Font Size: a A A

Value Of MRI Combined With Texture Analysis In Differentiating Breast Fibroadenoma From Invasive Ductal Carcinoma

Posted on:2020-10-22Degree:MasterType:Thesis
Country:ChinaCandidate:L H MengFull Text:PDF
GTID:2404330590456170Subject:Imaging and nuclear medicine
Abstract/Summary:PDF Full Text Request
Objective:(1)To explore the value of histogram combined with dynamic enhanced TIC curve in differentiating breast fibroadenoma from invasive ductal carcinoma.(2)Discuss the value of differential diagnosis between breast fibroadenoma and invasive ductal carcinoma based on gray level co-occurrence matrix and DCE-MRI.Methods:1.A retrospective analysis was made of 47 female patients(15 cases of breast fibroadenoma and 32 cases of invasive ductal carcinoma)who underwent DCE-MRI from January 2016 to June 2018 without any treatment and confirmed by pathology.Bilateral mammary plain and dynamic enhanced scanning were performed with8-channel phased array coil of PHILIPS Achieva 3.0T MR imager.Draw the time-signal intensity curve at the self-contained workstation and extract the texture feature parameters of gray histogram of T2 WI,enhanced scanning phase 1,3 and 10 images in MaZada software.Among them,the type of dynamic enhancement curve was compared between benign and malignant using2 test,and the parameters of gray histogram were compared between benign and malignant using two independent samples t test.Then the sensitivity,specificity and accuracy of the combined diagnosis were obtained for benign and malignant breast lesions.2.Same as above,select the enhanced scan first,third,and tenth images to be exported in DICOM format,extract the gray level co-occurrence matrix texture featureparameters in MaZada software,use two independent sample t test,and draw ROC on the statistically different texture feature parameters.Curve and calculate the area under the curve.Results:1.DCE-MR time-signal intensity curve comparison between breast fibroadenoma and invasive ductal carcinoma: TIC curve of breast fibroadenoma was mainly type I(66.7%)and invasive ductal carcinoma was mainly type III(75.0%).The difference was statistically significant(P < 0.05).The histogram parameters of breast fibroadenoma and invasive ductal carcinoma were compared: Kurtosis and Skewness of breast fibroadenoma had statistical significance in phase 3 of enhanced scanning(P=0.013,P=0.015,P<0.05),while the other parameters had no statistical significance in other sequences(P>0.05).Dynamic enhancement,histogram parameters and comparison of the two methods in the diagnosis of breast fibroadenoma and invasive ductal carcinoma: The specificity and accuracy of the combined diagnosis of breast fibroadenoma and invasive ductal carcinoma were higher than those of the single application,and the sensitivity was lower.2.Angular second moment(AngScMom)of breast fibroadenoma had statistical significance in phase 1 and phase 3 of enhanced scanning(P < 0.05);Entropy,Entrp and Dif Entropy had statistical significance in each phase(P < 0.05);Sum Varnc had statistical significance in phase 10 of enhanced scanning(P < 0.05),and there was no difference between the two groups in other parameters.The difference was statistically significant(P > 0.05).Conclusion:DCE-MRI combined with texture analysis technology is valuable in the diagnosis of breast fibroadenoma and invasive ductal carcinoma,and can provide more evidence for the diagnosis of breast cancer in clinic.
Keywords/Search Tags:Texture analysis, breast fibroadenoma, invasive ductal carcinoma, DCE-MRI
PDF Full Text Request
Related items