ABSTRACT
Objective:To explore the value of radiomics features extracted from multi-parameter MRI (mp-MRI) in preoperative prediction of triple negative breast cancer (TNBC) and non triple negative breast cancer (NTNBC).Methods:The clinical and preoperative-MRI data of 371 patients with breast cancer confirmed by surgical pathology from January 2017 to July 2019 in Henan Provincial People′s Hospital were retrospectively analyzed. Based on the results of immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) from postoperative pathological specimens, the cancer was classified as TNBC and NTNBC. Patients were randomly assigned to a training set ( n=250) and a validation set ( n=121). Quantitative radiomics features were extracted from three-dimensional lesions based on dynamic contrast enhanced-T 1WI (DCE-T 1WI) and fat-suppressed T 2WI sequences, and 32 quantitative radiomics features were selected by Mann-Whitney U test, elastic network, and support vector machine recursive feature elimination. Three radiomics signatures were constructed by using the algorithm of support vector machine based on the quantitative radiomics features extracted from fat-suppressed T 2WI, DCE-T 1WI and the mp-MRI of their combination. The prediction performances were evaluated by receiver operating characteristic (ROC) curve and the area under the ROC curve, accuracy, sensitivity, and specificity were calculated. Results:There were 61 patients with TNBC and 310 patients with NTNBC. The clinicopathological characteristics between NTNBC and TNBC were statistically different in the pathological grade (χ2=24.544, P<0.001). Other clinicopathological characteristics (age, maximum diameter of mass, vascular tumor thrombus, axillary lymph nodes) were not statistically differences between NTNBC and TNBC ( P>0.05). The radiomics signature presenting the best performance for predictive TNBC and NTNBC were based on mp-MRI radiomics features. The area under the ROC curve, accuracy, sensitivity, and specificity were 0.91[95% confidence interval (CI) 0.881-0.932], 86.0%, 84.4% and 86.3% in training set, and 0.84 (95%CI 0.807-0.868), 75.2%, 68.7% and 76.1%, in validation set, respectively. Conclusion:Radiomics based on mp-MRI features can be a effectively potential tool for predictive TNBC and NTNBC breast cancer and provide scientific basis for clinicians to make treatment decisions.
ABSTRACT
Objective:To explore the value of mammography-based radiomics for preoperative prediction of axillary lymph node metastasis in breast carcinoma.Methods:The clinical and X-ray data of female patients with pathologically confirmed breast cancer in Henan People′s Hospital from June 2013 to July 2017 were analyzed retrospectively. A total of 214 patients, aged 30-85 (53±11) years, were randomly divided into training set ( n=153) and verification set ( n=61) according to the ratio of 3∶1. According to pathological findings of the axillary lymph node metastasis, 99 cases were divided into positive group and 115 cases into negative group. The lesions were segmented and extracted in X-ray images of mediolateral oblique (MLO) and cranial caudal (CC). Three, nine and seven axillary lymph node metastasis related histologic features were selected from the high dimensional features of CC, MLO and CC combined MLO images by lasso regression model. According to the characteristics of imaging and clinical characteristics, the prediction model was constructed. The prediction ability of the model was verified by 10% cross validation. Results:The lymph node in positive group was larger than negative groups, the difference was statistically significant ( t=2.611, P<0.05). In the validation set, the area under curve (AUC) values of CC, MLO, CC combined with MLO images, clinical features and clinical features combined with CC and MLO images were 0.680, 0.723, 0.740, 0.558 and 0.714, respectively. Among them, CC combined with MLO images had the highest prediction efficiency, and AUC values were higher than CC alone, MLO images and CC combined with MLO images. Conclusions:Quantitative radiomics features of breast tumor extracted from digital mammograms are helpful for preoperatively predicting axillary lymph node metastasis. Future larger studies are needed to further evaluate these findings.
ABSTRACT
In order to solve the pathological grading of hepatocellular carcinomas (HCC) which depends on biopsy or surgical pathology invasively, a quantitative analysis method based on radiomics signature was proposed for pathological grading of HCC in non-contrast magnetic resonance imaging (MRI) images. The MRI images were integrated to predict clinical outcomes using 328 radiomics features, quantifying tumour image intensity, shape and text, which are extracted from lesion by manual segmentation. Least absolute shrinkage and selection operator (LASSO) were used to select the most-predictive radiomics features for the pathological grading. A radiomics signature, a clinical model, and a combined model were built. The association between the radiomics signature and HCC grading was explored. This quantitative analysis method was validated in 170 consecutive patients (training dataset: = 125; validation dataset, = 45), and cross-validation with receiver operating characteristic (ROC) analysis was performed and the area under the ROC curve (AUC) was employed as the prediction metric. Through the proposed method, AUC was 0.909 in training dataset and 0.800 in validation dataset, respectively. Overall, the prediction performances by radiomics features showed statistically significant correlations with pathological grading. The results showed that radiomics signature was developed to be a significant predictor for HCC pathological grading, which may serve as a noninvasive complementary tool for clinical doctors in determining the prognosis and therapeutic strategy for HCC.
Subject(s)
Humans , Carcinoma, Hepatocellular , Diagnostic Imaging , Liver Neoplasms , Diagnostic Imaging , Magnetic Resonance Imaging , Neoplasm Grading , Methods , ROC CurveABSTRACT
Objective To explore the feasibility of MR lymphography (MR-LG) in detecting internal mammary sentinel lymph node (SLN) for rabbit models of VX2 breast cancer.Methods Totally 75 New Zealand white rabbits were randomly divided into 2 groups to establish VX2 breast cancer (tumor group) and inflammation models (inflammation group),respectively.MR-LG examinations and lymph node dissection were performed in all the rabbits.The results of internal mammary SLN and lymphatic vessels mapping were recorded,and the influence factors were analyzed.Results The success rate of model establishment was 97.33％ (73/75).The detection rate of internal mammary SLN and lymphatic vessels was 15.94％ (11/69) and 75.36％ (52/69),respectively,there was no statistical difference between the two groups (both P＞0.05).The rabbit's weight,time of tumor growth,tumor size,number,size and status of axillary SLN,as well as presence or not of internal mammary lymph vessels detection showed no statistical difference between detection and non-detection of internal mammary SLN (all P＞0.05),while the tumor size and axillary lymph node number showed statistically significant differences between detection and not detection of internal lymphatic vessels (both P＜0.05).The accuracy,sensitivity,specificity,false negative and false positive rate of internal mammary SLN detection with MR-LG was 76.81％ (53/69),39.13％(9/23),95.65％(44/46),60.86％(14/23) and 4.35％(2/46),respectively.Conclusion MR-LG can effectively detect internal mammary SLN and lymphatic vessels for rabbit models of VX2 breast cancer.
ABSTRACT
Objective To study the value of CT lymphography (CT-LG) in evaluating axillary sentinel lymph nodes (SLN) of rabbit with breast cancer using 3D and curved planar reformation (CPR).Methods CT-LG examinations were completed on the VX2 rabbit models bearing breast cancer established by tumor tissue mass suspension injection.All raw data were transferred to a postprocessing workstation, then the 3D and CPR were used to show lymphatic vessel and lymph nodes.On the CT-LG images, one or several lymph nodes displayed firstly in the lymphatic drainage pathway was defined as the SLN which was then confirmed by the biopsy.Results (1) 20 rabbits bearing tumor underwnet CT-LG examinations, and both SLNs and draining lymphatic vessels were detected using 3D and CPR techniques.70% (14/20) of experimental rabbits had fine images, and 16 SLNs and 7 non-SLNs were demonstrated on 14 rabbits on CT-LG images with good quality.(2) 8 continuous lymphatic vessels displayed well, exhibiting a homogeneous and continuous tube with uniform contrast-medium filling on CPR images.Most lymphatic vessels demonstrated well, and the 14 vessels with local interrupt or fuzzy on 3D images showed continuous tubular contrast-filling with heterogeneous high-density on the CPR images from the injection sites to axillary SLN.Among 6 rabbits with poor images, 3 rabbits showed the SLN but without lymphatic vessels on 3D images, 2 showed only initial segment of lymphatic vessel,and 1 had no lymphatic vessel and SLN.(3) A total 24 SLNs were obtained by SLNB,including 18 metastatic SLNs and 6 negative ones.On CT-LG images, there were statistical differences between the metastatic and negative SLNs in the maximum diameter and filling defect (P=0.041 and P=0.001, respectively).However, the short diameter, shape and margin of SLNs had no statistical differences between the two groups (all P>0.05).Conclusion On CT-LG examinations, 3D and CPR images may show axillary SLNs and lymphatic vessels effectively,and CPR can demonstrate more details than 3D.The maximum diameter and fillingdefect of SLN are helpful for distinguishing metastatic and negative SLNs.
ABSTRACT
Objective To explore the value of mammogram texture analysis in prediction of metastasis of axillary lymph nodes for breast carcinoma.Methods Mammograms and clinical data of 171 patients with breast carcinoma confirmed by pathology were retrospectively analyzed,and all patients underwent axillary lymph node dissection (ALND).Then the patients were divided into axillary lymph node metastasis group and non-metastasis group according to the result of ALND.The texture features of these lesions were statistically analyzed,including gray-level histogram texture parameters (mean value,standard deviation,skewness,kurtosis and variance) and gray-level co-occurrence matrix texture parameters (energy,entropy,correlation,inertia,inverse difference moment and contrast).Results In all of 171 breast cancer patients,96patients had axillary lymph node metastasis,while 75 patients had no metastasis.Mammograms showed negative axillary lymph nodes in 119 patients and positive axillary lymph nodes in 52 patients,and the sensitivity and specificity of mammograms in the diagnosis of positive axillary lymph nodes was 48.96％ (47/96) and 93.33％ (70/75),respectively.Mammogram texture analysis showed the values of energy,entropy,inverse difference moment and correlation in axillary lymph node metastasis group were higher than those in non-metastasis group,while the values of inertia and contrast in the axillary lymph node metastasis group were lower than those in non-metastasis group (all P＜0.05).The rest texture parameters had no significant differences between two groups (all P＞0.05).Area under curve (AUC) for texture parameters of energy,entropy,inertia,inverse difference moment,correlation and contrast was 0.610,0.610,0.374,0.599,0.612 and 0.421 (all P＜0.05),respectively.AUC of mammography,mammogram texture features,and the combination of mammography and texture features was 0.711,0.676 and 0.787 (all P＜0.05),respectively.The sensitivity and specificity of mammogram texture features,the combination of mammography and texture features in diagnosis of axillary lymph nodes metastasis was 62.5％ and 64.6％,66.7％ and 82.7％,respectively.Conclusion Mammogram texture parameters are helpful for predicting axillary lymph node metastasis,and the combination of mammography and texture features can improve diagnostic efficiency of axillary lymph node metastasis.
ABSTRACT
Objective To study the radiological appearance and pathological features of breast phyllodes tumors(PTs), and to enhance the knowledge of the tumor.Methods Clinical and imaging findings were reviewed retrospectively in 65 women with PTs confirmed by surgical pathology.61 patients had preoperative MRI and/or sonography, and 44 had preoperative mammography.Then these findings were compared with the pathological results.Results (1)The percentages of benign, borderline and malignant PTs were 32.3%(21/65), 43.1%(28/65) and 24.6%(16/65), respectively,and the mean maximum diameters of these lesions were (4.75±2.62) cm, (5.37±3.11) cm and (4.96±3.82) cm, respectively.There were no statistical difference in lesion''s size among these three types(F=0.247,P=0.782).(2)On the mammograms of these 44 PTs(including 16 cases of benign type, 18 cases of borderline and 10 cases of malignant type), the appearances of radiolucent halo and calcification among these three types showed statistically significant differences(both P<0.05).(3)On the sonography of these 61 PTs(including 19 cases of benign type, 26 cases of borderline and 16 cases of malignant type), the shape among these three types showed statistically significant difference(P<0.05).(4)On the MRI of these 61 PTs(including 21 cases of benign type, 24 cases of borderline and 16 cases of malignant type), internal hypointense septations on T2WI and time intensity curves showed statistically significant differences among these three types(both P<0.05).Conclusion Some radiological appearances of PTs can be used to help determine the histologic subtypes of breast PTs preoperatively.
ABSTRACT
Objective To explore the MSCT manifestations of chromophobe renal cell carcinoma (CRCC)and to improve its accu-racy of preoperative diagnose.Methods Clinical and MSCT finding were retrospectively reviewed in 14 patients with CRCC,which were confirmed by surgical pathology or biopsy.Results (1)On plain scanning,most of the CRCC lesions showed round or oval (86.7%,13/1 5)and isodensity (73.3%,1 1/1 5)mass;and the border of 66.7% (10/1 5)lesions were clear.53.3% (8/1 5)of the lesions were heterogeneous with lamellar cystic low density (8 lesions)and coarse ringed and/or foliated calcifications (4 lesions). (2)On contrast-enhanced sacanning,the CRCC lesions showed mild to moderate (86.7%,13/1 5)and heterogeneous (60%,9/1 5) enhancement,and 73.3% (1 1/1 5)of the lesions were persistent enhanced.(3)According to the pathology,80% (12/1 5)of the le-sions was the typical type,20% (3/1 5)was the eosinophilic type,and 0 was the hybrid type.Conclusion CRCC demonstrates cer-tain characteristics signs at MSCT examination.Lesion mostly shows well-circumscribed round or oval mass in the renal parenchy-ma,mild to moderate,heterogeneous and continuous enhancement on the contrast scanning.The diagnosis of CRCC should be con-sidered especially when the lesion has cystic change and coarse ringed and/or foliated calcifications.
ABSTRACT
Objective To describe the radiological features of mastitis.Methods Imaging findings were retrospectively reviewed in 58 women with mastitis,which was confirmed by pathological results.The imaging features were described according to the BI-RADS (breast imaging reporting and data system).All 58 patients had preoperative MRI,while 49 of them were examined with sonograms and 30 with mammograms preoperatively.Results Overall,no remarkable findings were reported on 6 mammography and 2 sonography examinations,but all 58 lesions were identified on MRI in our study.Asymmetrical density (n =16,53.3％) on mammograms and solitary or multiple separated/contiguous hypoechoic mass-like lesions (n =34,69.4％) on ultrasound were the most common signs of mastitis.On enhanced MR images,82.8 ％ (48/58) patients showed non-mass-like enhanced lesions.Multiple,regional enhancement (66.7％,32/48) and separated or contiguous,clumped,rim-like enhancement (63.1％,31/48) were the most common manifestations in non-mass-like enhanced lesions.Of the 58 patients,type Ⅰ and type Ⅱ timesignal intensity curve were detected in 48.3％ and 50.0％ of the patients,respectively.BI-RADS 0 had the highest frequency in reports on mammography (33.3％,10/30),while category 4a was most frequent on sonography (36.7％,18/49) and also on MRI (56.9％,33/58).Conclusion The imaging findings from mammography and US are non-specific for mastitis,therefore,MR can be helpful in the diagnosis,especially in the presence of non-mass-like enhancement that are multiple,regional,separated,or contiguous,clumped,and tim-like.
ABSTRACT
Objective To evaluate the value of quantitative 3T dynamic contrast enhanced MRI in the diagnosis of breast lesions. Methods One-hundred and eighteen patients suspected of breast lesions underwent MRI examination. A 3.0 T MR scanner was used to obtain the quantitative MR pharmacokinetic parameters: Ktrans( volume transfer constant), Kep (exchange rate constant) and Ve (extravascular extracellular volume fraction). The mean Ktrans, Kep and Ve of malignant, benign and normal glandular tissues were calculated and compared each other using LSD method. Independent sample t test was used between invasive ductal carcinoma and ductal carcinoma in situ (microinvasion included). Finally, the areas under the ROC curve (AUC) of Ktrans, Kep and Ve between malignant and benign lesions were compared. Results The mean Ktrans, Kep and Ve of malignant lesions (n=87) were (1.010±0.580) min-1, (1.634 ± 1.481) min-1 and (0.735 ±0.273); the mean Ktrans, Kep and Ve of benign lesions (n=23) were (0.331±0.192) min - 1, (0.417±0.324) min - 1 and (0.847±0.291); and the mean Ktrans, Kep and Ve of normal glandular tissues (n =83) were (0.051 ±0.028) min-1, (0.133±0.125) min-1 and (0.597±0.354), respectively. There were significant differences between normal glandular tissues and benign lesions, normal glandular tissues and malignant lesions, benign and malignant lesions in Ktrans (t=9.681, 11.189, 5. 590, respectively, P ＜ 0. 01 ), normal glandular tissues and malignant lesions, benign and malignant lesions in Kep(t =5. 287, 3. 874, P＜0. 05). There were a statistic differences between normal glandular tissues and benign lesions, normal glandular tissues and malignant lesions in Ve(t =2. 932, 2. 562 ,P ＜0. 05). There were no significant differences between normal glandular tissues and benign lesions in Kep, benign and malignant lesions in Ve ( t = 0. 760, 0. 832, P ＞ 0.05 ),invasive ductal carcinoma and ductal carcinoma in situ (microinvasion included) in Ktrans, Kep and Ve(t =0.834,0.075,0.454,P＞0.05). The areas under the ROC curve (AUC) of Ktrans, Kep and Ve between malignant and benign lesions were 0. 934, 0. 941 and 0. 659. The sensitivity of Ktrans, Kep and Ve were 77.01% ,91.95% ,56. 32% and the specificity of Ktrans, Kep and Ve were 95. 65%, 86. 96%, 78.26% for the differential diagnosis of breast lesions if taken the maximum Youden's index as cut-off. Conclusion The differential diagnosis of benign and malignant breast lesions by Ktrans, Kep is applicable.
ABSTRACT
ObjectiveTo study the MR Imaging features of breast phyllodes tumor (PT),and to correlate it with pathological results.MethodClinical and MRI findings were retrospectively reviewed in twenty-seven women with 28 PTs lesions confirmed by surgical pathology.Statistical analyses were one-way ANOVA for size analysis,Fisher exact test for analysis of MR appearances and Spearman correlation to study the relationship between MRI findings and BI-RADS categories.Results( 1 ) The histologic findings were benign,borderline and malignant PTs in 14.3％ (4/28),53.6％ (15/28) and 32.1％ (9/28) of lesions,respectively.(2) The mean maximum-diameter were (6.4 ± 3.9) cm,(5.7 ± 2.2) cm in borderline type and (4.8 ± 1.8)cm in benign type respectively.The results showed differences in lesion's size among the three type (F = 287.541,P =0.000),especially between malignant and benign type (P = 0.033 ).(3)Internal non-enhanced septations and silt-like changes on enhanced images,as well as time-signal curve on MRI correlated significantly with the histological grade ( P ＜ 0.05 ).( 4 ) If the category BI-RADS ≥ 4a was considered to be a suspicious sign for malignant lesion,the diagnostic accuracy of MRI would be 96.4％ (27/28),and the BI-RADS category of the MRI could reflect the PT's histological grade with a low correlation coefficient ( r = 0.382,P = 0.045 ).Conclusion The findings of PT on MRI have some characteristics,with tumor size and several MRI features correlating with the histological grade of breast PT.
ABSTRACT
Objective To evaluate the feasibility of sentinel lymph node(SLN) mapping with CT lymphography (CT-LG). Methods Twenty-five patients with confirmed breast cancer and no palpably axillary lymph node underwent CT-LG examination. The first one or more lymph nodes along the lymph duct draining from the injection sites to axilla were defined as SLNs, and then the LG results were compared with sentinel lymph node biopsy (SLNB). The over- and underestimation of LG were evaluated. The quality of LG imaging was classified Grade Ⅰ and Ⅱ according to lymph duct appearance on volume rendering (VR). The body mass index (BMI) ＞ 25 was considered obesity. Fisher exact test was used for the statistics. Results (1)Of 25 patients, 5 had local mastectomy history. BMI ＜ 25 was found in 20 cases, and ≥25 was in 5 cases. (2) All SLNs were showed by CT-LG, and Grade Ⅰ and Ⅱ imaging quality were achieved in 21 cases ( 84. 0% ) and 4 cases ( 16. 0% ), respectively. The obese patient tended to have a poor imaging quality ( P ＜ 0. 05 ). (3) Fifty-six SLNs and 45 lymph ducts in all 25 patients were identified on CT-LG. Compared with the results of SLNB, 7 cases ( 28. 0% ) and 9 cases ( 36. 0% ) were over- and underestimated respectively by CT-LG due to obesity and local mastectomy ( P ＜ 0. 05 ). (4) Fifty-two negative SLNs in 18 patients and 15 positive SLNs in 7 patients were confirmed by pathology through SLNB, while 56 SLNs were delineated on CT-LG with 43 negative and 13 positive. The shape in 32. 6% of the negative SLNs (14/43) and 76. 9% of the positive SLNs (10/13) was round,the difference was significant (P＜0. 05). The filling defect on the center in 9.3% of negative SLNs (4/43) and 23. 1% of positive SLNs (3/13) was demonstrated, and irregular filling defect on the margin was found only in 30.8% of positive SLNs (4/13). 3 SLNs in 2 patients combined with small satellite lymph nodes on CT-LG were also confirmed to have tumor infiltration. Conclusion CT-LG can clearly demonstrate the breast lymphatic pathway and may potentially be used for breast SLN mapping, while the imaging quality can be influenced by the obesity and local mastectomy.
ABSTRACT
llary-like mural nodules of PTC showed less enhancement on post-contrast CT and cervical lymph node metastases were more frequently seen in PTC.
ABSTRACT
Objective To evaluate the diagnostic values of the breast imaging reporting and data system-MRI (BI-RADS-MRI)description about non-masslike enhancement by data mining. Methods Fifty-five patients with non-masslike enhancement lesions showed on breast contrast-enhanced MRI were evaluated using two data mining algorithms (Logistic regression and decision tree) and 10-fold cross-validation methods. Results There were 28 malignant and 27 benign lesions. The most frequent findings of the malignant lesions were clustered ring enhancement and clumped enhancement [ 12 and 4 lesions, respectively; 84. 2% (16/19) in decision trees, partial regression coefficients in Logistic model were 2. 128 and 1.723, respectively], whereas homogenous, stippled, reticular internal and linear ductal enhancement were the most frequent findings in benign lesions [ 4、9、1 and 7 lesions, respectively; 72. 4% (21/29) in decision tree, partial regression coefficients in Logistic model were 0.357 (homogenous), 1. 861 (stippled) and 18. 870( reticular), respectively]. 10-fold cross-validation indicated that decision tree (C5.0) achieved an accuracy of 69.3% with a sensitivity of 66.7% and a specificity of 71.7% in comparison to the Logistic regression model with an accuracy of 57. 0%, a sensitivity of 43.3% and a specificity of 71.7%. Conclusions The diagnosis efficacy of non-masslike enhancement interpretation according to BI-RADS-MRI is not high. It is very important to find more potential features of non-masslike enhancement to improve the diagnosis accuracy.