磁共振成像纹理特征分析在胰腺浆液性囊腺瘤及黏液性囊腺瘤鉴别诊断中的价值

doi:10.3881/j.issn.1000-503X.2018.02.008

Acta Academiae Medicinae Sinica

Acta Academiae Medicinae Sinica ›› 2018, Vol. 40 ›› Issue (2): 187-193.doi: 10.3881/j.issn.1000-503X.2018.02.008

• Original Articles • Previous Articles Next Articles

Value of Magnetic Resonance Imaging Texture Feature Analysis in the Differential Diagnosis between Pancreatic Serous Cystadenoma and Mucinous Cystadenoma

WANG Botao¹, HE Lei¹, LIU Gang², LIU Mengqi^{1, 2}, CHEN Zhiye^{1, 2}()

¹Department of Radiology,Hainan Branch of Chinese PLA General Hospital,Sanya,Hainan 572013,China
²Department of Radiology,Chinese PLA General Hospital,Beijing 100853,China

Received:2017-07-25 Online:2018-04-28 Published:2018-05-03
About author:The first two author contributed equally to this article

Abstract

Abstract:

Objective To investigate the difference of tumor textures presented on T1-weighted image(T1WI) and T2-weighted image(T2WI) between pancreatic serous cystadenoma and mucinous cystadenoma. Methods Subjects including 24 patients with serous cystadenoma and 20 patients with mucinous cystadenoma who were treated in Hainan Branch of Chinese PLA General Hospital were enrolled in this study.The texture analysis was performed on T1WI and T2WI,and the texture parameters included angular second moment,contrast,correlation,inverse difference moment(IDM),and entropy.Independent sample t-test and Mann-Whitney U test were performed for the between-group comparison,the regression model was established by Binary Logistic regression analysis,and receiver operating characteristics curve(ROC) was drawn.Results The contrast and IDM of T1WI and the contrast,correlation,and IDM of T2WI showed significant differences between serous cystadenoma and mucinous cystadenoma(T1WI∶ $P angular second moment$ =0.714,P_contrast=0.003,P_correlation=0.069,P_IDM=0.004,P_entropy=0.818;T2WI∶ $P angular second moment$ =0.122,P_contrast=0.000,P_correlation=0.000,P_IDM=0.024,and P_entropy=0.081).The area under the ROC curve was 0.736-0.846 for the positive texture variables,and the area under the ROC curve was 0.879 for the variables in the model of Logistic regression.Binary Logistic regression analysis demonstrated that the contrast was a specific imaging variable for the differential diagnosis between serous cystadenoma and mucinous cystadenoma.Conclusion The texture analysis of T1WI and T2WI may be a simple and effective tool in the differential diagnosis between pancreatic serous cystadenoma and mucinous cystadenoma.

Key words: pancreas, serous cystadenoma, mucinous cystadenoma, texture analysis, magnetic resonance imaging

CLC Number:

R445.2

WANG Botao, HE Lei, LIU Gang, LIU Mengqi, CHEN Zhiye. Value of Magnetic Resonance Imaging Texture Feature Analysis in the Differential Diagnosis between Pancreatic Serous Cystadenoma and Mucinous Cystadenoma[J].Acta Academiae Medicinae Sinica, 2018, 40(2): 187-193.

Figures/Tables 7

Fig 1

Table 1

Comparisons of demographic and clinicopathological features between 24 patients with serous cystadenoma and 20 patients with mucinous cystadenoma"

分组Group	女/男^a Female/male^a(n)	年龄( $x ?$ ±s,岁)^a Age( $x ?$ ±s,years)^a	肿瘤直径[中位数 (四分位数间距),cm]^b Tumor size [median(QR),cm]^b	单囊型/多囊型^a Single cystic/ Polycystic^a (n)	胰腺头颈/体尾部^a Head and neck of pancreas/ body and tail of pancreas^a (n)
浆液性囊腺瘤 Serous cystadenoma(n=24)	19/5	51.9±14.4	2.8(1.8)	5/19	12/12
黏液性囊腺瘤 Mucinous cystadenoma(n=20)	15/5	49.1± 8.5	3.6(1.7)	6/14	6/14
χ²/U	0.108	0.809	142.500	0.122	1.072
P	0.743	0.423	0.021	0.727	0.300

Table 1

Table 2

Comparisons of five texture parameters of T1WI between 24 patients with serous cystadenoma and 20 patients with mucinous cystadenoma"

分组Group	角二阶矩 [中位数(四分位数间距)]^a Angular second moment [median(QR)]^a	对比度[中位数 (四分位数间距)]^a Contrast [median(QR)]^a	自相关 [中位数(四分位数间距)]^a Correlation[median(QR)]^a	逆差距 Inverse difference moment( $x ?$ ±s)	熵Entropy ( $x ?$ ±s)
浆液性囊腺瘤 Serous cystadenoma (n=24)	0.006(0.010)	37.411(53.260)	0.006(0.020)	0.252±0.087	5.155±0.817
黏液性囊腺瘤 Mucinous cystadenoma (n=20)	0.008(0.010)	13.219(25.710)	0.017(0.020)	0.342±0.112	5.204±0.520
U/t	224.500	115.000	163.000	3.025	0.231
P	0.714	0.003	0.069	0.004	0.818
95% CI	0.007~0.012	24.457~45.922	0.011~0.019	0.260~-0.326	4.968~5.388

Table 2

Table 3

Table 4

Comparisons of five texture parameters of T2WI between 24 patients with serous cystadenoma and 20 patients with mucinous cystadenoma"

分组Group	角二阶矩[中位数 (四分位数间距)]^a Angular second moment [median(QR)] ^a	对比度[中位数 (四分位数间距)]^a Contrast [median(QR)] ^a	自相关[中位数 (四分位数间距)]^a Correlation [median(QR)] ^a	逆差距 Inverse difference moment( $x ?$ ±s)	熵[中位数 (四分位数间距)]^a Entropy [median(QR)] ^a
浆液性囊腺瘤 Serous cystadenoma(n=24)	0.005(0.010)	114.194(98.540)	0.002(0.000)	0.297±0.248	5.714(1.420)
黏液性囊腺瘤 Mucinous cystadenoma(n=20)	0.008(0.030)	14.714(25.320)	0.009(0.040)	0.456±0.192	5.320(1.720)
U/t	174.500	74.000	86.000	2.345	166.000
P	0.122	0.000	0.000	0.024	0.081
95% CI	0.024~0.135	57.201~106.150	0.006~-0.021	0.300~-0.441	4.346~5.323

Table 4

Table 5

Fig 2

References 17

[1]	Adsay NV.Cystic neoplasia of the pancreas:pathology and biology[J]. J Gastrointest Surg,2008,12(3):401-404. DOI:10.1007/s11605-007-0348-z.
[2]	张蕊,彭晓刚,赵雪峰,等. 磁共振在胰腺囊性肿瘤鉴别诊断中的价值[J]. 中国临床医学影像杂志,2016,27(2):110-113.
[3]	刘润幸. 使用SPSS作多变量观察值的ROC曲线分析[J]. 中国公共卫生,2003,19(9):1151-1152. DOI:10.3321/j.issn:1001-0580.2003.09.078.
[4]	Bi Y,Xie BJ.The diagnostic value of multislice CT in cystic neoplasms of the pancreas[J]. Chin J Gastroenterol Hepatol,2014,23(2):234-237.
[5]	赵义军,耿小平. 胰腺囊腺瘤的诊治进展[J]. 肝胆外科杂志,2015,23(2):152-154.
[6]	周晓霞,张景,刘卓,等. 基于MR图像女性阿尔茨海默症海马纹理特征研究[J]. 北京生物医学工程,2010,29(3):230-234. DOI:10.3969/j.issn.1002-3208.2010.03.03.
[7]	Bastidas-Rodriguez MX,Prieto-Ortiz FA,Espejo E.Fractographic classification in metllic materials by using computer vision[J]. Eng Fail Anal,2016,59:237-252. DOI:10.1016/j.engfailanal.2015.10.008.
[8]	Gao S,Peng Y,Guo H,et al.Texture analysis and classification of ultrasound liver images[J]. Biomed Mater Eng,2014,24(1):1209-1216.
[9]	Zhang Y.MRI texture analysis in multiple sclerosis[J]. Int J Biomed Imaging,2012,2012(2):762804. DOI:10.1155/2012/762804.
[10]	Chen Z,Chen X,Liu M,et al.Texture features of periaqueductal gray in the patients with medication-overuse headache[J]. J Headache Pain,2017,18(1):14. DOI:10.1186/s10194-017-0727-0.
[11]	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. DOI:10.3881/j.issn.1000-503X.2017.02.002.
[12]	Holli K.Texture analysis as a tool for tissue characterization in clinical MRI[J]. Plos One,2011,8(12):e84744-e84744.
[13]	刘延彤,李茂星,叶建军. 计算机辅助诊断在医学影像诊断中的新进展[J]. 西北国防医学杂志,2006,27(6):450-451. DOI:10.3969/j.issn.1007-8622.2006.06.017.
[14]	薄华,马缚龙,焦李成. 图像纹理的灰度共生矩阵计算问题的分析[J]. 电子学报,2006,34(1):155-158,134. DOI:10.3321/j.issn:0372-2112.2006.01.032.
[15]	郝敬军,赵林,庄伟雄,等. 胰腺浆液性囊腺瘤与黏液性囊腺瘤的CT诊断与鉴别[J]. 医学影像学杂志,2017,27(2):273-276.
[16]	Tixier F,Hatt M,Le Rest CC,et al.Reproducibility of tumor uptake heterogeneity charact-erization through textural feature analysis in 18F-FDG PET[J]. J Nucl Med,2012,53(5):693-700. DOI:10.2967/jnumed.111.099127.
[17]	孙振球. 医学统计学 [M]. 2版. 北京:人民卫生出版社,2005:663-667.

参数Parameter	曲线下面积 Area under curve	临界值 Critical value	95% CI	敏感性 Sensitivity	特异性 Specificity	临界值以上 Above critical value	临界值以下 Below critical value
T2WI
对比度Contrast	0.846	40.891	58.284-108.299	0.800	0.875	浆液性囊腺瘤 Serous cystadenoma	黏液性囊腺瘤 Mucinous cystadenoma
自相关Correlation	0.821	0.008	0.000- 0.136	0.550	0.917	黏液性囊腺瘤 Mucinous cystadenoma	浆液性囊腺瘤 Serous cystadenoma
逆差距 Inverse difference moment	0.746	0.248	0.306- 0.460	0.900	0.708	黏液性囊腺瘤 Mucinous cystadenoma	浆液性囊腺瘤 Serous cystadenoma
T1WI
对比度Contrast	0.760	13.578	25.097- 48.200	0.550	0.917	浆液性囊腺瘤 Serous cystadenoma	黏液性囊腺瘤 Mucinous cystadenoma
逆差距 Inverse difference moment	0.736	0.317	0.252- 0.350	0.650	0.792	黏液性囊腺瘤 Mucinous cystadenoma	浆液性囊腺瘤 Serous cystadenoma
Logistic回归模型 Model of Logistic regression	0.879	-	0.359- 0.554	0.900	0.792	-	-

自变量 Independent variable	回归系数 Regression coefficient	标准误 Standard error	Wald χ²	P
对比度Contrast(T2WI)	-0.019	0.007	7.650	0.006
对比度Contrast(T1WI)	-0.040	0.019	4.571	0.033
常数Constant	2.418	0.850	8.086	0.004

Acta Academiae Medicinae Sinica

Home

About Journal

Current Issue

Archive

Search

Contact Us

中文

Value of Magnetic Resonance Imaging Texture Feature Analysis in the Differential Diagnosis between Pancreatic Serous Cystadenoma and Mucinous Cystadenoma

RichHTML

PDF (PC)

Like

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 7

References 17

Related Articles 15

Metrics

Recommended 10

[1]	CAI Ning,FU Qiang,ZHANG Yanyang,YU Xinguang. Effect of Flurbiprofen Axetil on Low-frequency Fluctuation Amplitude of Resting-state Functional Magnetic Resonance Imaging in Trigeminal Neuralgia [J]. Acta Academiae Medicinae Sinica, 2019, 41(2): 228-233.
[2]	LIU Mengqi,LIU You,CHEN Zhiye. Ependymoma in Sellar Region:Report of One Case [J]. Acta Academiae Medicinae Sinica, 2019, 41(1): 139-142.
[3]	MA Hongli,Lü Furong,XIAO Zhibo,DU Jinchao,SHENG Bo. Comparison of Prenatal Diagnostic Value between Fast Imaging Employing Steady State Acquisition and Single Shot Fast Spin Echo Sequence in Diagnosis of Placental Invasion [J]. Acta Academiae Medicinae Sinica, 2019, 41(1): 86-92.
[4]	XU Jia,WANG Xuan,YOU Yan,XUE Huadan,WANG Qin,WANG Shitian,JIN Zhengyu. Gadolinium Ethoxybenzyl Diethylene-triamine-pentaacetic-acid-enhanced Magnetic Resonance Imaging Perfusion and Multiple Hepatobiliary-phase Imaging for Quantitative Assessment of Liver Fibrosis in Rat Models [J]. Acta Academiae Medicinae Sinica, 2018, 40(6): 809-816.
[5]	WANG Botao,LIU You,WANG Kun,XU Huan,CHEN Zhiye. Magnetic Resonance Imaging Features of the Supratentorial Extra-axial Hemangioblastoma:A Case Report [J]. Acta Academiae Medicinae Sinica, 2018, 40(6): 847-849.
[6]	XIE Xiaohong,LIU Mengqi,LI Baige,CHEN Zhiye. Spinal Schwannoma with Entering-and-exiting-nerve Sign:Report of One Case [J]. Acta Academiae Medicinae Sinica, 2018, 40(6): 850-852.
[7]	SU Baiyan,ZHOU Kang,SHI Haifeng,SUN Hao,ZHANG Gumuyang,XUE Huadan,JIN Zhengyu. Predictive Value of Texture Analysis in the Treatment of Magnetic Resonance-guided Focused Ultrasound Surgery for Symptomatic Uterine Fibroids [J]. Acta Academiae Medicinae Sinica, 2018, 40(5): 673-679.
[8]	XU Aqiao,ZHAO Zhenhua,YANG Jianfeng,WEI Mingzhu,ZHENG Jing,LU Junmin,HU Hongjie,ZHANG Shengjian. Evaluation of Uterine Tumor Angiogenesis with Quantitative Magnetic Resonance Imaging Perfusion Parameters [J]. Acta Academiae Medicinae Sinica, 2018, 40(5): 705-709.
[9]	LI Juan,CHANG Xiaoyan,ZHU Liang,DAI Menghua,XUE Huadan. Visually Isoattenuating Pancreatic Acinar Cell Carcinoma:Report of One Case [J]. Acta Academiae Medicinae Sinica, 2018, 40(5): 714-718.
[10]	LI Lihui,HUANG Houbin,CHEN Zhiye. Contrast Enhanced T2 Fluid-attenuated Inversion Recovery Imaging in Diagnosing Macular Edema Caused by Retinal Vein Occlusion:Report of One Case [J]. Acta Academiae Medicinae Sinica, 2018, 40(5): 719-722.
[11]	CUI Bao, SUN Xingwang, ZHU Yufen, DU Yuping, CUI Jinguo, MA Lu, CAI Jianming. Relationship between Angiotensin Converting Enzyme Gene Polymorphism and Carotid Atherosclerotic Plaque: A Study Based on Vessel Wall Magnetic Resonance Imaging [J]. Acta Academiae Medicinae Sinica, 2018, 40(4): 493-500.
[12]	YANG Shuai, WANG Xiaoyi, LIAO Weihua, XIAO Haiqing, ZHOU Gaofeng, HU Ping, PENG Xianjing. Magnetic Resonance Imaging Features and Their Pathological Mechanisms of Hashimoto’s Encephalopathy [J]. Acta Academiae Medicinae Sinica, 2018, 40(4): 501-509.
[13]	WU Peng, ZHANG Yujiao, QIN Fang, LIU Lan, CUI Shujun. Diffusion Tensor Imaging in Early Diagnosis and Prognostic Prediction of Intervertebral Disc Change in Patients with Brucella Spondylitis [J]. Acta Academiae Medicinae Sinica, 2018, 40(4): 519-527.
[14]	ZHOU Ping, YAN Chaoqun, ZHANG Shuai, HUO Jianwei, LIU Cunzhi. Research Advances in Insomnia Disorder Based on Magnetic Resonance Imaging [J]. Acta Academiae Medicinae Sinica, 2018, 40(4): 543-548.
[15]	XU Huan, ZHANG Xiaohuan, WANG Kun, FAN Wenping, WANG Xue, YUN Jie, CHEN Zhiye. Diagnostric Value of Contrast-enhanced T2 Fluid-attenuated Inversion Recovery Imaging for Splenic Hemangioma:A Case Report [J]. Acta Academiae Medicinae Sinica, 2018, 40(4): 573-576.