文献速递乳腺技术与安全MR

同时多层读出分段回波平面成像是否有助于预测乳腺癌分子亚型

Is simultaneous multi-slice readout-segmented echo-planar imaging valuable for predicting molecular subtypes of breast cancer?
PubMed 文献 · Eur J Radiol · 2022-03-02

回顾性纳入125例单发浸润性乳腺癌患者,术前行同时多层读出分段扩散加权回波平面成像,由两名放射科医师独立进行全肿瘤直方图与纹理分析。HER2阳性型ADC均值、中位数、第5百分位数高于luminal B型,三阴性型ADC偏度高于luminal A型;ADC第5百分位数在淋巴结阳性与阴性乳腺癌间有显著差异。

要点

  • 基于同时多层读出分段扩散加权成像的全肿瘤直方图参数与乳腺癌分子亚型相关。
  • HER2阳性型ADC均值、中位数、第5百分位数高于luminal B型。
  • 三阴性型ADC偏度高于luminal A型,联合参数鉴别非luminal与luminal型AUC为0.712。

核心结论

  • 基于同时多层读出分段扩散加权成像的全肿瘤直方图与纹理分析参数可能提供侵袭性乳腺癌分子亚型的生物学信息。
  • ADC第5百分位数在淋巴结阳性与淋巴结阴性乳腺癌间有显著差异。

研究设计与样本

  • 回顾性研究,纳入125例单发浸润性乳腺癌患者(平均年龄52.81±12.01岁,范围24-84岁),均在本机构行术前MRI(含同时多层读出分段扩散加权成像)并接受手术。
  • 由两名放射科医师独立在同时多层读出分段扩散加权成像的ADC图上进行全肿瘤直方图与纹理分析。
  • 扫描参数:TR 3800.0 ms,TE 90.0 ms,视野340 mm×206 mm,平面内矩阵244×148,层厚5 mm,读出分段数5,b=0、800 s/mm²,扫描时间3分钟,层面加速因子2。
  • 采用Mann-Whitney U检验或Kruskal-Wallis检验比较直方图与纹理参数,对差异显著的参数绘制ROC曲线并计算AUC、敏感度、特异度、阳性预测值、阴性预测值。

诊断标准与关键数值

  • HER2阳性型ADC第95百分位数高于luminal A型(p=0.016)。
  • HER2阳性型ADC均值(p=0.025)、ADC中位数(p=0.025)、ADC第5百分位数(p=0.041)均高于luminal B型。
  • 三阴性型ADC偏度高于luminal A型(p=0.021)。
  • HER2阳性型与luminal型间ADC均值、中位数、第5百分位数、第95百分位数差异有统计学意义,联合AUC为0.762(敏感度72.41%,特异度75.00%,阳性预测值27.27%,阴性预测值95.45%)。
  • 三阴性型ADC偏度高于luminal型(p=0.012),AUC为0.688(敏感度96.15%,特异度36.78%,阳性预测值31.25%,阴性预测值96.97%);非luminal型ADC偏度与第95百分位数高于luminal型(p=0.003、0.002),联合AUC为0.712(敏感度92.11%,特异度50.23%,阳性预测值41.18%,阴性预测值92.50%)。

临床应用与阅片要点

  • ADC第5百分位数预测腋窝淋巴结状态的AUC为0.659(敏感度65.38%,特异度64.86%)。

局限性

  • 原文未报告局限性。

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 investigate the feasibility of simultaneous multi-slice readout-segmented diffusion-weighted echo-planar imaging (SMS rs-EPI DWI) in predicting invasive breast cancer molecular subtypes using whole-tumor histogram and texture analyses.

METHODS: In our retrospective study, 125 patients (mean age, 52.81 ± 12.01 years; range, 24-84 years) with single invasive breast cancer who underwent preoperative MRI with SMS rs-EPI DWI and surgery at our institution were included. Two radiologists independently performed whole-tumor histogram and texture analyses on the apparentdiffusioncoefficient (ADC) map of SMS rs-EPI DWI (TR, 3800.0 ms; TE 90.0 ms; field of view, 340 mm × 206 mm; in-plane matrix, 244 × 148; section thickness, 5 mm; readout segments, 5, b = 0, 800 s/mm2; scan time, 3 mins; slice acceleration factor, 2). The Mann-Whitney U test or Kruskal-Wallis test were used to compare histogram and texture parameters to assess their potential in differentiating molecular subtypes of breast cancer and predicting axillary lymph node status. Receiver operating characteristic (ROC) curves were generated for parameters with significant differences, and the area under the curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated.

RESULTS: Compared to luminal A breast cancer, the HER2-positive breast cancer subtype showed higher ADC_95th percentile values (p = 0.016). In addition, HER2-positive breast cancer yielded significantly higher ADC_mean (p = 0.025), ADC_median (p = 0.025), and ADC_5th percentile (p = 0.041) values than luminal B breast cancer. The ADC_skewness value was higher for triple-negative breast cancer (TNBC) than Luminal A breast cancer (p = 0.021). The difference in ADC_mean, ADC_median, ADC_5th percentile and ADC_95th percentile values between the HER2-positive breast cancer and luminal types was statistically significant, with a combined AUC for these parameters of 0.762 (sensitivity 72.41%, specificity 75.00%, PPV 27.27 %, and NPV 95.45 %). The ADC_skewness value for TNBC was higher than for luminal types (p = 0.012) and yielded an AUC value of 0.688 (sensitivity 96.15%, specificity 36.78%, PPV 31.25%, and NPV 96.97 %). Additionally, significantly higher ADC_skewness and ADC_95th percentile values (p = 0.003, 0.002, respectively) were reported for non-luminal types than luminal types yielding a combined AUC value of 0.712 (sensitivity 92.11%, and specificity 50.23%, PPV 41.18%, and NPV 92.50%). The AUC value for ADC_5th percentile for axillary lymph node status prediction was 0.659 (sensitivity 65.38%, specificity of 64.86%).

CONCLUSIONS: The whole-tumor histogram and texture analyses parameters based on SMS rs-EPI DWI may provide biological information on aggressive molecular subtypes of breast cancer. In addition, ADC_5th percentile values were significantly different between lymph node-positive and lymph node-negative breast cancer.

来源:Eur J Radiol,查看原文。版权归原作者所有。

引用

  1. Li, Zhu, Sun, et al. Is simultaneous multi-slice readout-segmented echo-planar imaging valuable for predicting molecular subtypes of breast cancer? Eur J Radiol. 2022. PMID: 35259708. doi:10.1016/j.ejrad.2022.110232

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。