共 50 条
Differentiation between primary CNS lymphoma and glioblastoma: qualitative and quantitative analysis using arterial spin labeling MR imaging
被引:38
作者:
You, Sung-Hye
[1
,2
]
Yun, Tae Jin
[1
,2
]
Choi, Hye Jeong
[1
,2
]
Yoo, Roh-Eul
[1
,2
]
Kang, Koung Mi
[1
,2
]
Choi, Seung Hong
[1
,2
]
Kim, Ji-Hoon
[1
,2
]
Sohn, Chul-Ho
[1
,2
]
机构:
[1] Seoul Natl Univ, Inst Radiat Med, Med Res Ctr, Seoul, South Korea
[2] Seoul Natl Univ Hosp, Dept Radiol, 101 Daehangno, Seoul 110744, South Korea
关键词:
Lymphoma;
Glioblastoma;
Brain neoplasms;
Magnetic resonance imaging;
Perfusion imaging;
NERVOUS-SYSTEM LYMPHOMA;
CONTRAST-ENHANCED-MRI;
BRAIN-TUMORS;
BLOOD-FLOW;
PERFUSION MRI;
MASS LESIONS;
GRADE;
GLIOMAS;
D O I:
10.1007/s00330-018-5359-5
中图分类号:
R8 [特种医学];
R445 [影像诊断学];
学科分类号:
1002 ;
100207 ;
1009 ;
摘要:
To evaluate the diagnostic performance of arterial spin labelling perfusion weighted images (ASL-PWIs) to differentiate primary CNS lymphoma (PCNSL) from glioblastoma (GBM). 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 (nCBF(intratumoral) and nCBF(peritumoral), 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. In qualitative analysis, scores 5 (CBFintratumoral > CBFGM, 68.8% [64/93]) and 4 (CBFintratumoral ae CBFGM, 47.6% [10/21]) were the most frequently reported scores for GBM and PCNSL, respectively. In quantitative analysis, both nCBF(intratumoral) and nCBF(peritumoral) in PCNSL were significantly lower than those in the GBM (nCBF(intratumoral), 0.89 +/- 0.59 [mean and SD] vs. 2.68 +/- 1.89, p < 0.001; nCBF(peritumoral), 0.17 +/- 0.08 vs. 0.45 +/- 0.28, p < 0.001). nCBF(peritumoral) demonstrated the best diagnostic performance (area under the ROC curve: visual scoring, 0.814; nCBF(intratumoral), 0.849; nCBF(peritumoral), 0.908; p < 0.001 for all). Interobserver agreements for visual scoring (weighted kappa = 0.869), nCBF(intratumoral_GM) (ICC = 0.958) and nCBF(peritumoral_GM) (ICC = 0.947) were all excellent. ASL-PWI performs well in differentiating PCNSL from GBM in both qualitative and quantitative analyses. ASL-PWI performs well (AUC > 0.8) in differentiating PCNSL from GBM. The visual scoring template demonstrated good diagnostic performance, similar to quantitative analysis. nCBF (peritumoral) demonstrated better diagnostic performance than nCBF (intratumoral) or visual scoring.
引用
收藏
页码:3801 / 3810
页数:10
相关论文
共 50 条