CT纹理分析在误诊的实性肺结节鉴别诊断中的应用
|
摘要
目的探讨CT纹理分析方法在误诊的肺实性结节鉴别诊断中的应用价值。方法回顾性分析CT误诊、经手术和病理证实的89例肺实性结节患者资料,包括良性病变误诊为肺癌54例和肺癌误诊为良性病变35例。采用MaZda软件对患者的CT图像进行纹理分析,分别用3种纹理特征提取方法(Fisher系数,Fisher;分类错误概率联合平均相关系数,POE+ACC;交互信息,MI)选择出前10个最有鉴别意义的纹理特征参数。采用原始数据分析(raw data analysis,RDA)、主要成分分析(principal component analysis,PCA)、线性分类分析(linear discriminant analysis,LDA)和非线性分类分析(nonlinear discriminant analysis,NDA)评估3种特征提取方法和三联法(Fisher+POE+ACC+MI,FPM)鉴别良、恶性肺实性结节的错判率(misclassified rate,MCR)。结果 Fisher、POE+ACC和MI这3种纹理特征提取方法选择的特征参数鉴别良、恶性肺实性结节的MCR均较低,FPM法可进一步降低MCR,用LDA分析3种特征提取方法鉴别良、恶性肺结节的MCR最低;用LDA分析FPM法(LDA-FPM)可使MCR进一步降低至0。结论利用CT图像纹理分析的方法有助于对误诊的良、恶性实性肺结节进行鉴别。
Objective To investigate the value of CT texture analysis in differential diagnosis of misdiagnosed pulmonary solid nodules. Methods Eighty-nine patients with solid pulmonary nodules which were misdiagnosed by preoperative CT,confirmed by surgery and pathology,were retrospectively reviewed.Among them,54 cases of benign nodules were misdiagnosed as lung cancers and 35 cases of lung cancers misdiagnosed as benign diseases.Texture analysis was performed for CT images by extracting texture features using MaZda software.The feature selection methods included Fisher coefficient(Fisher),classification error probability combined with average correlation coefficients(POE+ACC),mutual information(MI) and the combination of the above 3 methods.These methods were used to identify the 10 most significant texture features in the discrimination of benign and malignant pulmonary solid nodules.The different statistical methods including raw data analysis(RDA),principal component analysis(PCA),linear discriminant analysis(LDA)and nonlinear discriminant analysis(NDA)were used to evaluate the misclassified rate(MCR)of the 3 methods and their combination(Fisher+POE+ACC+MI,FPM)for distinguishing between benign and malignant pulmonary nodules.Results The 3 texture feature selection methods(Fisher,POE+ ACC,MI)had a low MCR for distinguishing between benign and malignant pulmonary nodules.A combination of these 3 methods could further reduce MCR.Using LDA,a lowest MCR was achieved for the feature selection methods of Fisher,POE+ACC,MI,and even MCR=0 for FPM. Conclusions CT texture analysis can be used to distinguish benign and malignant pulmonary solid nodules with a good performance.
Objective To investigate the value of CT texture analysis in differential diagnosis of misdiagnosed pulmonary solid nodules. Methods Eighty-nine patients with solid pulmonary nodules which were misdiagnosed by preoperative CT,confirmed by surgery and pathology,were retrospectively reviewed.Among them,54 cases of benign nodules were misdiagnosed as lung cancers and 35 cases of lung cancers misdiagnosed as benign diseases.Texture analysis was performed for CT images by extracting texture features using MaZda software.The feature selection methods included Fisher coefficient(Fisher),classification error probability combined with average correlation coefficients(POE+ACC),mutual information(MI) and the combination of the above 3 methods.These methods were used to identify the 10 most significant texture features in the discrimination of benign and malignant pulmonary solid nodules.The different statistical methods including raw data analysis(RDA),principal component analysis(PCA),linear discriminant analysis(LDA)and nonlinear discriminant analysis(NDA)were used to evaluate the misclassified rate(MCR)of the 3 methods and their combination(Fisher+POE+ACC+MI,FPM)for distinguishing between benign and malignant pulmonary nodules.Results The 3 texture feature selection methods(Fisher,POE+ ACC,MI)had a low MCR for distinguishing between benign and malignant pulmonary nodules.A combination of these 3 methods could further reduce MCR.Using LDA,a lowest MCR was achieved for the feature selection methods of Fisher,POE+ACC,MI,and even MCR=0 for FPM. Conclusions CT texture analysis can be used to distinguish benign and malignant pulmonary solid nodules with a good performance.
引文
[1] MOMEN W,JOSEPH G,RANJIT G,et al.Evidence for the treatment of patients with pulmonary nodules:when is it lung cancer?ACCP evidence-based clinical practice guidelines (2nd edition) [J].Chest,2007,132(3):94S-107S.
[2] WANG YX,GONG JS,SUZUKI K,et al.Evidence based imaging strategies for solitary pulmonary nodule[J].J Thorac Dis,2014,6(7):872-887.
[3] BASHIR U,SIDDIQUE MMS,MCLEAN E,et al.Imaging heterogeneity in lung cancer:techniques,applications,and challenges[J].Am J Roentgenol,2016,207(3):534-543.
[4] METZ S,GANTER C,LORENZEN S,et al.Multiparametric MR and PET imaging of intratumoral biological heterogeneity in patients with metastatic lung cancer using voxel-by-voxel analysis[J].PLoS One,2015,10(7):e0132386.
[5] RAMI R,BOLEJACK V,CROWLEY J,et al.The IASLC lung cancer staging project:proposals for the revisions of the t descriptors in the forthcoming eighth edition of the TNM classification for lung cancer[J].J Thorac Oncol,2015,10(7):990-1003.
[6] 石晓磊,王明泉.DICOM图像格式与BMP图像格式的转换[J].微计算机信息,2010,26(26):195-197.
[7] 张伟竹,华婷,徐婷婷,等.常规MRI纹理分析鉴别乳腺良、恶性病变的价值初探[J].中华放射学杂志,2017,51(8):588-591.
[8] COLLEWET G,STRZELECKI M,MARIETTE F.Influence of MRI acquisition protocols and image intensity normalization methods on texture classification[J].Magn Reson Imaging,2004,22(1):81-91.
[9] MATERKA A.MaZda user’s manual[EB/OL].[2002-10-21].http://www.eletel.p.lodz.pl/mazda/download/mazda _manual.pdf.
[10] NG F,KOZQRSKI R,GANESHAN B,et al.Assessment of tumor heterogeneity by CT texture analysis:can the largest cross-sectional area be used as an alternative to whole tumor analysis?[J].Eur J Radiol,2013,82(2):342-348.
[11] DASH M,LIU H.Feature selection for classification[J].Intell Data Anal,1997,1(1):131-156.
[12] TRUONG MT,KO JP,ROSSI S,et al.Update in the evaluation of the solitary pulmonary nodule[J].Radiographics,2014,34(6):1658-1679.
[13] GURNEY JW,LYDDON DM,MCKAY JA.Determining the likelihood of malignancy in solitary pulmonary nodules with Bayesian analysis.Part Ⅱ.Application[J].Radiology,1993,186(2):415-422.
[14] ELBAYOUMI E,SILVESTRI GA.Bronchoscopy for the diagnosis and staging of lung cancer[J].Semin Respir Crit Care Med,2008,29(3):261-270.
[15] GANESHAN B,GOH V,MANDEVILLE HC,et al.Non-small cell lung cancer:histopathologic correlates for texture parameters at CT[J].Radiology,2013,266(1):326-336.
[16] MACKAY JW,MURRAY PJ,KASMAI B,et al.MRI texture analysis of subchondral bone at the tibialplateau[J].Eur Radiol,2016,26(9):3034-3045.
[17] MANNIL M,BURQSTALLER JM,THANABALASINGAM A,et al.Texture analysis of paraspinal musculature in MRI of the lumbar spine:analysis of the lumbar stenosis outcome study (LSOS) data[J].Skeletal Radiol,2018,47(7):947-954.
[18] YUAN M,ZHANG YD,PU XH,et al.Comparison of a radiomic biomarker with volumetric analysis for decoding tumour phenotypes of lung adenocarcinoma with different disease-specific survival[J].Eur Radiol,2017,27(11):4857-4865.
[19] DENNIE C,THORNHILL R,SETHIVIRMANI V,et al.Role of quantitative computed tomography texture analysis in the differentiation of primary lung cancer and granulomatous nodules[J].Quant Imaging Med Surg,2016,6(1):6-15.
[20] HARALICK RM,SHANMUGAM K,DINSTEINI,et al.Textural features for image classification[J].Trans Syst Man Cybernet,1973,3(6):610-621.
[21] CHAE HD,PARK CM,PARK SJ,et al.Computerized texture analysis of persistent part-solid ground-glass nodules:differentiation of pre-invasive lesions from invasive pulmonary adenocarcinomas[J].Radiology,2014,273(1):285-293.
[22] ERASMUS JJ,CONNOLLY JE,MCADAMS HP,et al.Solitary pulmonary nodules.Part Ⅰ.Morphologic evaluation for differentiation of benign and malignant lesions[J].Radiographics,2000,20(1):43-58.
[23] MCNITTGRAY MF,WYCKOFF N,SAYRE JW,et al.The effects of co-occurrence matrix based texture parameters on the classification of solitary pulmonary nodules imaged on computed tomography[J].Comput Med Imaging Graph,1999,23(6):339-348.
[24] YE JP,XIONG T.Computational and theoretical analysis of null space and orthogonal linear discriminant analysis[J].J Mach Learn Res,2006,7(6):1183-1204.
[2] WANG YX,GONG JS,SUZUKI K,et al.Evidence based imaging strategies for solitary pulmonary nodule[J].J Thorac Dis,2014,6(7):872-887.
[3] BASHIR U,SIDDIQUE MMS,MCLEAN E,et al.Imaging heterogeneity in lung cancer:techniques,applications,and challenges[J].Am J Roentgenol,2016,207(3):534-543.
[4] METZ S,GANTER C,LORENZEN S,et al.Multiparametric MR and PET imaging of intratumoral biological heterogeneity in patients with metastatic lung cancer using voxel-by-voxel analysis[J].PLoS One,2015,10(7):e0132386.
[5] RAMI R,BOLEJACK V,CROWLEY J,et al.The IASLC lung cancer staging project:proposals for the revisions of the t descriptors in the forthcoming eighth edition of the TNM classification for lung cancer[J].J Thorac Oncol,2015,10(7):990-1003.
[6] 石晓磊,王明泉.DICOM图像格式与BMP图像格式的转换[J].微计算机信息,2010,26(26):195-197.
[7] 张伟竹,华婷,徐婷婷,等.常规MRI纹理分析鉴别乳腺良、恶性病变的价值初探[J].中华放射学杂志,2017,51(8):588-591.
[8] COLLEWET G,STRZELECKI M,MARIETTE F.Influence of MRI acquisition protocols and image intensity normalization methods on texture classification[J].Magn Reson Imaging,2004,22(1):81-91.
[9] MATERKA A.MaZda user’s manual[EB/OL].[2002-10-21].http://www.eletel.p.lodz.pl/mazda/download/mazda _manual.pdf.
[10] NG F,KOZQRSKI R,GANESHAN B,et al.Assessment of tumor heterogeneity by CT texture analysis:can the largest cross-sectional area be used as an alternative to whole tumor analysis?[J].Eur J Radiol,2013,82(2):342-348.
[11] DASH M,LIU H.Feature selection for classification[J].Intell Data Anal,1997,1(1):131-156.
[12] TRUONG MT,KO JP,ROSSI S,et al.Update in the evaluation of the solitary pulmonary nodule[J].Radiographics,2014,34(6):1658-1679.
[13] GURNEY JW,LYDDON DM,MCKAY JA.Determining the likelihood of malignancy in solitary pulmonary nodules with Bayesian analysis.Part Ⅱ.Application[J].Radiology,1993,186(2):415-422.
[14] ELBAYOUMI E,SILVESTRI GA.Bronchoscopy for the diagnosis and staging of lung cancer[J].Semin Respir Crit Care Med,2008,29(3):261-270.
[15] GANESHAN B,GOH V,MANDEVILLE HC,et al.Non-small cell lung cancer:histopathologic correlates for texture parameters at CT[J].Radiology,2013,266(1):326-336.
[16] MACKAY JW,MURRAY PJ,KASMAI B,et al.MRI texture analysis of subchondral bone at the tibialplateau[J].Eur Radiol,2016,26(9):3034-3045.
[17] MANNIL M,BURQSTALLER JM,THANABALASINGAM A,et al.Texture analysis of paraspinal musculature in MRI of the lumbar spine:analysis of the lumbar stenosis outcome study (LSOS) data[J].Skeletal Radiol,2018,47(7):947-954.
[18] YUAN M,ZHANG YD,PU XH,et al.Comparison of a radiomic biomarker with volumetric analysis for decoding tumour phenotypes of lung adenocarcinoma with different disease-specific survival[J].Eur Radiol,2017,27(11):4857-4865.
[19] DENNIE C,THORNHILL R,SETHIVIRMANI V,et al.Role of quantitative computed tomography texture analysis in the differentiation of primary lung cancer and granulomatous nodules[J].Quant Imaging Med Surg,2016,6(1):6-15.
[20] HARALICK RM,SHANMUGAM K,DINSTEINI,et al.Textural features for image classification[J].Trans Syst Man Cybernet,1973,3(6):610-621.
[21] CHAE HD,PARK CM,PARK SJ,et al.Computerized texture analysis of persistent part-solid ground-glass nodules:differentiation of pre-invasive lesions from invasive pulmonary adenocarcinomas[J].Radiology,2014,273(1):285-293.
[22] ERASMUS JJ,CONNOLLY JE,MCADAMS HP,et al.Solitary pulmonary nodules.Part Ⅰ.Morphologic evaluation for differentiation of benign and malignant lesions[J].Radiographics,2000,20(1):43-58.
[23] MCNITTGRAY MF,WYCKOFF N,SAYRE JW,et al.The effects of co-occurrence matrix based texture parameters on the classification of solitary pulmonary nodules imaged on computed tomography[J].Comput Med Imaging Graph,1999,23(6):339-348.
[24] YE JP,XIONG T.Computational and theoretical analysis of null space and orthogonal linear discriminant analysis[J].J Mach Learn Res,2006,7(6):1183-1204.