增强CT动脉期图像纹理分析对膀胱乳头状瘤与低级别乳头状癌鉴别诊断的价值
|
摘要
目的探讨增强CT动脉期图像纹理分析方法在膀胱乳头状瘤与低级别乳头状癌鉴别诊断中的价值。方法收集经病理检查证实的膀胱乳头状瘤和膀胱低级别乳头状癌患者43例,分别在增强CT动脉期图像对每个病灶分析300个纹理特征参数,并选取其中最具鉴别意义的10个参数(包括:Ang Sc Mom、Contrast、Entropy、DifVarnc、Horzl_GLevNonU、45dgr_RLNonUni、135dr_RLNonUni、Wav En LL_s-1、Wav En HL_s-2和Wav En HL_s-3)。采用独立样本t检验、Mann-Whitney U检验和受试者工作特征曲线比较两组肿瘤之间的纹理特征参数差异及各纹理特征参数的诊断效能。结果所选取的最具鉴别意义的10个纹理特征参数在膀胱乳头状瘤和低级别乳头状癌组间比较差异均具有统计学意义(P均<0. 05)。对纹理特征参数的诊断效能进行分析,AUC介于0. 76-0. 89之间,准确率介于78. 5%-91. 4%之间,灵敏度介于69. 5%-91. 5%之间,特异度介于70. 6%-93. 0%之间。结论基于增强CT动脉期图像的纹理特征参数在膀胱乳头状瘤和低级别乳头状癌中存在差异,纹理特征分析作为一种客观、定量的分析方法可应用于膀胱乳头状瘤和低级别乳头状癌的鉴别诊断。
Objective To evaluate the value of enhanced CT arterial phase image texture analysis in the differential diagnosis of bladder papilloma and low grade papillary carcinoma. Methods A total of 43 patients with pathological-confirmed bladder papilloma and low grade papillary carcinoma were collected. 300 texture features were analyzed in each lesion,respectively. 10 of the most discriminative texture features including Ang Sc Mom,Contrast,Entropy,Dif Varnc,Horzl_GLevNonU,45 dgr_RLNonUni,135 dr_RLNonUni,Wav En LL_s-1,Wav En HL_s-2 and Wav En HL_s-3 were selected. The differences of texture features between two groups and the diagnostic efficacy of each texture feature were compared by independent sample t test,Mann-Whitney U test and receiver operating characteristic curve. Results The results showed that the 10 most significant texture features were significantly different between the two groups( P all <0. 05). The AUC of each texture feature was between 0. 76 and 0. 89,the accuracy was between 78. 5% and 91. 4%,and the sensitivity was between 69. 5% and 91. 5%, and the specificity was between 70. 6% and 93. 0%,respectively. Conclusions The texture features of enhanced CT arterial phase images are different in the diagnosis of bladder papilloma and low grade papillary carcinoma. As an objective and quantitative analysis method,texture features analysis can be used in the differential diagnosis of bladder papilloma and low grade papillary carcinoma.
Objective To evaluate the value of enhanced CT arterial phase image texture analysis in the differential diagnosis of bladder papilloma and low grade papillary carcinoma. Methods A total of 43 patients with pathological-confirmed bladder papilloma and low grade papillary carcinoma were collected. 300 texture features were analyzed in each lesion,respectively. 10 of the most discriminative texture features including Ang Sc Mom,Contrast,Entropy,Dif Varnc,Horzl_GLevNonU,45 dgr_RLNonUni,135 dr_RLNonUni,Wav En LL_s-1,Wav En HL_s-2 and Wav En HL_s-3 were selected. The differences of texture features between two groups and the diagnostic efficacy of each texture feature were compared by independent sample t test,Mann-Whitney U test and receiver operating characteristic curve. Results The results showed that the 10 most significant texture features were significantly different between the two groups( P all <0. 05). The AUC of each texture feature was between 0. 76 and 0. 89,the accuracy was between 78. 5% and 91. 4%,and the sensitivity was between 69. 5% and 91. 5%, and the specificity was between 70. 6% and 93. 0%,respectively. Conclusions The texture features of enhanced CT arterial phase images are different in the diagnosis of bladder papilloma and low grade papillary carcinoma. As an objective and quantitative analysis method,texture features analysis can be used in the differential diagnosis of bladder papilloma and low grade papillary carcinoma.
引文
[1] Yi X,Guan X,Chen C,et al. Adrenal incidentaloma:machine learning-based quantitative texture analysis of unenhanced CT can effectively differentiate s PHEO from lipid-poor adrenal adenoma[J]. J Cancer,2018,9(19):3577-3582.
[2]郭翌,周世崇,余锦华,等.影像组学的前沿研究与未来挑战[J].肿瘤影像学,2017,262(2):81-90.
[3] Ginsburg SB,Zhao J,Humphries S,et al. Texturebased quantification of centrilobular emphysema and centrilobular nodularity in longitudinal CT scans of current and former smokers[J]. Acad Radiol,2016,23(11):1349-1358.
[4]刘露,吴霜,伍兵. CT图像纹理分析鉴别肝上皮样血管内皮瘤与结肠癌肝转移瘤的初步研究[J].中国普外基础与临床杂志,2018,25(4):483-487.
[5]陈文静,牟玮,张文馨,等. MR动态增强图像纹理分析判断乳腺结节良恶性的价值[J].中国医学影像技术,2017,33(5):647-651.
[6]周海龙,张古沐阳,石冰,等.定量CT纹理分析鉴别诊断透明细胞型与非透明细胞型肾癌[J].中国医学影像技术,2017,33(12):1768-1733.
[7] Kocak B,Yardimci AH,Bektas CT,et al. Textural differences between renal cell carcinoma subtypes:Machine learning-based quantitative computed tomography texture analysis with independent external validation[J].Eur J Radiol,2018,107(8):149-157.
[8] Hodgdon T,McInnes MD,Schieda N,et al. Can quantitative CT texture analysis be used to differentiate fatpoor renal angiomyolipoma from renal cell carcinoma on unenhanced CT images Radiology,2015,276(3):787-796.
[9] Xu F,Ma X,Wang Y,et al. CT texture analysis can be a potential tool to differentiate gastrointestinal stromal tumors without KIT exon 11 mutation[J]. Eur J Radiol,2018,107(6):90-97.
[10] Lubner MG,Malecki K,Kloke J,et al. Texture analysis of the liver at MDCT for assessing hepatic fibrosis[J].Abdom Radiol(NY),2017,42(8):2069-2078.
[11]王县平,杨学辉,严海员,等.螺旋CT增强扫描在膀胱癌术前分期中的临床价值研究[J].中国CT和MRI杂志,2017,15(11):95-97.
[12] MochH,Humphrey PA,Ulbright TM,et al. WHO classification of tumours of the urinary system and male genital organ[M]. Lyon IARC Press,2016.
[13] Szczypiński P M, Strzelecki M, Materka A, et al. MaZda-A software package for image texture analysis. Comput Meth Prog Bio,2009,94(1):66-76.
[14] Guyon I,Elisseeff A. An introduction to variable and feature selection[J]. JMLR,2003,3(12):1157-1182.
[15]冷媛媛,黄燕涛,仲建全,等.多层螺旋CT在膀胱内翻性乳头状瘤中的诊断价值[J].山西医药杂志,2018,47(14):1660-1662.
[16]闫鹏飞.传统磁共振影像特征在脑膜瘤病理级别的术前鉴别中的作用:定性及定量方法探讨[D].武汉:华中科技大学,2017.
[17] Varghese BA,Chen F,Hwang DH,et al. Differentiation of Predominantly Solid Enhancing Lipid-Poor Renal Cell Masses by Use of Contrast-Enhanced CT:Evaluating the Role of Texture in Tumor Subtyping[J]. AJR Am J Roentgenol,2018,21(2):1-9.
[18]王永芹,黄子星,袁放,等. CT平扫图像纹理分析对肝癌与肝血管瘤鉴别诊断的初步研究[J].中国普外基础与临床杂志,2017,24(2):254-258.
[19]王波涛,刘刚,樊文萍,等.纹理特征分析在肝囊肿及肝血管瘤磁共振成像鉴别诊断中的价值[J].中国医学科学院学报,2017,39(2):169-176.
[20]陈文静,牟玮,张文馨,等. MR动态增强图像纹理分析判断乳腺结节良恶性的价值[J].中国医学影像技术,2017,33(5):647-651.
[2]郭翌,周世崇,余锦华,等.影像组学的前沿研究与未来挑战[J].肿瘤影像学,2017,262(2):81-90.
[3] Ginsburg SB,Zhao J,Humphries S,et al. Texturebased quantification of centrilobular emphysema and centrilobular nodularity in longitudinal CT scans of current and former smokers[J]. Acad Radiol,2016,23(11):1349-1358.
[4]刘露,吴霜,伍兵. CT图像纹理分析鉴别肝上皮样血管内皮瘤与结肠癌肝转移瘤的初步研究[J].中国普外基础与临床杂志,2018,25(4):483-487.
[5]陈文静,牟玮,张文馨,等. MR动态增强图像纹理分析判断乳腺结节良恶性的价值[J].中国医学影像技术,2017,33(5):647-651.
[6]周海龙,张古沐阳,石冰,等.定量CT纹理分析鉴别诊断透明细胞型与非透明细胞型肾癌[J].中国医学影像技术,2017,33(12):1768-1733.
[7] Kocak B,Yardimci AH,Bektas CT,et al. Textural differences between renal cell carcinoma subtypes:Machine learning-based quantitative computed tomography texture analysis with independent external validation[J].Eur J Radiol,2018,107(8):149-157.
[8] Hodgdon T,McInnes MD,Schieda N,et al. Can quantitative CT texture analysis be used to differentiate fatpoor renal angiomyolipoma from renal cell carcinoma on unenhanced CT images Radiology,2015,276(3):787-796.
[9] Xu F,Ma X,Wang Y,et al. CT texture analysis can be a potential tool to differentiate gastrointestinal stromal tumors without KIT exon 11 mutation[J]. Eur J Radiol,2018,107(6):90-97.
[10] Lubner MG,Malecki K,Kloke J,et al. Texture analysis of the liver at MDCT for assessing hepatic fibrosis[J].Abdom Radiol(NY),2017,42(8):2069-2078.
[11]王县平,杨学辉,严海员,等.螺旋CT增强扫描在膀胱癌术前分期中的临床价值研究[J].中国CT和MRI杂志,2017,15(11):95-97.
[12] MochH,Humphrey PA,Ulbright TM,et al. WHO classification of tumours of the urinary system and male genital organ[M]. Lyon IARC Press,2016.
[13] Szczypiński P M, Strzelecki M, Materka A, et al. MaZda-A software package for image texture analysis. Comput Meth Prog Bio,2009,94(1):66-76.
[14] Guyon I,Elisseeff A. An introduction to variable and feature selection[J]. JMLR,2003,3(12):1157-1182.
[15]冷媛媛,黄燕涛,仲建全,等.多层螺旋CT在膀胱内翻性乳头状瘤中的诊断价值[J].山西医药杂志,2018,47(14):1660-1662.
[16]闫鹏飞.传统磁共振影像特征在脑膜瘤病理级别的术前鉴别中的作用:定性及定量方法探讨[D].武汉:华中科技大学,2017.
[17] Varghese BA,Chen F,Hwang DH,et al. Differentiation of Predominantly Solid Enhancing Lipid-Poor Renal Cell Masses by Use of Contrast-Enhanced CT:Evaluating the Role of Texture in Tumor Subtyping[J]. AJR Am J Roentgenol,2018,21(2):1-9.
[18]王永芹,黄子星,袁放,等. CT平扫图像纹理分析对肝癌与肝血管瘤鉴别诊断的初步研究[J].中国普外基础与临床杂志,2017,24(2):254-258.
[19]王波涛,刘刚,樊文萍,等.纹理特征分析在肝囊肿及肝血管瘤磁共振成像鉴别诊断中的价值[J].中国医学科学院学报,2017,39(2):169-176.
[20]陈文静,牟玮,张文馨,等. MR动态增强图像纹理分析判断乳腺结节良恶性的价值[J].中国医学影像技术,2017,33(5):647-651.