Font Size: a A A

Imaging Characteristics And Radiomics Features Of Breast Ductal Carcinoma In Situ:correlation With Clinical Risk Assessment

Posted on:2021-05-06Degree:MasterType:Thesis
Country:ChinaCandidate:M R ZhaoFull Text:PDF
GTID:2494306470977469Subject:Clinical Medicine
Abstract/Summary:PDF Full Text Request
Purpose:Analyzing the correlation between the X-ray,ultrasound and MRI features of ductal carcinoma in situ(DCIS)and its clinical risk,and investigating the feasibility of predicting the clinical risk gradation of DCIS by radiomics signatures and clinical MRI characteristics.Methods:A retrospective analysis of 129 patients(132 lesions)with DCIS and DCIS with microinvasion(DCIS-MI)confirmed by pathology from January 2015 to June 2019 was conducted.The X-ray,ultrasound and MRI features were analyzed systemically according to the Breast Imaging Reporting and Data System(BI-RADS).132 lesions were classified into high-risk group and low-risk group.The lesions were considered as positive results with diagnoses as BI-RADS 4B,4C,5.Using t test,chi-square test,and Mann-Whitney U test to compare the differences in imaging characteristics between the two groups.P<0.05 was considered statistically significant The ITK-SNAP software was used to manually segment the lesions in dynamic contrast-enhanced MRI(DCE-MRI)and extract the radiomics signatures.Comparing the sensitivity,specificity,accuracy and area under receiver operating characteristic curve(AUC)of the prediction methods based on radiomics signatures,clinical MRI characteristics and radiomics combined with clinical MRI features.Result:There were 48 low-risk group lesions(36.4%),and 84 high-risk group lesions(63.6%).The mean age of high-risk group and low-risk group patients was(45.8 ± 8.7)years old and(49.5 ± 9.9)years old,respectively.There was significant difference between the two groups(P=0.025).MRI was performed in 132 lesions of DCIS,mammography in 121 lesions and ultrasound in 131 lesions.X-ray,ultrasound and MRI all had higher diagnostic accuracy for high-risk group(P<0.05).In X-ray,66 cases(59.5%)of DCIS showed calcification.In terms of calcification morphology,the high-risk group showed fine pleomorphic,fine linear or fine-linear branching calcification(81.4%,35/43),while the low-risk group showed round,coarse heterogeneous and amorphous calcification(58.3%,14/23).In terms of calcification distribution,high-risk group were mostly distributed in line or segment(48.8,21/43),while low-risk group cases were mostly showed grouped distribution(69.6%,16/23).The differences were statistically significant in the above calcification features(P<0.05).Ultrasound showed internal vascularity accounted for about 76.7%(56/73)of the high-risk group and 45.5%(15/33)of the low-risk group,with a statistically significant difference(P=0.006).On MRI,the lesions of high-risk group were more extensive(P<0.001).The frequency of segmental enhancement of low-risk group and high-risk group was 22.7%(10/44)and 51.8%(43/83),respectively.The early internal enhancement patterns of high-risk group were mostly heterogeneous(81.9%,68/83),while the low-risk group was mostly clumped(40.9%,18/44).In high-risk group,type Ⅰ,type Ⅱ andⅢ TIC were demonstrated in 11(13.3%)cases and 72(86.7%)cases.While in low-risk group,14(31.8%)cases showed type Ⅰ,and 30 cases(68.2%)showed type Ⅱ and typeⅢ.All of the above MRI features were significant(P<0.05).All classifiers that predicted the risk of DCIS were trained and validated by tenfold cross-validation.Extracting 1163 radiomics signatures of the 132 lesions based on DCE-MRI.The following three methods were used in this study.The sensitivity,specificity,accuracy and AUC of the radiomics signatures were 0.74,0.93,0.80 and 0.78,respectively,of clinical MRI characteristics were 0.87,0.81,0.86 and 0.80,of radiomics combined with clinical MRI features were 0.74,0.98,0.84 and 0.83,respectively.Conclusions:There is a certain correlation between the X-ray,ultrasound and MRI features of DCIS and DCIS-MI with their clinical risk.Compared with the radiomics signatures and clinical MRI characteristics,the ability of radiomics combined with clinical MRI features to predict the clinical risk of DCIS is better.It is helpful to distinguish the high-risk group from the low-risk group and provide reliable imaging basis for clinical treatment,so as to avoid the over-diagnosis and over-treatment of low-risk DCIS,and to take active clinical treatment for high-risk DCIS as soon as possible.
Keywords/Search Tags:Ductal carcinoma in situ(DCIS), Magnetic Resonance Imaging(MRI), Radiomics, Prognosis, Prediction
PDF Full Text Request
Related items