动脉自旋标记MR成像定性及定量分析鉴别原发性中枢神经系统淋巴瘤与胶质母细胞瘤
本研究分析经病理证实的原发性中枢神经系统淋巴瘤(PCNSL,n=21)与胶质母细胞瘤(GBM,n=93)的动脉自旋标记灌注加权图像(ASL-PWI),采用视觉评分和定量参数(肿瘤内及瘤周正常化CBF)进行鉴别。定量分析中PCNSL的nCBFintratumoral和nCBFperitumoral均显著低于GBM(p<0.001),nCBFperitumoral诊断效能最佳(AUC 0.908)。结论认为ASL-PWI在定性和定量分析中均能较好地区分PCNSL与GBM。
要点
- 视觉评分中GBM最常见为5分(肿瘤内CBF>灰质CBF,68.8%[64/93]),PCNSL最常见为4分(肿瘤内CBF≈灰质CBF,47.6%[10/21])。
- 定量分析PCNSL的nCBFintratumoral为0.89±0.59,显著低于GBM的2.68±1.89(p<0.001)。
- nCBFperitumoral诊断效能最佳(AUC 0.908),优于nCBFintratumoral(AUC 0.849)和视觉评分(AUC 0.814)。
核心结论
- ASL-PWI在定性和定量分析中均能较好地区分PCNSL与GBM,nCBFperitumoral诊断效能优于nCBFintratumoral和视觉评分。
研究设计与样本
- 回顾性分析经病理证实的PCNSL(n=21)和GBM(n=93)患者的ASL-PWI图像,采用视觉评分和定量参数(nCBFintratumoral和nCBFperitumoral)进行比较,以ROC曲线评估诊断准确性,采用加权kappa或ICC评估观察者间可靠性。
关键影像表现/征象
- 视觉评分中GBM最常见为5分(肿瘤内CBF>灰质CBF,68.8%[64/93]),PCNSL最常见为4分(肿瘤内CBF≈灰质CBF,47.6%[10/21])。
- 定量分析中PCNSL的nCBFintratumoral为0.89±0.59,GBM为2.68±1.89,PCNSL显著低于GBM(p<0.001)。
- PCNSL的nCBFperitumoral为0.17±0.08,GBM为0.45±0.28,PCNSL显著低于GBM(p<0.001)。
诊断标准与关键数值
- 视觉评分基于ASL-CBF图将肿瘤分为5个等级,5分为肿瘤内CBF>灰质CBF,4分为肿瘤内CBF≈灰质CBF。
- nCBFintratumoral和nCBFperitumoral分别以对侧灰质CBF进行正常化得出。
- 诊断PCNSL与GBM的AUC:视觉评分0.814,nCBFintratumoral 0.849,nCBFperitumoral 0.908(均p<0.001)。
- 观察者间一致性:视觉评分加权kappa=0.869,nCBFintratumoral ICC=0.958,nCBFperitumoral ICC=0.947。
临床应用与阅片要点
- ASL-PWI视觉评分与定量分析均可用于鉴别PCNSL与GBM,视觉评分诊断效能与定量分析相似。
- 瘤周正常化CBF(nCBFperitumoral)诊断效能优于肿瘤内正常化CBF和视觉评分。
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
OBJECTIVES: To evaluate the diagnostic performance of arterial spin labelling perfusion weighted images (ASL-PWIs) to differentiate primary CNS lymphoma (PCNSL) from glioblastoma (GBM).
METHODS: ASL-PWIs of pathologically confirmed PCNSL (n = 21) or GBM (n = 93) were analysed. For qualitative analysis, tumours were visually scored into five categories based on ASL-CBF maps. For quantitative analysis, normalised CBF values were derived by contralateral grey matter (GM) in intra- and peritumoral areas (nCBFintratumoral and nCBFperitumoral, respectively). Visual scoring scales and quantitative parameters from PCNSL and GBM were compared. In addition, the area under the receiver-operating characteristic (ROC) curve was used to determine the diagnostic accuracy of ASL-PWI for differentiating PCNSL from GBM. Weighted kappa or intraclass correlation coefficients (ICCs) were used to assess reliability between two observers.
RESULTS: In qualitative analysis, scores 5 (CBFintratumoral>CBFGM, 68.8% [64/93]) and 4 (CBFintratumoral ≈ CBFGM, 47.6% [10/21]) were the most frequently reported scores for GBM and PCNSL, respectively. In quantitative analysis, both nCBFintratumoral and nCBFperitumoral in PCNSL were significantly lower than those in the GBM (nCBFintratumoral, 0.89 ± 0.59 [mean and SD] vs. 2.68 ± 1.89, p < 0.001; nCBFperitumoral, 0.17 ± 0.08 vs. 0.45 ± 0.28, p < 0.001). nCBFperitumoral demonstrated the best diagnostic performance (area under the ROC curve: visual scoring, 0.814; nCBFintratumoral, 0.849; nCBFperitumoral, 0.908; p < 0.001 for all). Interobserver agreements for visual scoring (weighted kappa = 0.869), nCBFintratumoral_GM (ICC = 0.958) and nCBFperitumoral_GM (ICC = 0.947) were all excellent.
CONCLUSIONS: ASL-PWI performs well in differentiating PCNSL from GBM in both qualitative and quantitative analyses.
KEY POINTS: • ASL-PWI performs well (AUC > 0.8) in differentiating PCNSL from GBM. • The visual scoring template demonstrated good diagnostic performance, similar to quantitative analysis. • nCBFperitumoral demonstrated better diagnostic performance than nCBFintratumoral or visual scoring.
引用
- You SH, Yun TJ, Choi HJ, et al. Differentiation between primary CNS lymphoma and glioblastoma: qualitative and quantitative analysis using arterial spin labeling MR imaging. Eur Radiol. 2018. PMID: 29619520. doi:10.1007/s00330-018-5359-5
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。