基于纹理分析的儿童室管膜瘤和髓母细胞瘤的鉴别诊断
|
摘要
目的通过对儿童后颅窝常见肿瘤中室管膜瘤和髓母细胞瘤的MRI图像进行基于Gabor滤波的纹理分析,用支持向量机(SVM)对提取的特征进行训练分类,并对分类结果进行评价。资料与方法选取22例室管膜瘤和23例髓母细胞瘤图像的肿瘤部分为感兴趣区(ROI),对ROI进行5个尺度8个方向的Gabor滤波,观察滤波后图像提取均值、对比度、熵、角度方向二阶矩4组共160个纹理特征,分析160个纹理特征在不同肿瘤之间的差异。利用SVM对具有显著性差异的纹理特征进行训练并分类。结果 Gabor滤波后提取的160个纹理特征中,114个特征在2种肿瘤间差异有统计学意义(P<0.05),利用SVM,室管膜瘤与髓母细胞瘤分类准确率达(87.03±4.22)%。结论基于Gabor滤波的纹理特征分析能够有效实现儿童后颅窝肿瘤中室管膜瘤和髓母细胞瘤的分类,可以作为一种临床诊断的辅助方法。
Purpose To analyze the textural features by Gabor filtering of MRI of ependymoma and medulloblastoma, common pediatric posterior fossa tumors, to train and classify the extracted features using support vector machine(SVM), and to evaluate the results of classification. Materials and Methods A total of 22 cases of ependymoma and 23 cases of medulloblastoma were selected. The area of tumor in the 45 tumor images was set as the region of interest(ROI), and were filtered by Gabor filtering at 5 scales and 8 directions. After filtering the images, the four groups of 160 texture features were extracted, including mean, con, ent, and asm. The differences in texture features between different tumors were analyzed. SVM was used to train and classify the texture features with statistical differences between the two tumors. Results Among the 160 features extracted by Gabor filtering, 114 features had statistically significant differences(P<0.05). The accuracy of classification of ependymoma and medulloblastoma by SVM was(87.03±4.22)%. Conclusion The texture analysis based on Gabor filtering can effectively classify ependymoma and medulloblastoma in pediatric posterior fossa tumors, and can be used as an auxiliary method for clinical diagnosis.
Purpose To analyze the textural features by Gabor filtering of MRI of ependymoma and medulloblastoma, common pediatric posterior fossa tumors, to train and classify the extracted features using support vector machine(SVM), and to evaluate the results of classification. Materials and Methods A total of 22 cases of ependymoma and 23 cases of medulloblastoma were selected. The area of tumor in the 45 tumor images was set as the region of interest(ROI), and were filtered by Gabor filtering at 5 scales and 8 directions. After filtering the images, the four groups of 160 texture features were extracted, including mean, con, ent, and asm. The differences in texture features between different tumors were analyzed. SVM was used to train and classify the texture features with statistical differences between the two tumors. Results Among the 160 features extracted by Gabor filtering, 114 features had statistically significant differences(P<0.05). The accuracy of classification of ependymoma and medulloblastoma by SVM was(87.03±4.22)%. Conclusion The texture analysis based on Gabor filtering can effectively classify ependymoma and medulloblastoma in pediatric posterior fossa tumors, and can be used as an auxiliary method for clinical diagnosis.
引文
[1]李娅,张勇,程敬亮.囊性成分在鉴别儿童髓母细胞瘤和室管膜瘤中的价值.中国医学影像学杂志,2017,25(3):196-197.
[2]Castellano G,Bonilha L,Li LM,et al.Texture analysis of medical images.Clin Radiol,2004,59(12):1061-1069.
[3]朱碧云,陈卉.医学图像纹理分析的方法及应用.中国医学装备,2013,10(8):77-81.
[4]Mallat SG.A theory for multiresolution signal decomposition:the wavelet representation.IEEE Trans Pattern Anal Mach Intell,1989,11(7):674-693.
[5]Gabor D.Theory of communication.IEEE,1946,93:429-457.
[6]Ahmadian A,Mostafa A,Abolhassani MD,et al.A texture classification method for diffused liver diseases using Gabor wavelets.Conf Proc IEEE Eng Med Biol Soc,2005,2(2):1567-1570.
[7]Ukil A.Support vector machine.Computer Science,2002,1(4):1-28.
[8]Siegel R,Naishadham D,Jemal A.Cancer statistics,2013.CA Cancer J Clin,2013,63(1):11-30.
[9]Fetit AE,Novak J,Rodriguez D,et al.Radiomics in paediatric neuro-oncology:a multicentre study on MRItexture analysis.NMR Biomed,2018,31(1):e3781.
[10]李娅,李蒙蒙,张勇,等.磁共振图像的纹理分析在儿童常见后颅窝肿瘤鉴别诊断中的应用.中华医学杂志,2016,96(23):1853-1856.
[11]Fetit AE,Novak J,Peet AC,et al.Three-dimensional textural features of conventional MRI improve diagnostic classification of childhood brain tumours.NMR Biomed,2015,28(9):1174-1184.
[12]Gutierrez DR,Awwad A,Meijer L,et al.Metrics and textural features of MRI diffusion to improve classification of pediatric posterior fossa tumors.Am J Neuroradiol,2014,35(5):1009-1015.
[2]Castellano G,Bonilha L,Li LM,et al.Texture analysis of medical images.Clin Radiol,2004,59(12):1061-1069.
[3]朱碧云,陈卉.医学图像纹理分析的方法及应用.中国医学装备,2013,10(8):77-81.
[4]Mallat SG.A theory for multiresolution signal decomposition:the wavelet representation.IEEE Trans Pattern Anal Mach Intell,1989,11(7):674-693.
[5]Gabor D.Theory of communication.IEEE,1946,93:429-457.
[6]Ahmadian A,Mostafa A,Abolhassani MD,et al.A texture classification method for diffused liver diseases using Gabor wavelets.Conf Proc IEEE Eng Med Biol Soc,2005,2(2):1567-1570.
[7]Ukil A.Support vector machine.Computer Science,2002,1(4):1-28.
[8]Siegel R,Naishadham D,Jemal A.Cancer statistics,2013.CA Cancer J Clin,2013,63(1):11-30.
[9]Fetit AE,Novak J,Rodriguez D,et al.Radiomics in paediatric neuro-oncology:a multicentre study on MRItexture analysis.NMR Biomed,2018,31(1):e3781.
[10]李娅,李蒙蒙,张勇,等.磁共振图像的纹理分析在儿童常见后颅窝肿瘤鉴别诊断中的应用.中华医学杂志,2016,96(23):1853-1856.
[11]Fetit AE,Novak J,Peet AC,et al.Three-dimensional textural features of conventional MRI improve diagnostic classification of childhood brain tumours.NMR Biomed,2015,28(9):1174-1184.
[12]Gutierrez DR,Awwad A,Meijer L,et al.Metrics and textural features of MRI diffusion to improve classification of pediatric posterior fossa tumors.Am J Neuroradiol,2014,35(5):1009-1015.