• 作者：A.J. Revert Ventura^a ; ^{ajrevert@telefonica.net} ; R. Sanz-Requena^b ; L. Martí ; -Bonmatí ; ^b ; ^c ; J. Jornet^d ; J. Piquer^e ; A. Cremades^f ; J.M. Carot^g
• 关键词：Perfusió ; n RM ; Perfusió ; n cerebral ; Glioblastoma ; Aná ; lisis de histogramas
• 刊名：Radiolog铆a
• 出版年：2010
• 期刊代码：pca298_00338338
• 类别：med
• 出版时间：September-October 2010
• 卷：52
• 期：5
• 页码：432-441
• 文件大小：347 K

Objectives

To classify the tumor areas in patients with grade IV astrocytoma by calculating and statistically analyzing quantitative MRI perfusion parameters.

Material and methods

We applied two models of MRI perfusion, the unicompartmental and the pharmacokinetic models, in 15 patients diagnosed with grade IV astrocytoma. In the unicompartmental model, we quantified cerebral blood volume (CBV), mean transit time (MTT), and cerebral blood flow (CBF). In the pharmacokinetic model, we measured the permeability constant (K^trans), the extraction coefficient (k_ep), the fraction of the volume in the interstitial space (v_e), the fraction of the volume in the vessels (v_p), the permeability in the first pass (K_fp), and the vascular volume in the first pass (v_pfp). For each parameter, histograms were obtained for the total tumor area, for the peritumoral area, and for the healthy tissue. The statistical analysis included an analysis of variance for each parameter and a discriminant analysis.

Results

The most significant differences between the regions were obtained with CBV, CBF, K^trans, and v_pfp; of these, CBV had the best results. The best classificatory function on the discriminant analysis was the combination of K^trans and CBV. The analysis of the shape of the histogram showed statistically significant differences for the kurtosis of K^trans and k_ep, as well as for the skewness of CBV, CBF, K^trans, and v_pfp.

Conclusion

When parameters are considered individually, CBV is the one that best enables differentiation between tumor, peritumoral, and healthy tissue. The classificatory function generated from CBV and K^trans results in improved classification by areas.