Chen Gang, Jespersen Sune, Pedersen Michael, Pang Qi, Horsman Michael R, Stødkilde Jørgensen Hans
MR Research Center, Aarhus University Hospital, Aarhus, DK8200 Aarhus N, Denmark.
Anticancer Res. 2005 Sep-Oct;25(5):3399-405.
Texture analysis can be used to identify and discriminate between objects in an image and has already been used for the characterization of diseases. The present study evaluated the potential of texture analysis for monitoring tumor cell viability during treatment with an anti-vascular drug.
Texture analysis of T2-weighted magnetic resonance images of C3H mammary carcinomas in mice was used to study tumor cell viability during treatment with the anti-vascular drug combretastatin A-4 disodium phosphate (CA4DP). Texture analysis was done by MaZda, and boot strap statistics were used to analyze the results.
The results revealed significant differences in the textures of T2-weighted images when non-tumor tissue and tumor tissue were compared. Furthermore, selected texture parameters were significantly different when image data from before and after CA4DP administration were analyzed.
Texture analysis is helpful in the diagnosis of different types of tissues, especially when these tissues can not be recognized by human observers. Some functional changes, such as perfusion changes in the tumor, can also be detected by texture analysis.
纹理分析可用于识别和区分图像中的物体，并且已经用于疾病的特征描述。本研究评估了纹理分析在使用抗血管药物治疗期间监测肿瘤细胞活力的潜力。
使用小鼠C3H乳腺癌T2加权磁共振图像的纹理分析来研究在用抗血管药物磷酸二钠康普瑞他汀A-4（CA4DP）治疗期间的肿瘤细胞活力。纹理分析由MaZda完成，并使用自展统计分析结果。
结果显示，当比较非肿瘤组织和肿瘤组织时，T2加权图像的纹理存在显著差异。此外，分析CA4DP给药前后的图像数据时，选定的纹理参数有显著差异。
纹理分析有助于不同类型组织的诊断，特别是当这些组织无法被人类观察者识别时。一些功能变化，如肿瘤中的灌注变化，也可以通过纹理分析检测到。
