLUSION: Kinetic assessment with visually identified peak enhancement was acceptable for differentiating benign from malignant lesions.
引用
- Mori N, Pineda FD, Tsuchiya K, et al. Fast Temporal Resolution Dynamic Contrast-Enhanced MRI: Histogram Analysis Versus Visual Analysis for Differentiating Benign and Malignant Breast Lesions. AJR Am J Roentgenol. 2018. PMID: 30063374. doi:10.2214/AJR.17.19225
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。